Injecting prior to turbo comp' impellers

[JohnA]

This is starting to look quite interesting. :smile:

I've booked mine on a hub dyno on the 17th of September, where I will do a few runs testing various W.I. combinations, along with W.I. and fuelling matches too. (lean, normal and rich fuelling with W.I.)
All precompressor of course, 15-10% water/fuel.

These dynos produce very repeatable results, so hopefully I'll have some meaningful graphs afterwards.

[hotrod]

I like the idea of a coaxial spray nozzle !!!
That support ring looks nice --- you planning on making those up in various diameters ?

How about a design with a slightly larger outer ring width (broader) designed to be slipped into a flexible hose and held in place by clamping the outside of the hose with a radiator style clamp. That would allow other placements besides precompressor like centrally located in a front mount intercooler tube.

One other possiblility would be to include the feed line into one of the support struts so the support strut brings the hose through the outer ring, with a bulkhead hose fitting on it. That would pretty much elminate hose whip I would think.

Larry

[Richard L]

Probably need someone to chime in to give some suggestion as to what was happening. At the sametime, a slightest knock was registered.
Here is a superimposed image of base (faint orange) + Pre-turbo only (full colour)
http://www.aquamist.co.uk/forum/gallery/gtart/pre-turbo+base.jpg

What caused the a/f ratio to change? injecting about 75cc/min water per 400FWHP.

[JohnA]

So what are looking here at?

I can see two AFR lines, three torque curves (without a labeled axis), and two power curves.
Is that right?

I've had some interesting dyno printouts myself this weekend, turns out I need to use MORE water he he... :cool:

[Greenv8s]

I can see ... three torque curves (without a labeled axis), and two power curves. Is that right?


I think there are two torque curves, and two boost curves. (The second grey line is so feint that it is almost invisible.)

[JohnA]

I think there are two torque curves, and two boost curves. (The second grey line is so feint that it is almost invisible.)
boost curves?
I see none of those, there is not even a boost scale...

[Richard L]

Here is a better one.


http://www.aquamist.co.uk/forum/gallery/gtart/pre-turbo+base2.jpg

[Greenv8s]

Some further thoughts:

Given the very close correlation between the two runs at lower rpms, I'm inclined to think that the variation at the top represents a genuine change in performance (rather than just measurement error, for example).

The behaviour around 7500 rpm is particularly interesting. Without water injection, the torque and boost diverge and the AFR shows that it richens up. Is this a MAP based system? If so, this might be evidence that the charge temperature is going up at this point (less mass air flow for a given absolute pressure). With upstream WI the effect is there but is much smaller.

[Richard L]

A close-up of that region...

http://www.aquamist.co.uk/forum/gallery/gtart/pre-turbo+base2mag.jpg

[JohnA]

So we're talking here, what, 10bhp extra only at the highest revs?

My dyno printouts are quite different, the precompressor injection effects are striking.
I guess it depends on how much 'over the edge' the compressor is. Well, that's what I thought until I saw further gains with W.I. well inside the compressor's efficient island.

You HAVE to lean the mixture though, or else you lose power (I did that too on the dyno)

[Richard L]

We didn't have much time, we just funished tuning the non-WI map and just switch on the pre-turbo as "wait and see". We have to move on to post turbo (will print later) and combined etc.

We may not have inject enough water. I guess we were injecting some 4 litre of fuel per minute so w/f ratio is 150cc/4000cc X 70% DC = 2.6%

Not really a great deal of water. We wil be continuing as soon as we have time.

[Richard L]

So we're talking here, what, 10bhp extra only at the highest revs?

My dyno printouts are quite different, the precompressor injection effects are striking.
I guess it depends on how much 'over the edge' the compressor is. Well, that's what I thought until I saw further gains with W.I. well inside the compressor's efficient island.

You HAVE to lean the mixture though, or else you lose power (I did that too on the dyno)

Any chance of posting your dyno plot?

[JohnA]

We may not have inject enough water. I guess we were injecting some 4 litre of fuel per minute so w/f ratio is 150cc/4000cc X 70% DC = 2.6%
Yeah, if you run as rich as most people do (looking by the AFRs on that diagram), then to get 700horses you may have to use as much as 4000cc/min.

But where did the 150cc/min come from?
None of the 'normal' Aquamist nozzles gives out so little, right?

As for my dyno graphs, I've got to scan them first and highlight the relevant curves, there are a few runs on each printout.
I'm gonna need another long session though, running more boost, and more water *hint* *hint*

[Richard L]

We are running twe 0.3mm jet, one in front of each turbo. At 100% DC, each jet passes 75cc/min of water. Since we are only running 70% fuel dc, so 75+75 x70% = 105cc/min.

Fuel flow is estimated at 4000cc/min. so w/f ratio is 105/4000= 2.6%.

As soon as I have time, I will start increase the jet size. The effect will be more dramatic for sure. At 2.6%, the effect is already very noticeable. Doesn't really know what was going on.

Do you have your dyno plot at hand? If you can't post it, please describe it.

http://www.aquamist.co.uk/forum/gallery/gtart/32.jpg

Ay 2.6% w/f and 11.5a/f ratio, the amount injected is less than a rainy day !! Quick calculation: 0.22% of water of total mass inhaled by the engine.

[JohnA]

We are running twe 0.3mm jet, one in front of each turbo. At 100% DC, each jet passes 75cc/min of water. Since we are only running 70% fuel dc, so 75+75 x70% = 105cc/min.
0.3 jets, are they now available then? I thought 0.4 is the smallest (what I was using)
Also in my experience the 0.4 flows more than the rated 150cc/min. I use up easily half a litre after a blast on the road, or a few dyno runs. And that's injecting at 14psi

Fuel flow is estimated at 4000cc/min. so w/f ratio is 105/4000= 2.6%.

my w/f ratio was higher than that, and still was nowhere near enough :lol:

I'd need two things at this stage:

1. the Aquamist turbine gizmo that measures waterflow. I wouldn't dare lean out further without it.

2. that screw-in adaptor to secure the nozzle on a silicon hose. Can't really swap nozzles on the dyno without one.

Let's hope that the Aquamist R&D fund can stretch that far, he he... :wink:


Do you have your dyno plot at hand? If you can't post it, please describe it.
I saw clear power loss when injecting water with stock fuelling (no surprises there, eh?)
I saw clear power gains when leaning the fuelling and using WI on top
I saw greater power gains when injecting water at 8psi instead of 14psi (that WAS a suprise to me, I expected it to be the same or worse)

It took me a few hours on the dyno overall, this is not a cheap sport. I definately need more runs now, knowing what I know.

[Richard L]

0.3 jets, are they now available then? I thought 0.4 is the smallest (what I was using)
Also in my experience the 0.4 flows more than the rated 150cc/min. I use up easily half a litre after a blast on the road, or a few dyno runs. And that's injecting at 14psi


The 0.4mm jet flow 180cc/min - it has been for a number of years.


I'd need two things at this stage:

1. the Aquamist turbine gizmo that measures waterflow. I wouldn't dare lean out further without it.

2. that screw-in adaptor to secure the nozzle on a silicon hose. Can't really swap nozzles on the dyno without one.

Let's hope that the Aquamist R&D fund can stretch that far, he he... :wink:


1. The flow sensor is available on its own, require a volrtage meter and DDS3 is not used.

2. The adaptor is a press-on type, only abailable for 6mm thick hose - too loose on 5mm hose.


I saw clear power loss when injecting water with stock fuelling (no surprises there, eh?)
I saw clear power gains when leaning the fuelling and using WI on top
I saw greater power gains when injecting water at 8psi instead of 14psi (that WAS a suprise to me, I expected it to be the same or worse)


If pre-turbo cooler rate is increased, do you think it is able to use up the excess fuel to gain extra power? If this is the case, pre-turbo has in-build failsafe mechanism.

Richard

[JohnA]

The 0.4mm jet flow 180cc/min - it has been for a number of years.I'd swear it is over 200cc/min, 250cc/min perhaps.
But I haven't measured it precisely with the precompressor setup running in real situation, so can't tell for sure.
The voltage is different under those conditions, and so is the pressure difference across the nozzle.
Holding it full boost for 15 or 30seconds in one go is out of the question too, I only have it injecting once both turbos are on-song, and speeds would get silly indeed.

I've swapped it for a 0.6mm now and see how we're doing.


2. The adaptor is a press-on type, only abailable for 6mm thick hose - too loose on 5mm hose.
I'm not sure what the thickness of this hose is, it is not silicone though, more like plastic-rubbery. I've managed to hack some traces of thread (courtesy of the Aquamist tap!) on a steel hoseclip, which is secured right on top of the hose hole.
So the nozzle is much stabler now, it should do the trick for a while, until you guys bring press-on adaptors for thinner hoses.

There is always vacuum in that hose, so it cannot pop off, it's just some mechanical stability that it needs

If pre-turbo cooler rate is increased, do you think it is able to use up the excess fuel to gain extra power? If this is the case, pre-turbo has in-build failsafe mechanism.

Richard
What excess fuel, you have to get rid of all excess fuel for water injection to work properly. That's what I've been saying all along :wink:
You have to run LEAN under full boost.
It's not for the faint of heart.

[Richard L]

Can you return the 0.4mm jet and we can test it for you, I will replace it with a new one.

0.6mm: - we have got that far yet, you must be getting some exceptional response, based on the 0.3mm jet we tried.

Excess fuel: - almost all factory turbo cars run rich, so if pre-turbo is so effective to lean-off a/f ratio, then there is little tuning needed to gain power plus point is the in-built fail-safe mechanism?

[JohnA]

...0.6mm: - we have got that far yet, you must be getting some exceptional response, based on the 0.3mm jet we tried.

The power you're making is about 200-250bhp higher than what I'm working on, so my water/fuel ratio is even higher.

Also I am on *stock* turbos, that are outside their efficiency range from 1 bar upwards, so my gains could potentially be higher.
almost all factory turbo cars run rich, so if pre-turbo is so effective to lean-off a/f ratio, then there is little tuning needed to gain power plus point is the in-built fail-safe mechanism?
It all depends on HOW MUCH rich we're talking about.
Mine for example runs AFRs of 10.0 at 1.1 bar with stock fuelling.
Simply adding modest precomp WI (trigger at 14psi) on to this setup loses me 10bhp and 10lbft (around 3.5% loss). Boost also drops by 0.05~ 0.1bar with W.I.

If you're already running rich, W.I. makes things worse
Running even richer (after an ECU reset) doesn't even stop it from detonating, mine pulled timing immediately despite the WI (at 8psi) and running rich as hell. Resetting the stock ECU also resets the timing advance to the max values, so first time round the ECU 'hunts' for det, which it found.

I was running 99RON fuel by the way, Tesco's finest, nowhere near race fuel --- I want realistic figures, without fancy FMICs, airfilters, open bonnets etc.

[JohnA]

After a few timed runs on the 0.6mm nozzle, the results are not very encouraging.
The difference of water injection in total is minimal 1~2%, barely measureable in fact. You only get on-the-wheels figures with the accelerometer, but using the same stretch of road same day etc gives very repeatable results (not nice curves as the dyno though!)

I thought that maybe the higher boost pressures would be happy with the bigger nozzle, but apparently I went too far.
It didn't even *feel* different either (the 0.4mm nozzle felt different back to back)
Good thing perhaps, because it used around 400cc during 5 runs of 40-100, quite a lot actually...

Fitting the 0.5mm next time, and see how that goes...
(I've measured again the flow of the 0.5mm nozzle, it is 240cc/min with the engine on idle, it would be a bit more with a strong vacuum inside the pipe, wouldn't it?)

[Richard L]

I wonder if you need to make some changes on your igntion as well as fuel to make it more effective, with the 0.6mm.

It is interesting that a 0.4mm is more effective. I am sure the vacuum will make it flow more.

Richard

[JohnA]

I wonder if you need to make some changes on your igntion as well as fuel to make it more effective, with the 0.6mm.
I don't have direct control over the ignition right now, apart from resetting the ECU so that the max advance 'targets' are re-established.

It is interesting that a 0.4mm is more effective.
I've got the 0.5mm now, datalogged some runs on the accelerometer.
More boost too, 18-19psi
Haven't transfered the results yet.
I am sure the vacuum will make it flow more.
Richard
I wonder if we can use the pressure formula we use for fuel injectors:

If the ERL pump operates at 10bar normally, we could say that at 1 bar the pressure across the nozzle is 9 bar.
Precompressor, the equivalent pressure would be 10.5bar perhaps.

Any ideas why the above thinking might be wrong?

[JohnA]

After entering the results on Excel, here is a summary from the accelerometer datalogging (mainly from 40-100 runs, power estimated at 10mph intervals)

Baseline: car running stock fuelling, stock boost and no WI

Stage1: fuelling trimed to an average 12-12.5 AFR, 7psi extra boost, no WI

Stage2: as stage1, with 0.5mm nozzle injecting at 13psi boost (when both turbos are singing)

Average bhp gains from Baseline to Stage1: 10.36%

Average bhp gains from Stage1 to Stage2: 7.85%

Timewise, on that particular stretch of road (not totally straight and slightly uphill perhaps) 40-100 through the gears, same way every time:

baseline: 9.94secs
Stage1: 8.61 secs
Stage2: 7.97secs

[Richard L]

After entering the results on Excel, here is a summary from the accelerometer datalogging (mainly from 40-100 runs, power estimated at 10mph intervals)

Baseline: car running stock fuelling, stock boost and no WI

Stage1: fuelling trimed to an average 12-12.5 AFR, 7psi extra boost, no WI

Stage2: as stage1, with 0.5mm nozzle injecting at 13psi boost (when both turbos are singing)

Average bhp gains from Baseline to Stage1: 10.36%

Average bhp gains from Stage1 to Stage2: 7.85%

Timewise, on that particular stretch of road (not totally straight and slightly uphill perhaps) 40-100 through the gears, same way every time:

baseline: 9.94secs
Stage1: 8.61 secs
Stage2: 7.97secs

John, those were impressive gains. It seemed that there is a limit how much water can be injection pre-turbo before the gain becomes a loss.

Testing on the sam epiece of road is pretty consist than most people would give credit for. You have spare yourself a great deal of dyno time and cost.

I am not sure if it will help if I lend you a flowmeter, all you need is a DVM to get the flowrate reading, evey little thing helps. PM me your address.

Richard

[JohnA]

John, those were impressive gains. It seemed that there is a limit how much water can be injection pre-turbo before the gain becomes a loss.
Oh, it definately works.
I saw that on the dyno as well, there is no question about that.
The devil is in the detail though, I want to know *how much* and *when* so I can optimise the setup for maximum gains with minimal water usage.

Testing on the sam epiece of road is pretty consist than most people would give credit for. You have spare yourself a great deal of dyno time and cost.
The weather helps lately, if it is wet it will not be consistent enough (and bloody dangerous on a public road!)
The dyno gives you a nice curve with AFR info as well, no wheelspin, it's very good.
The road datalogging is good for testing more roughly, to see if you're moving in the right direction.

I am not sure if it will help if I lend you a flowmeter, all you need is a DVM to get the flowrate reading, evey little thing helps. PM me your address.
Flow meter is one, the other is trimming WI according to the fuel injection pulse. At the moment it's not a huge problem because my WI is between 4K and 6.5K so the 'dynamic range' isn't that wide. My water/fuel ratio ranges from 10% to 15% or so (with my basic aquamist system)

I'll send you a PM when I've decided on what sort of setup I'll go for in the near future. Maybe using 2x0.3mm nozzles perhaps...

[ctischmick]

OK so I read through all 19 pages today...
I went from excited to mixed feelings...
Someone should really update the first page with the more recent results...

People were predicting 30% gains, but in reality you are seeing about 7-10%...

I think I'm going to try it on my car soon.... We'll see how it goes... I have some good ideas.... I love that little contraption you made...

Also could you please tell me what size threading on a bolt, is equivalent to an Aquamist nozzle? I think I already have a pre made tap for an Aquamist nozzle on this used upp pipe I have laying around...

[JohnA]

People were predicting 30% gains, but in reality you are seeing about 7-10%....
people in the past have predicted the end of the world (more than once!) but I don't feel let down because it hasn't materialised yet. :D

[Richard L]

Here is a plot when we run 50% methanol + some timing and fuel trim, the gain is more noticeable - still using the same jet sizes - Port +Pre-turbo.

We will move on to bigger jets as time permits very soon.

http://www.aquamist.co.uk/forum/gallery/gtart/all+methanol.jpg

[JohnA]

Richard, does this run SIX 0.4mm nozzles at the throttle body or is it a typo?

[Richard L]

The skyline runs 6x 0.4mm at each port and 2x 0.3mm pre-turbo.




http://www.aquamist.co.uk/forum/gallery/gtart/50.jpg

[JohnA]

I see that the AFR at high revs hovers around 11:1

Have you tried to go leaner? I got most of the gains when being leaner than 12:1, all other factors being the same.

Also is it running race fuel? If so, the gains from these four nozzles might be more modest, especially if the FMIC is really good.

What sort of charge temps does it register after a few seconds at full boost?

[Richard L]

I see that the AFR at high revs hovers around 11:1

Have you tried to go leaner? I got most of the gains when being leaner than 12:1, all other factors being the same.

Also is it running race fuel? If so, the gains from these four nozzles might be more modest, especially if the FMIC is really good.

What sort of charge temps does it register after a few seconds at full boost?

At the time of the test, we have no information of the actual WI percentage to fuel so we were a bit careful until we have calculated the correct afr.

Knowing the sto for methanol is 6.6:1 and gasoline 14.7:1, the correct a/f ratio of the mix has to reflex the mix, is that correct?

We will resume the test as soon as time permitting. This time we will plan it first compared to tuning the engine for race in 12 hours ahead. The next test will all the temperature sensors installed and logged. At preset I have a few big projects to complete.

The entire test was run on pump fuel - race fuel was not permitted on the race.

[JohnA]

At the time of the test, we have no information of the actural percentage to fuel so we were a bit careful until we have calculated the correct afr.

True, but at 50% methanol the Water/Fuel ratio has to be over 20% for the shift in the stoich point to become an issue, right?
Also the more the alcohol percentage, the less the atomisation deficiencies of petrol, and the cooler the burn, so one can go even closer to stoich at full boost.
The entire test was run on pump fuel - race fuel was not permitted on the race.
Pump fuel, nice, so we can see some of the 'octane boost' effects in action! :D

[Richard L]

When time permitting, I will get a set of logged data published here so you can see the effect.

Please don't treat the published power output is the final result - there are lots and lots of testing, furling and timing optimisation to be done yet.

I was merely show the dramatic effect on the a/f ratio with just water alone - still puzzled.

I am hoping to see a lot more gain form now on - since we have the use of a dyno and a twin pump aquamist system. Almost unlimited supply of methanol, nitromethane etc - just needs to find time.

Watch this space.

[JohnA]

...Watch this space.

http://www.max-boost.co.uk/stuff/alex_watching.jpg

[Richard L]

...Watch this space.

http://www.max-boost.co.uk/stuff/alex_watching.jpg


I getting to like this picture a lot.

Nothing to report yet.

[espritGT3]

[quote="Richard L"][quote="JohnA"]I see that the AFR at high revs hovers around 11:1

Have you tried to go leaner? I got most of the gains when being leaner than 12:1, all other factors being the same.

Richard,

Of course it will vary the amount of fuel and water being injected, but on my personal car toggling the WI on and off results in about a 0.04 increase in lambda (~0.6 A/F), and I'm running nowhere near 20% water/fuel.

WI appears to have a significant impact on lambda readings. I mention this only to point out that the Skyline may not be running as rich as it may first appear.

It would be interesting if you could perform a similar test to gather some data on the WI/ effect on lambda readings for comparison.

Mike

[espritGT3]

[quote="Richard L"][quote="JohnA"]I see that the AFR at high revs hovers around 11:1

Have you tried to go leaner? I got most of the gains when being leaner than 12:1, all other factors being the same.]

Of course it will vary the amount of fuel and water being injected, but on my personal car toggling the WI on and off results in about a 0.04 increase in lambda (~0.6 A/F), and I'm running nowhere near 20% water/fuel.

WI appears to have a significant impact on lambda readings. I mention this only to point out that the Skyline may not be running as rich as it may first appear.

It would be interesting if you could perform a similar test to gather some data on the WI/ effect on lambda readings for comparison.

Mike
espritGT3

[redvictor]

this is a superb topic with excellent info.I have some questions.Firstly i will introduce myself.
I have been running a 1550hp nitrous 600ci V8 in a class called Street Eliminator.Best time and speed in the 1/4 was 8.2 at 177mph in a 3000lb street legal car.Bearing in mind we have to use Optimax(with a commercial booster) and block treaded tyres. We have used the ERL system on the car but found after a while we didn't need it with the nitrous as detonation wasn't a problem(even with 600+ nitrous injected in 2 stages). I am now putting on 2 x 88mm turbos on a 572ci V8 that has to be run in the same class. We will be shooting for 2000+hp on Optimax and tetraboost. I am also going to use an air to air cooler because of events like the king of europe which has a 20min turn around and it would be impossible to use ice water in an air to water system. I will be removing the port WI system and moving the nozzles around.i also have a plate under the throttle body with 4 nozzles in that i will keep.these are .8mm. the nozzle system uses .4mm x 8.I will also inject a couple of .4 nozzles after the intercooler.Now,the pre turbo injection seems like a fine idea but what nozzle size would i need,and how many? How far away from the compressor are you injecting from? straight shot at it or against the tube wall?Also has anyone injected WI just inside the intake pressure port of the turbo? I know from previous experience that a 50/50 mix of water/methanol works great injecting with an rpm switch,and a sureflo pump and solenoid(which i still have). Also bear in mind that the water injection system would only be on for about 10seconds max in the burnout and low 8 to high 7 sec in the run.
Your input and info would be highly appreciated on this one.
BTW,a piccy of one of the turbos for the install...and yes,Holset have sized the turbo for the application,and are giving me their valuable input...
ANDY FROST.
http://pic.piczo.com/img/i32663166_31036_2.jpg
PS,could someone post a pic of a pre turbo nozzle placement?

Visit us at www.redvictor1racing.co.uk

[JohnA]

Injecting straight into the 'eye' of the compressor is not a bad way to go.

PS
that's a small mug you've got there (or is it a big turbo? hard to tell, lol...)

[redvictor]

thanks for the reply John.Now if i'd have looked at the rest of the topic before i'd posted i could have got most of my answers...
Did Richard ever get those centre jet aluminium castings done?
The thing is that injecting into the centre of the turbo on this application would be quite difficult without something of that nature being used.......unless i use a 90deg piece of tubing that turns close to the impeller...then that would restrict the airflow i guess... maybe i'll get my machinist to knock something up...
Btw John,turbo is 10" across on the inlet,cup is normal... :lol:

[cheekychimp]

Red,

How do you find the Holset turbos for spool up and do you think their legendary durability makes them better suited to this type of pre turbo injection?

Paul

[redvictor]

I'll tell you when they are running..... I picked Holset for 3 reasons.Firstly they are on a million diesel engines that last forever,and they were able to custom build and size for my application,and they are here in the UK........
IMO the 360deg bearing that is in these turbos is so much more durable than any roller bearing.It's force fed with a constant oil supply for starters.

[sdminus]

how is the car getting on ? Are you still injecting pre turbo ?

Scott

[cheekychimp]

I was a bit wary of posting this since this is a water injection specific thread, but this is something very relevant to me so I wanted to ask you guys since you are very knowledgeable.

I have in my possession a propane injection kit that was put together for me by a very well known injection specialist best known for their nitrous systems.

Having read through this thread I was wondering how effective it would be to inject propane pre turbo. Injected as a liquid under pressure it would turn gaseous 'instantly' especially since it boils at less than zero (centigrade). Whilst it may not offer much cooling as a result, it would consequently have no opportunity to condense in the intercooler, would offer a denser air mixture at the turbo, increase octane content, would not interfere with a hotwire MAS in blow through and could theoretically be used in tandem with water/methanol injection at the throttle body.

What do you guys think?

[rarson]

That's not a bad idea. Anyone else want to comment on that?

[Richard L]

Propane is normally injected as a gas once it leaves the nozzle (restrictor). Almost turns to gas instantly due to pressure drop on the other side of the orifice.

Propane in gaseous form has little or no coolng effect. The resultant charge temperature will be the ratio of propane mass and air mass. Since you will not be injecting a huge amount of propane relative to air (mass% <10P/90A), the drop in charge air temperature is not that great - most people are reading into the marketing hype too much.

The flash point is at its most explosive when the A/P mixture is at around stoichometic and feeding the mixture pass a hot wire sensor can be risky. I am not an expert on this but I would ask the propane injection specialist first before implementing the set up on your car.

Richard

[JohnA]

This article towards the end mentions pre-compressor injection considered for hypersonic planes of the future. :lol:

http://www.afa.org/magazine/dec2003/1203spaceplane.asp

I guess we're on the right track, eh?

[Richard L]

This article towards the end mentions pre-compressor injection considered for hypersonic planes of the future. :lol:

http://www.afa.org/magazine/dec2003/1203spaceplane.asp

I guess we're on the right track, eh?

Excellent find.

I will gibve them a call :lol:

[EvoTio]

This article towards the end mentions pre-compressor injection considered for hypersonic planes of the future. :lol:

http://www.afa.org/magazine/dec2003/1203spaceplane.asp

I guess we're on the right track, eh?

This is some very interesting reading.

[redvictor]

it's been a while since i posted here but i have heard on good authority that high end turbo cars on pump gas are going with pre turbo injection.Richard,did you get those castings done? if not have you got an adaptor that will allow the plastic pipe to go through the inlet pipe without leakage of air?
some pics of the install as it is now,water injection is into the plate under the throttle body and into the intake pipe(not shown)....
http://www.phasechangeracing.com/pics/frosty28.jpg
http://www.phasechangeracing.com/pics/frosty27.jpg
http://www.phasechangeracing.com/pics/frosty25.jpg
http://www.phasechangeracing.com/pics/frosty29.jpg
http://www.phasechangeracing.com/pics/frosty8.jpg
http://www.phasechangeracing.com/pics/andyheader2.jpg
http://upload2.postimage.org/192001/intercoolerpic08.jpg (http://upload2.postimage.org/192001/photo_hosting.html)
http://upload2.postimage.org/192003/intercoolerpic05.jpg (http://upload2.postimage.org/192003/photo_hosting.html)
http://upload2.postimage.org/125934/_cid_002c01c62c44se8a0fa20s0502a8c0_acer b0474dc4.jpg (http://upload2.postimage.org/125934/photo_hosting.html)
http://upload2.postimage.org/192007/turboside10.jpg (http://upload2.postimage.org/192007/photo_hosting.html)http://upload2.postimage.org/192028/intercoolerpic07.jpg (http://upload2.postimage.org/192028/photo_hosting.html)
http://upload2.postimage.org/192029/intercoolerpic06.jpg (http://upload2.postimage.org/192029/photo_hosting.html)http://upload2.postimage.org/192038/intercoolerpic03.jpg (http://upload2.postimage.org/192038/photo_hosting.html)

[redvictor]

WI nozzle placement and BOV....
http://upload2.postimage.org/197201/bov04.jpg (http://upload2.postimage.org/197201/photo_hosting.html)
http://upload2.postimage.org/197205/bov02.jpg (http://upload2.postimage.org/197205/photo_hosting.html)
http://upload2.postimage.org/197210/winozzles04.jpg (http://upload2.postimage.org/197210/photo_hosting.html)
http://upload2.postimage.org/197215/winozzles03.jpg (http://upload2.postimage.org/197215/photo_hosting.html)
http://upload2.postimage.org/197217/bov05.jpg (http://upload2.postimage.org/197217/photo_hosting.html)

[Richard L]

Andy, what a nice set of pictures.

I look forward to read them carefully in the weekend.

Thanks for posting them.

[Richard L]

I was not able to see the images in too much details.

Did you allow the full pattern of the jet be developed - the core shape of the spray should not be touching any bore surfaces.

Here is a spray pattern of our jet (if they are the ones you used)

http://www.aquamist.co.uk/forum/85.JPG

[redvictor]

mmmmm,they look like the ones i have.......i'd say the bottom one no...the 2 side ones yes...maybe i can delete the bottom one and rely on the 2 side ones? don't forget there is 4 in the plate under the throttle body....and we can add a boss at the top of the pipe for a straight shot into the throttle body.....did you ever make those bosses for the pre turbo WI jet? i;d like to try that.....

[rarson]

Quick question: if the nozzle forms a "cone" spray pattern in a fine mist, wouldn't that mean that spraying directly at the turbo, even if it was mounted right on the inlet, would result in the water droplets hitting the blades?

I'm trying to imagine spraying at the compressor nut. I'm thinking a "stream" instead of a fog would be necessary, in order to fling the water off to aid atomization.

[NAnderson]

Digging this thread back up because it's quite interesting and informative.

This summer I'll be running a dual stage water/meth setup (1/3 water-2/3 meth) with one nozzle before the turbo, spraying directly into the compressor and one nozzle after the intercooler, spraying into the upper intercooler pipe. My question centers around the order of activation of these two nozzles. Should I activate the pre-turbo nozzle first (~15 psi) to cool the incoming charge and the post-intercooler nozzle second (~20-25 psi) to help with knock supression? Or am I thinking backwards on this? Any input or suggestions would be GREATLY appreciated.

I'm shooting for ~30-35 psi on a 2.3L (stroked 2.0L Mitsu 4G63) and would like to run 93 octane, but have access to 100 octane if necessary. I'd be glad to supply any additional info about the setup if it's needed.

[JohnA]

... My question centers around the order of activation of these two nozzles. Should I activate the pre-turbo nozzle first (~15 psi) to cool the incoming charge and the post-intercooler nozzle second (~20-25 psi) to help with knock supression? ...
Your logic makes sense.
You should also think about the relative size of the nozzles, as they are directly related. The more the first one sprays, the lower the compressor discharge temps are bound to be. Also some of that water will still make it to the chambers, reducing the need for a bigger nozzle post-intercooler. :wink:

Ambient temps and fuel octane also play a major role obviously, in the winter with race fuel you'll need smaller nozzles compared to pump fuel in the summer.

I'm experimenting with a similar setup by the way.

[NAnderson]

Your logic makes sense.
You should also think about the relative size of the nozzles, as they are directly related. The more the first one sprays, the lower the compressor discharge temps are bound to be. Also some of that water will still make it to the chambers, reducing the need for a bigger nozzle post-intercooler. :wink:
I'll actually be running two similarly sized nozzles, but being that the pre-compressor nozzle won't have boost working against it it'll see full pump pressure (if not a touch more due to vacuum). The post-compresor nozzle will have anywhere from 25-35 psi of boost pushing against it and thus won't "seem" as large as it actually is.

So John, would you recommend activating the pre-compressor nozzle first, and the post-intercooler nozzle second?

[JohnA]

..I'll actually be running two similarly sized nozzles, but being that the pre-compressor nozzle won't have boost working against it it'll see full pump pressure (if not a touch more due to vacuum). The post-compresor nozzle will have anywhere from 25-35 psi of boost pushing against it and thus won't "seem" as large as it actually is.

That's very true. The precomp nozzle will flow a bit more compared to a post-comp nozzle of the same size.
There is actually an older thread where we calculated that difference.
So John, would you recommend activating the pre-compressor nozzle first, and the post-intercooler nozzle second?
It depends on how hard you drive your turbo(s) I suppose.
If they are pushed way out of their efficiency islands then I'd start precomp injection a bit before they start getting outside the manufacturer's max boost intentions. But all this is experimental you see, we are all learning from our own (and those of others) mistakes. Hopefully.

[NAnderson]

It depends on how hard you drive your turbo(s) I suppose.
If they are pushed way out of their efficiency islands then I'd start precomp injection a bit before they start getting outside the manufacturer's max boost intentions. But all this is experimental you see, we are all learning from our own (and those of others) mistakes. Hopefully.
This will be a single 50-trim Garrett/Mitsu hybrid turbo (Forced Performance "Green").
http://linux.forcedperformance.net/merchant2/merchant.mv?Screen=PROD&Store_Code=FP&Product_Code=NTDSMFPGREEN&Category_Code=DSM-Turbo

At higher boost levels on a 2.3L the compressor will definitely start moving out of its efficiency range. Looks like I'll inject with the post-intercooler nozzle first, and then at higher boost levels I'll kick on the pre-compressor nozzle to move back into a more efficient island on the compressor map. God, I love this stuff! :lol:

[hotrod]

Looks like I'll inject with the post-intercooler nozzle first, and then at higher boost levels I'll kick on the pre-compressor nozzle to move back into a more efficient island on the compressor map.

That is the way I would do it. Only do the pre-turbo spray when it will do you the most good --- to get that last 10% or so of flow out of the turbo compressor.

Larry

[NAnderson]

http://www.aquamist.co.uk/forum/gallery/precooler/21.jpg

http://www.aquamist.co.uk/forum/gallery/precooler/22.jpg
Richard, just wondering if you've made any progress on these or if you plan on releasing them any time in the future. If so, you've got one interested buyer here! Also, what diameter(s) do they come in?

I'm dead-set on successfully using pre-compressor injection, just don't want to be replacing the compressor wheel every year or so. My setup as it stands now is very close to spraying directly into the compressor wheel, but there may be a bit of spray pattern to intake pipe collision as the nozzle is a bit farther away from the wheel than I'd like.

[Borgue]

hey guys,

i have been trying to tune my 1985 mazda rx7 which uses an elford turbo kit.

the set up uses an S.U. Carb before the turbo, which is a garret T3 A.I resurch which has sheilded beraings to protect from fuel contamination to the oil.

as a stock car the setup delivers 160Hp at 5psi of boost.

curretly my car is running 16 psi with no intercooler using pre turbo methanol injection, there is no lag. from cruse to full boost in one second.

the methanol has allowed me to use loads of timing and act as a fuel suppliment to allow me to assist in fueling the car.

i hope to get the car on a dyno soon to test its strength.

http://upload3.postimage.org/230613/elford1.jpg (http://upload3.postimage.org/230613/photo_hosting.html)


http://upload3.postimage.org/230615/elford5.jpg (http://upload3.postimage.org/230615/photo_hosting.html)

any one else playing around with anything similar?
:roll:
Ollie

[JohnA]

So you inject pure methanol pre-compressor?

It evaporates pretty damn fast, but I wonder if there are any safety concerns with pure meth. :?

I've been researching the subject heavily again, even bought papers on the subject. Those are aimed at aircraft or power generator compressors, but the fundamentals are the same. Demineralised water is the only substance used though in the experiments, so how methanol behaves is anyone's guess.

The water droplets boil by the way inside the compressor, once they enter the areas of extremely low pressure. So water never makes it into the diffuser, only steam. We're talking water/air ratios of up to 5% here, which is far more than any of us have used - I've only gone up to 1.5% (by weight)

Steam is useless for in-cylinder cooling by the way, there is no phase change left that we can take advantage of. I was suspecting that, hence my setup with an extra nozzle after the intercooler. Steam actually displaces oxygen in the chambers (not good) but because this happens after the compressor no oxygen molecules have been traded off. This may explain the boost increase which is observed as soon as precomp injection kicks in)

[sdminus]

Olly may not be able to reply at this time but he does inject pure meth pre compressor as well as meth with no intercooler. It works very well. He runs no BOV which limits the gear changes but over all gives him loads of scope.

Its true steam does nothing but methanol atomises at a lower temp and does not change state after atomisation
Scott

[simple]

JohnA,
Just trying to get my head around what you were talking about the steam.
1. port injection is good as cylinder get to see water droplets 5-20 micron in size
2. evaporative cooling is good as it cools the air down until saturation point is reached
3. steam what produced in intake system and not absorbed by air is bad as it ?replaces? some of the air in the system?
Basically speaking all excesses water that does not get absorbed due to the system been past saturation point and which boils up is bad?

[maxc]

Water replaces air too. It's harder for the frame front to travel around larger water particals.

[cheekychimp]

Okay, this may be somewhat of a new approach in some respects, but could those of you with some thermodynamics expertise bear with me on this one and give me some input.

Going right back to the beginning of this thread and trying to incorporate as much of the theory as I can from the wealth of knowledge that has followed on, I have grasped these main concepts.

(1) Pre-compressor injection is unique in that it can increase compressor efficiency by pushing compression away from adiabatic towards isothermic (I hope I got those terms correct);

(2) Whilst larger volumes of liquid injected pre compressor may reduce temperatures by a greater amount at the compressor, a point of diminishing returns can be reached because (a) this may reduce intercooler efficiency resulting in intercooler outlet temps increasing and (b) saturation occurs whereby no more water can evaporate and consequently no further heat can be absorbed.

(3) Therefore, ideally we want to inject just enough water to achieve (1) without inducing either of the diadvantages of (2);

(4) Water has the greatest ability of all substances injected to absorb latent heat; and finally

(5) In order to prevent or at least substantially reduce compressor blade damage we want to inject water droplets of the smallest size possible (ideally 10 microns or less and certainly no larger than 50 microns)


So, a few questions.

(1) Does a mist of water vapour at ambient temperature absorb any less heat when injected than say a mist of water that is chilled to 10 degrees celsius immediately prior to injection?

(2) If the answer to (1) is that cooler water absorbs more heat, then would water injected at it's lowest liquid temperature (i.e. 1 degree celsius) be more or less effective in absorbing heat than the same volume of an alcohol and water mixture that could be injected at below the freezing point of water?

The reason I ask this is that people have been injecting propane instead of water pre compressor for years. Propane enrichens the octane level of the final fuel mixture as well as expanding at a phenomenal rate which reputedly also helps distribute fuel in the combustion chamber better, which in turn leads to fewer hotspots.

As such it also suppresses detonation but NOT by reducing the temperature of the intake charge. Injected for the most part as a vapour all the cooling effect takes place at the propane container which becomes cold to the touch, but the cooling effect on the intake tract is minimal.

Okay, this is where it starts getting interesting. Forget safety issues for now, I am not interested in those for the purposes of this discussion. During actual application yes, but NOT here. I have a Propane Injection System which uses a solenoid at the injection point (like a nitrous system). The bottle is inverted and the delivery line once purged of vapour remains under pressure prior to injection. Propane is thus injected as a liquid. Propane 'boils' at -40 degrees (that coincidentally is roughly the same in either fahrenheit or degrees).

Since it becomes a gas almost instantaneously it probably won't absorb much latent heat - accepted

It does however have an octane rating of around 110 which means it really can replace any fuel it displaces although it takes up a considerable volume of 'airspace'.

So suppose we used a nitrous wet fogger injector which simultaneously injected liquid propane at 40 degrees below zero together with a water and alcohol mixture (composed of the greatest percentage of water possible that would not freeze) and injected it pre compressor.

I can see a number of possible issues;

(1) Catastrophic impeller damage created by a phenomenal change in temp from several hundred degrees to -40
(2) The mixture is unable to absorb heat;
(3) Temperature differential results in condensation and compressor damage;
(4) Loss of intercooler efficiency; and
(5) The fogging effect even under 400 plus psi fails to atomize the liquid to less than 50 microns.

However, given the number of devices used to reduce air intake temperatures and the advantages of isothermic compression, I would think that there must be some merit in this idea even if the injection point needs to be re-evaluated.

I'd value your comments.

Paul.

[JohnA]

JohnA,
Just trying to get my head around what you were talking about the steam.
1. port injection is good as cylinder get to see water droplets 5-20 micron in size
2. evaporative cooling is good as it cools the air down until saturation point is reached
3. steam what produced in intake system and not absorbed by air is bad as it ?replaces? some of the air in the system?
Basically speaking all excesses water that does not get absorbed due to the system been past saturation point and which boils up is bad?

two different things:

Port injection is a better version of injecting at the throttle: it provides mainly in-cylinder cooling

Precomp injection provides minimal in-cylinder cooling, but it reduces charge temps a lot and allows the turbo to spin slower for the same boost.

One does not exclude the other - i.e. you could have both. (I do)

[hotrod]

Interesting concept!

You would also have the issue of ice crystals forming during the evaporation process, the alcohol will evaporate more rapidly than the water so you get a distillation process as the mist moves down stream. An ice crystals that form would have a more agressive errosive effect on the compressor than the liquid water, so that is one issue to avoid.

With Propane injection you also have a balancing act going on between charge density due to cooling and oxygen dilution by the propane. This increases the volume of gas that must be compressed. As a result of that I would think you are looking for a happy medium where you get the benefits of some charge cooling, some increase in fuel octane, and better vaporization in the combustion chamber without over diluting the intake charge the turbo needs to compress.

Like you mention the cooling effect occurs at the point the liquid changes to gas, so by moving the solenoid to the injection point the cooling would move to the nozzle.
years ago I used to work in a gas station and one of my dutied every night was to blow off the compressed air in a pressure washer tank. In that case the point of expansion was at the outlet valve and that is where the frost formed. It might be ideal in your application to run the liquid propane through a heat exchanger as it expands to capture that cooling power and use a finned heat sink to cool the air charge.

Larry

[cheekychimp]

It might be ideal in your application to run the liquid propane through a heat exchanger as it expands to capture that cooling power and use a finned heat sink to cool the air charge.

Larry

Can you expand upon this Larry, and tell me how the heat exchanger would be linked to cool the air charge.

[hotrod]

Just brain storming a bit.

First you would need to do a bit of experimentation to find out how much cooling occurs near the solenoid and the line between it and the spray nozzle. If that section of line gets really cold you could wrap it in a coil around a segment of tubing and insulate the outside and put some fins on the inside. If that was placed in line with the air intake it would help cool the air charge by using cooling from the expanding propane that would normally be wasted.

If most of the cooling occurs at the propane bottle you would need to determine if there was any practical way to use that cooling to pre-cool the intake air by passing the air charge around the propane bottle. I can imagine a way to do it if you were using the smaller consumer propane bottles intended for propane torches, but not if you were using a proper DOT rated propane bottle in the trunk.

You could also use a wasted propane (or nitrous/CO2) system that just blew off some of the propane, nitrous or CO2 through a heat exchanger, but that would add some concerns about where the extra propane goes and would it be a fire hazard. This would be similar to the cryo cooler systems.

Larry

[cheekychimp]

You would also have the issue of ice crystals forming during the evaporation process, the alcohol will evaporate more rapidly than the water so you get a distillation process as the mist moves down stream. An ice crystals that form would have a more agressive errosive effect on the compressor than the liquid water, so that is one issue to avoid.

What if I injected methanol only? -40 wouldn't freeze alcohol right? I find this stuff interesting but I don't understand physics at all. Water has the ability to absorb far more latent heat than alcohol, but alcohol evaporates faster! So which is better in this situation?

[JohnA]

..Water has the ability to absorb far more latent heat than alcohol, but alcohol evaporates faster! So which is better in this situation?
It depends on the efficiency of your compressors and the effectiveness of your current intercooling (as well as how often your cylinder pressures exceed 'safe' levels, mainly Compression Ratio relating to boost)
You might want mostly alcohol before the cylinders and mostly water into the cylinders.

[RICE RACING]

i have only ever used total pre turbo injection (15% of fuel flow).

what i have noticed is no negative performance difference in heat exchanger, mostly increases

same boost
no WI 59deg c charge at end of 400m test
with pre turbo WI 45to 47 deg c same test

the water may boil in compressor but it condenses again leading into IC and def out of it, i can see suspended water vapour clouds in charge pressurized water tank which takes feed just before throttle body. So i get the turbo efficency increase and also in chamber cooling as well for what i have seen.

in practice have run up to 28psi boost on 9:1 comp wankel engine where others are limited to around 20psi maximum on very heavy AFR's and retarded timing, so it def seems to be working. power is there too over different cars which means its not a one off fluke result.

i use a special water/air atomizer on my "ghetto" systems, as a result no compressor wear at all after many 1000's of km in field use. hope thats of some use ?

[JohnA]

Very nice :smile:

[NAnderson]

i use a special water/air atomizer on my "ghetto" systems, as a result no compressor wear at all after many 1000's of km in field use. hope thats of some use ?
Could you shed some more light on this "atomizer?" A website or picture possibly? I'm just curious as to what others are using for nozzles for pre-compressor injection.

[fperra]

So much good information here, but no recent follow-up.

[untubbed_20]

http://www.spray.com.au/ss-redir-prod.htm
if you look under the catalogue you will find some info on the SUE18
and the SUE25BDF that others use

go catalogue> air atomizing nozzles>18J and 14J series
than they are either in pressure spray setups- internal mix
pressure spray setups- external mix
cheers
darren

[cheekychimp]

Theory:-

i) By using a turbo exhaust housing with a smaller AR we can help a large turbo spool faster, but the trade off is that the housing is unlikely to be able to flow enough air in the upper rpms to be able to realize the full potential of the compressor.

ii) By using pre-turbo water injection to move compression away from adiabatic towards isothermic we improve compressor efficiency enabling it to flow more air.

Question:-

If we reach a point whereby the exhaust housing is already creating a restriction to airflow, will pre-turbo injection

a) Have NO noticeable effect because the housing is unable to support any further increase in flow?; OR

b) Make the air denser thereby allowing more air to pass through the housing?

[cheekychimp]

I would also like to get hold of a table to give me an idea of what sort of volume of flow I can expect from running an air atomizer nozzle with 150 psi of liquid pressure and 400 psi of air/gas pressure.

I'm concerned that liquid flow will be dramatically reduced as a result of the very high pressure. I looked at the nozzles Darren posted and at 60 psi of liquid pressure, flow was reduced from 2.28 gph at 50 psi of air pressure to 1.11 gph at 70 psi.

[JohnA]

Precompressor injection sees no boost because it takes place always under mild vacuum (before the compressors obviously!)
So the pump not flowing enough is not really a problem.
Good atomisation and aiming at the eye of the compressor are more important considerations.

[cheekychimp]

Okay, so that might still work then. What about my post above. The consensus previously seemed to be that this was only really beneficial for a compressor that was being maxed out and was unable to flow any more air. What about the situation I described where the restriction is the size of the exhaust housing. Will making the air denser allow me to push more through a smaller aperture. In theory it seems logical that it would but I'm wondering if the end result will be as effective as on a compressor that is reaching the end of it's efficiency threshold.

The thing is that whilst I understand that one of the main reasons that this works on smaller compressors is because they start to generate so much heat when they move out of their efficiency island, it also seems highly likely that where the exhaust housing is 'too' small, similar problems of heat and friction are going to occur where the big turbo is forcing more air into the housing that can pass the restriction. I'm sure that cooling the charge air and making it denser would have an affect here, I'm just not sure how much.

If anyone can spread some info here I'd be very grateful. I'm intending to use a Holset turbo on my application because of the turbo's durability and it's reputation for spooling quickly. The main criticism that is levelled against the turbo with the smaller exhaust housing however is that it loses between 50-100hp of it's full potential at the upper rpm range as a trade off for the very fast spool up of such a large turbo.

The possibilty of having a 65 lb/min turbo reaching 22 psi by 3500 rpms or sooner and making 600 hp is pretty exciting.

[JohnA]

.. What about the situation I described where the restriction is the size of the exhaust housing. Will making the air denser allow me to push more through a smaller aperture. In theory it seems logical that it would but I'm wondering if the end result will be as effective as on a compressor that is reaching the end of it's efficiency threshold.
We'd need good experimental data for a this.
I've not come across any.
The possibilty of having a 65 lb/min turbo reaching 22 psi by 3500 rpms or sooner and making 600 hp is pretty exciting.
Indeed :smile:

[hotrod]

Based on my experience I would "guess" that the answer is it may help some.
It makes the compressor more effecient (it takes less work to reach a given flow)
That implies that there will be less resistance for the turbine to overcome to reach a given flow so it should spool a bit quicker.

I know in my setup when the pre-compressor injection was working the turbo seemed to hit very hard when it turned on.
Sooooo the seat of the pants dyno says it might work but only some before and after testing would give you solid data to work with.

If I remember the NACA studies correctly at 3% water to air flow by weight the compressor flow increased about 10% over its normal flow, so I would think it would not be unreasonable to think the compressor should come on boost 10% quicker. If you hit full boost at 3500 rpm without it then perhaps you would get to full boost at 3200 with the injection.

Only my gut reaction but it sounds reasonable to me.

Larry

[cheekychimp]

That's very interesting, but that brings two things to mind;

1) My initial reason for doing this was to see if I could gain any of the power back at the top end that this turbo allegedly loses from using a smaller exhaust housing. The basic problem is that whilst the turbo flows 65 lb/min at it's peak, the exhaust housing might flow lets say only 52-55 lb/min. So I'm just not sure if using water injection is going to bring the 55 lb/min threshold on earlier giving me less rpms in my powerband or if the air being denser means the smaller exhaust housing my be able to flow more than 55 lb/min if the air is denser and therefore more condensed/compressed. Surely it stands to reason that 55 lbs of heated air is going to take up more space than 55 lbs of denser, saturated, cooler air?

2) Bringing spool on quicker is something I hadn't considered but if I were to use this method, I had not initially intended having the water injection come on so early. If it was boost activated the activation point would have to be very low, perhaps 5-7 psi or perhaps rpm based. Making it rpm based would however not take into account different loads and airflow in different gears. I wonder if an airflow based activation point run off the MAS/MAF would be worthwhile?

I also intended to run the water in a mixer nozzle with propane injection. I'm not sure if injecting propane at such low rpms is a good idea in which case I'd need to consider a liquid only system or maybe an air based pump.

[hotrod]

For top end mass flow through the hot side your best bet is to try to keep the exhaust gas as hot as possible to bump up the choke flow point as high as you can. That might mean leaning things out a bit, or playing with ignition timing.

It's really hard to speculate without any hard data. I triggered my pre-compressor injection strictly off of manifold pressure.
On my second generation setup I was going to use both an RPM window switch and the MAP to do it so I was sure mixture speed in the inlet tract was maxed out before I started the pre-compressor injection.

If you have a simple means to do it, a mass air flow sensor input would probably be ideal as it would self adjust for rpm and boost and all those other variables.

Larry

[cheekychimp]

Alright, I'm going to need some help here. On the face of it it looks as if I can get a 60 psi compressor running off 12V DC (draw on the electrical system will have to be evaluated) which I can use to do this. What is really doing my head in here is trying to calculate flow. What really messes things up is that in air atomization nozzles (which seem to produce the smallest water droplets) increasing air pressure will reduce the amount of water flow (presumably because of the liquid that the air displaces).

Now given that the tables for these nozzles suggest water pressure of around 50-70 psi and a Shurflo pump will produce around 100-150 psi we should be able to increase liquid flow, but would this then require a higher air pressure to successfully atomize the increased volume of water? Is 60 psi of air enough or should I be looking at more? I have located a 100 psi air pump but it flows less volume of air. I still can't get my head around whether pressure or volume is most critical for atomization.

There are probably too many variables here for anyone to give me a straight answer, but can someone point me in the right direction. Someone said that good atomization rather than flow is the key, but a certain amount of flow will presumably be necessary to pull sufficient heat out to make the compression isothermic (or as close to isothermic as possible).

So lets just look at flow for a moment. I can bench test how much water the system flows in a min, by experimenting with air and water pressures.

I have a 2.3 litre engine and a turbo rated at approximately 65 lbs/min (however that may have been reduced by the fitment of the smaller exhaust housing). How much water do I need to do what I am trying to do and would the volume of water injected need to be different if I set the injection point much lower in order to assist in spool up.

I have a lot of this stuff figured out and can do it, but I really need help working out the flow calculations.

My other concerns are: -

(a) Will injecting too much water cause it to puddle in the intercooler creating problems?
(b) Will it be necessary to inject more water post intercooler for knock suppression?

[hotrod]

If you are injecting pre-compressor at 2%-3% rate compared to air flow you will have about all the WI you need.

Suppose you spray at 2% air flow rate, and you have a 11.5:1 fuel air ratio at red line.

If you are flowing 65 lb/min, that is 29510 grams of air per min.
At 2% flow that means you would be injecting 590 grams / min of water.

If you have an 11.5:1 air fuel ratio than your fuel is 29510/11.5 = 2566 grams / min
Since most gasoline has a density of about .78 then that is 3289 cc/min
10% water to fuel would be 329 cc/min
15% water to fuel would be 493 cc/min


You would be spraying about 18% per min to fuel if you sprayed 2% of air by weight with a max power AFR of 11.5:1.

The beauty of using the air flow is is self corrects as you lean out the fuel.
the higher your AFR (say 12.5:1) the higher your percentage of water/fuel at a fixed water to air ratio.

Since many engines will report mass air flow directly in grams /sec it is trivial to find 2% of your max air flow in grams per min, which directly converts to water flow in cc/min.

If you delay WI turn on until 10 psi Manifold pressure or so the turbo will be well spooled before the water comes on.

Bottom line:
(a) Will injecting too much water cause it to puddle in the intercooler creating problems? --- possibly but that would quickly evaporate as the intercooler cools down as long as it was not a huge amount of over spray.


(b) Will it be necessary to inject more water post intercooler for knock suppression?
Probably not in my opinion.

On the atomizing nozzles I've looked up they most only need to have about 30 psi air supply pressure to work properly.
Not sure why you would want to run such high pressure unless you tapping a pre-existing high pressure supply.
Larry

[cheekychimp]

Larry thanks that really helps!

I guess I just figured higher pressure would atomize the water better. I know if you start getting paranoid about injecting water pre compressor you shouldn't do it. The problem is I have a brand new $1000 turbo about to go on the engine and I really want great atomization to avoid problems.

Had another thought. If you are using a MAF in draw through and injecting pre compressor with an atomizer nozzle, won't the air you inject be unmetered and make you go lean?

[EvoTio]

I completed my pre-compressor injection set up a couple of weeks ago on my 2003 Lancer Evolution VIII. I am still running the stock 16G turbo and am using a .3mm pre-compressor nozzle mounted on a custom axial mount. The .3mm nozzle that I?m using is an unreleased nozzle that Richard supplied me to experiment with. It has an internal progressive spring loaded valve so it opens about 2-3 psi later than my primary post intercooler .7mm nozzle. The .7mm is set to come on at 13 psi, so I figure that the pre-comp nozzle is spraying at about 15-16 psi. The maximum peak psi I see is 22-23 psi and it tapers down to 18 psi by redline. Here?s a link to pictures of my installation if anybody would like to check it out: http://www.aquamist.co.uk/phpBB2/viewtopic.php?t=654

I was able to go to my local Dragstrip yesterday to see what sort of results I would find. Unfortunately, my car fell off in power. I usually run 12.80?s @ 104-105 mph. My first pass was a 13.27 @ 100 mph. I then richened up my air/fuel mixture a little thinking that I may have leaned out with the pre-comp injection. I ran even worse, a 13.48 @ 99 mph. I decided to disconnect the pre-comp nozzle and run again on my regular .7mm post intercooler nozzle along with my previous fuel map. The car responded to a 12.90 @ 104 mph. The car was acting like the engine was receiving too much water mixture. My thoughts were that by using a small .3mm pre-comp injection nozzle, the water mixture would have evaporated by the time it exited the turbo. I was going to reduce the nozzle size post intercooler to see if the car would respond, but unfortunately the track was closed shortly after due to high winds. I?m hoping that maybe some of you that have had more experience with pre-compressor injection could offer some advice.

[Richard L]

This is a very good practical example of the pre-T injection.

Such a small amount of spray has such big effect, a mystery yet to be fully exploited! Pity you cannot complete more test before track is shut.

[hotrod]

You might find it would be better to cut back the size of you post intercooler jet to about a.5mm.
I found that in my setup I was over spraying quite a bit and have kept reducing the size of my pre-compressor jet.
I started with 4 gal/hr (240-250 ml/min) and ended up using a 2 gal/hr (120 ml/min) and I still think I was using a bit too much. I am re-engineering my setup and will be trying a couple other options.
I want to set the pre-compressor injection up so it only comes on at high boost pressure so I am planning on put a boost switch and another solenoid on the pre-compressor jet so it only comes on at boost pressures that I know will be maxing out the turbo's air flow (around 20 psi + ).

Larry

[EvoTio]

Larry, thanks for the response. I'll try a .5mm nozzle since I have one on hand. I also have a .4mm and .6mm also to play with. The only thing that concerns me is that since I was tuned for a .7mm post intercooler nozzle and a .5mm nozzle may run slightly lean for a moment until the pre-comp nozzle kicks in. The engine just had the symptoms of too much water mixture being ingested and reducing the post intercooler nozzle definitely makes sense. The first chance I get, I'll hit the dragstrip again. Luckily the track was kind enough to give the racers a free race pass for having to close down the track early so I can return and tune for free :smile:

Joe

[NAnderson]

I found that in my setup I was over spraying quite a bit and have kept reducing the size of my pre-compressor jet. I started with 4 gal/hr (240-250 ml/min) and ended up using a 2 gal/hr (120 ml/min) and I still think I was using a bit too much.
How did the car react when you were spraying too much pre-compressor through the turbo? Similar to an overly rich condition, where the car stumbles and chugs a bit? Did it hamper spool or boost response at all?

I ask because I've got a 3 gal/min nozzle planned for my pre-compressor setup and don't want to be spraying too much right out of the gate.

[ride5000]

I completed my pre-compressor injection set up a couple of weeks ago on my 2003 Lancer Evolution VIII. I am still running the stock 16G turbo and am using a .3mm pre-compressor nozzle mounted on a custom axial mount. The .3mm nozzle that I?m using is an unreleased nozzle that Richard supplied me to experiment with. It has an internal progressive spring loaded valve so it opens about 2-3 psi later than my primary post intercooler .7mm nozzle. The .7mm is set to come on at 13 psi, so I figure that the pre-comp nozzle is spraying at about 15-16 psi. The maximum peak psi I see is 22-23 psi and it tapers down to 18 psi by redline.

joe, i have a question about your setup...

you say that the pre comp nozzle has an internal spring valve that causes it to open 2-3psi "later" than the other nozzles in the car. i can't see how this spring valve inside the nozzle would be opening with 2-3psi more BOOST--it doesn't "see" the boost at all... it sees the pre compressor inlet pressure on one side, and the water injection pump's output rail on the other.

i CAN see the valve opening up once the RAIL pressure is 2-3psi higher than when the other nozzle(s) open up.

perhaps richard can chime in here and offer some insight as to the construction/use of the prototype nozzle?

the other advantage i can see to an internal 2-3psi check valve would be a decreased tendency to drip after flow is ceased... a "good thing" particularly pre compressor.

in any event, thanks for posting up your experience,
ken

[EvoTio]

I completed my pre-compressor injection set up a couple of weeks ago on my 2003 Lancer Evolution VIII. I am still running the stock 16G turbo and am using a .3mm pre-compressor nozzle mounted on a custom axial mount. The .3mm nozzle that I?m using is an unreleased nozzle that Richard supplied me to experiment with. It has an internal progressive spring loaded valve so it opens about 2-3 psi later than my primary post intercooler .7mm nozzle. The .7mm is set to come on at 13 psi, so I figure that the pre-comp nozzle is spraying at about 15-16 psi. The maximum peak psi I see is 22-23 psi and it tapers down to 18 psi by redline.

joe, i have a question about your setup...

you say that the pre comp nozzle has an internal spring valve that causes it to open 2-3psi "later" than the other nozzles in the car. i can't see how this spring valve inside the nozzle would be opening with 2-3psi more BOOST--it doesn't "see" the boost at all... it sees the pre compressor inlet pressure on one side, and the water injection pump's output rail on the other.

i CAN see the valve opening up once the RAIL pressure is 2-3psi higher than when the other nozzle(s) open up.

perhaps richard can chime in here and offer some insight as to the construction/use of the prototype nozzle?

the other advantage i can see to an internal 2-3psi check valve would be a decreased tendency to drip after flow is ceased... a "good thing" particularly pre compressor.

in any event, thanks for posting up your experience,
ken

Ken, I'll have to have Richard explain more about the nozzle design.

[Richard L]

I completed my pre-compressor injection set up a couple of weeks ago on my 2003 Lancer Evolution VIII. I am still running the stock 16G turbo and am using a .3mm pre-compressor nozzle mounted on a custom axial mount. The .3mm nozzle that I?m using is an unreleased nozzle that Richard supplied me to experiment with. It has an internal progressive spring loaded valve so it opens about 2-3 psi later than my primary post intercooler .7mm nozzle. The .7mm is set to come on at 13 psi, so I figure that the pre-comp nozzle is spraying at about 15-16 psi. The maximum peak psi I see is 22-23 psi and it tapers down to 18 psi by redline.

joe, i have a question about your setup...

you say that the pre comp nozzle has an internal spring valve that causes it to open 2-3psi "later" than the other nozzles in the car. i can't see how this spring valve inside the nozzle would be opening with 2-3psi more BOOST--it doesn't "see" the boost at all... it sees the pre compressor inlet pressure on one side, and the water injection pump's output rail on the other.

i CAN see the valve opening up once the RAIL pressure is 2-3psi higher than when the other nozzle(s) open up.

perhaps richard can chime in here and offer some insight as to the construction/use of the prototype nozzle?

the other advantage i can see to an internal 2-3psi check valve would be a decreased tendency to drip after flow is ceased... a "good thing" particularly pre compressor.

in any event, thanks for posting up your experience,
ken


I believe joe was trying to say the 0.3mm (pre-turbo), spraying at 65cc/min comes on at 15psi due to an inline 15psi checkvalve being installed. If his car is boosting 12psi, the 0.3mm will come on at 15 psi. The 0.3mm nozzle is a standard aquamist nozzle with a 25um filiter inside.

[HarryMann]

Having only just managed 14 pages of posts in my first read, maybe I should apologise for posting a reply/comment prematurely... but won't, as this is such a necessary and compelling thread! And I think I might be able to fill out some aerodynamic theory, as a lot of my thermodynamics has gone the way of the wind and tide :smile:

The step at the compressor inlet is almost certainly there to trip the boundary layer - if it's still laminar (likely), to fully turbulent. Many misconceptions around (not suggesting here, particularly) about laminar/turbulent boundary layers and separated/non-separated flow regimes - they're inter-related but very different things...

I've seen Mach No. (M), mentioned, it's important, but unless I missed it or comes after Page 14, not Reynolds No (Re), this being another non-dimensional parameter that determines the ratio of kinetic to viscous energy in the flow, and has a profound effect on boundary layer and flow separation. And of course the operation of turbo-machinery, as well as heat-exchange...

The effective Reynolds No. (Re) is low - and in turbo-machinery terms, perhaps very low, as the characteristic size of vehicle turbos is small (inlet hose diameter, impeller diameter) until things are changed dramatically as the flow really picks up speed (and is then immediately diffused).

We're talking 'dimpled' golf-ball behaviour here, where the payoff in delayed separation at the rear of a bluff body easily outweighs any extra frictional drag from a thick non-laminar boundary layer.

Suffice to say, that I'm not contributing directly to the debate on pre/post compressor WI, but suggesting that it might be worth establishing what's going on aerodynamically before the really complex bit - entry into the world of high-speed, but very small-scale turbo-machinery.

One possibility, is that a boundary-layer trip a bit further back up the inlet hose, might well help evaporation if injection is taking place well before the impellere eye - a wire, glued circumferentially around the inside of the hose, might do the trick... pulling flow energy down onto the surface (turbulent bl's have a much higher exchange of energy between the surface and their outermost extent).

One other point is a few mentions back there of sharp turns in the hose just before the turbo-compresor inlet... and a few raised eyebrows perhaps? Whilst a reasonably tight 90 degree bend probably isn't too much of a flow impediment, a common solution when a really tight bend is necessary is a banjo style design. An enlarged radius just prior to compressor entry being fed from the inlet hose at right angles to it - if that is a good enough picture in words. Acts a bit like a settling chamber before the flow changes direction

I have such a banjo elbow on my candidate vehicle (due an engine mounting within 4 or 5 inches of the compresor inlet), and thinking about it, might present an ideal area to inject straight into the impeller's eye from just a few inches away...

Enuff for now, must read-on...

[Richard L]

That added another dimension to the myth of pre-T injection.

It appeared people is getting mix results.

[mops]

OMG, I made it through all 24 pages... it took me a week at work - seriously, not a full time week, but 'when i have time week'.
some really good info here, but then again, there's some misrepresentation going on. Mind you might might be slightly off topic.

Particulary I want to comment on turbo sizing.

Everybody here seems to be talking about the compressor, and compressor maps, etc. Turbo is made of two parts compressor and turbine. they are conected by a common shaft therefore turbine is as important in making power as compressor being drive by that turbine. Nobody mentioned manifold design/type...

Exhaust housing size is mentioned in only 2 posts in the whle thread, while it is a very important factor.
Exhaust housing size, along with manifold sizing/design and exhaust system greatly affect power and responsiveness of the engin.
Assume good manifold deskign (tulbural, no sharp turns, equal length, full split pulse separation, with 2 wastegates) and ample exhaust system used.
The bigger the exhaust housing the less restrictive the exhaust gasses path, therefore greater power putput. That is aswell compounded by the fact that less restriction lowers EGT's therefore raises knock treshold.

it is common that oem turbo setups usually use smaller exhaust housing as that produces a quick throttle response making car feel more powerfull than it actually is...

ok, ok, where i'm getting at, with regards to pre turbo WI.
I'm saying that all thought I agree with the fact that preT injection will incerase flow of the compressor, i think the because air is dense, compressor is harder to turn therefore is turbine and that is more resistance to the engine.

And if your turbo is outside of efficency range, wi might incerase air flow, but chances are that your exhaust housing is maxed out in terms of flow therefore more air will not make much more power.

Many people here seem to confuse boost with power. power is a function of much much power is made in cylinders and how much is lost for running rngine (resistance of bearings, inertial losses on pistons which are constantly accelerating up and down, resistance of the exhaust gas flow. Generally speaking in a healty setup should produce nearly 200% of it's NA power at 1 bar of boost (i.e. 2 bat absolute pressure). and it's not exacly 100% per 1 bar, but as you incerase the boost you should see incerase in power. keep incerasing boost further and you will eventually get to the point of diminshing returns, where more boost doe not produce as much more power as at lower levels. so you work out a happy safe spot for you or look what can be changed to make that boost range more efficient... bigger exhaust housing, more efficient compressor, etc...

I guess what I'm saying is that with small oem setup type turbo your tuning abilities are limited because of the size of the turbo. and normally stock manifolds are cast and generally pretty bad design (see twin turbo supra stock manifold).

but with building efficient systems (low exhaust restrictions) great efficiency can be obtained. taking that idea to new level, there are some great benefits when we reac the point when boost pressure is higher than exhaust backpressure... then you can use wild cams on turbo motors, and that allows you to spool large turbos with small engine with great efficiency.... a example of bad way to do it is the 800hp skyline mentioned earlier in this thread which at 3 bar makes less power than at 2 bar. too small compressors ? possily. exhaust restrictions must be huge. pumping more air will not solve the problem. get exhaust housings 2-4 sizes larger and it will make 900 on the same boost... they will spool a bit later tho...

I think water (and other pre turbo injection) has it's place and should be further researched. But results may vary in every case. It seems like ideal candidate would be turbos with smallish compressor but large turbine... where compressor is limiting factor of the system... cooling charge air by convection (evaporatino of water or other fluid) is a way to go i think, but needs more experimentation...

I was thinking... maybe along with spraying large quantities of methanol, inject pure oxygen to the intake tract... yes, it might be expensive but power gains might be great too.... would work good on drag cars...

[masterp2]

Just getting familiar with this topic again, after being away for a year.

I am seeing some questions and observations that don't have answers, and I thought I would get involved. I have spent much of the last year working through a problem with coolant overheating on GM diesel work vehicles. Since some of my findings are applicable to this effort, I will elaborate.

What I found was that the smallish compressor was working in the northeast part of the map, and in some thermal conditions, "off the page". On further investigation, the temperature of the compressure discharge emerged at near 600 F. I calculated efficiency around the 50% mark. Needless to say, this is in a word, ridiculous! Further investigation led me to question what the effect of this heatup (and expansion) is on the downstream plumbing. Something told me that I was seeing extaordinary head losses with this huge heatup. In other words, the velocity in the IC/CAC plumbing would be increasing dramatically, leading to higher frictional losses. It turned out, the difference between the compressor discharge pressure (work) and the intake plenum pressure (downstream) was nearly 6 psi! This at a plenum requirement of 32 psi. On the diesel this is around 60-65 lb/min of airflow.

Now hopefully I have not lost you. The plumbing restrictions when considering the 2.5" IC plumbing and the IC itself, totalled 6 psi. (2.5" is way too small)

After working through the compressible flow equations in a 2.5" conduit, it turns out the increased discharge temp was creating much of this loss, via increased air flow velocity. With fluid flow, there is apoint where the force required to push the fluid (air) through the straw (or IC pipe) becomes exponential, the curve quickly rising vertically at some flow rate. When this happens it is time for a diameter increase from the engineers. (what in fact happened, is that GM reduced the diameter from 3.0" in previous models, and to this day there is no known explanation why)

Adding insult to injury, that high temp product was leading to dramatic ambient temp increases behind the grill. The IC sits in front of the radiator, and measured ambient in front of the radiator, on the hot side of the IC, was 240 degrees! So the CAC was acting like a torch to heat coolant. I calculated something like 290,000 BTU/hr of heat exchange with the IC. Nuts!

But back to hot discharge product. The predictions for this non-adiabatic behavior, show that velocity increases dramatically, and adds a lot of added restriction in the plumbing. Clearly cooling the charge increases density, reduces velocity, and hence pressure losses. This means that, for a given desired intake plenum boost pressure, less work (discharge pressure) is required. This compressor now works less, which means higher efficiency. The improvement is cyclical in nature.

it appears that pre-compressor WI, PCWI, can be a performer or a deterent to performance. If you already operate in the efficient islands, on a humid day, then cooling charge can move you to lower efficiency on the left. It also leads to huge condensation effect in the IC. So PCWI would have limited usefulness on a properly designed forced induction platform with large stock amounts of charge cooling. But the undersized compressor, operating on a dry day, with excessive air box temps, should benefit big.

From my point of view, I do not share the idealized concept that WI provides quasi-isothermal compression. I believe that all the inefficiencies of non-adiabatic compression are in place even with WI. But naturally the beneficial impact of cooler charge can be seen in my explanation. But I don't believe that WI improves the efficiency of the compressor, from a purist sense. This assumes that there is no appreciable evaporation prior to compressor, as is the case in an axial mount nozzle in front of the nut.

[masterp2]

Richard, I have tried PM'ing you and email. What am I doing wrong?

[masterp2]

What happened? Did everyone die? Or is it just cold everywhere but here?

[HarryMann]

I think theys did :D

Still trying pre-compressor injection... but nothing conclusive yet. BY lowering the onset pressure from 10 to 8psi on my 1.9TD diesel, I was fairly sure that a small effect was noticeable (0.3mm jet, not a lot of water) but then other things conspired against my tests

Have bought some temperature instrumentation to see if I can detect a reduction in charge temp after the compressor, or rather, how significant it is

[masterp2]

Harry, I am interested in your results. I have done some post compressor temp and pressure monitoring, but not with WI yet. I did it in conjunction with evaluating larger boost tubes. Used a nifty 4-channel temp datalogger.

If you have a VGT, and if it is like others I have seen, the WI should lower discharge pressure, while maintaining equal plenum pressure. Curious if you find this to be true. The effect should be magnified in systems with a lot of pressure drop, high IAT, and high boost/low efficiency/off map operation.

Anyway, all this means that for a given boost, is less wok that turbo must perform, and lower final charge temps, better charge density, hence better economy. For those that tow, and have a thermo-viscous fan, the fan can be kept off with lower CAC heat rejection, and this can be a huge benefit.

That's the theory anyway.

[HarryMann]

VGT?

The effect should be magnified in systems with a lot of pressure drop, high IAT, and high boost/low efficiency/off map operation.

That's me running a small A/R turbo that starts spinning early but soon reaches the turbine limit and probably chokes the compressor too.. hence high EGTs, and haven't been able to see any definite EGt eduction yet, but no doubt there must be some...

Going to look at my compressor vanes today, pretty well requires removing turbo... alos have had turbine/CHRA gas leaks (kKK's are bad for sealing that joint)

[JohnA]

Where did the EGT reductions come into play?
I've yet to find a documented case of EGT reduction anywhere in the (sparse) literature.

EGTs not increasing at even higher boost -- yes
EGTs decreasing ---> haven't seen that yet.

[HarryMann]

Well, true, I had been told not to expect EGT reductions of signifance, however, on a digital EGT a friend with same engine is seeing at least 30C reduction with post turbo Water Injection.

Mine are so high, that I thought I might see something of benefit in that area...

Non-intercooled 1.9 TD diesel, 12~15 psi boost

[JohnA]

Well, true, I had been told not to expect EGT reductions of signifance, however, on a digital EGT a friend with same engine is seeing at least 30C reduction with post turbo Water Injection....
Yes :smile: , small clarification here: it's not that you will NOT see any EGT drop, such a sweeping statement for WI would be *really* wierd :lol:
What I'm saying is that you won't necessarily see EGT drop in each and every case (although conventional wisdom would indicate otherwise)
I remember reading in the old lab tests where they were pouring LOTS of water (compared to fuel) that EGTs were not affected that much. Detonation is eliminated ofcourse.

[NAnderson]

Anybody else have input or experiences with their pre-compressor systems?

It's getting towards winter time here in the frozen north and I need ideas for next years setup. :cool:

[keithmac]

I`m going to give it a go, got some material ordered to knock up a pair of jet holders, similar to Richards install on the Skyline.

My car runs out of boost at 6000+ rpm (only holds 12psi to redline due to small 9b turbos), so first measure of any sucess for me will be higher boost. I`ve got some 1/4 mile info to compare results also, so the next test will be the trap speed with pre-turbo injection.

I`ll try and get some good pictures of the compressor wheels before fitting the injection, to give a starting point for the condition before and after.

How my car/ system is set up I`ll have to be using a 0.9-1.0 nozzle on each turbo to keep my tune close, can always re-jig if it overcomes the turbos, if it runs well on 1.0 jetting I can directly compair the results to mylast 1/4 mile info as I`ll be injecting the same amount of water/methanol.

[JohnA]

Keith,

Excellent, please take as many pictures and measurements as possible :cool:

[NAnderson]

Definitely keep us posted on your results. The more info the better as pre-compressor injection works great "on paper" but seems to need some tweaking and R&D in practice.

How my car/ system is set up I`ll have to be using a 0.9-1.0 nozzle on each turbo to keep my tune close...
You're planning on running a .9-1.0 nozzle on each 9B Mitsu turbo?! (6G72? Since you mentioned the 9B's) That nozzle seems awfully large for that small of a turbo and boost level. A few pages back hotrod runs through his experiences with injecting too much water pre-compressor. May want to take a browse back there.

[Cummins Express]

so I've kept up on this thread for a few years, and now have a variant for you all. masterp2 shares some interest with me from another site, as well as being an owner, like me, of a diesel. I run a compound compressed 5.9l cummins. Low pressure turbo is a BW BHTB3 currently peaking at 37 psi absolute. High pressure turbo is BW S3A with a self made variable turbine. post high pressure turbo boost peaks at 82 psi absolute. An OEM aftercooler is employed post secondary, and no current compression stage intercooling is present, chemical or otherwise.

I have experimented with precomp WI in the past, and have recently been motivated to continue the R&D, but have some questions requiring your input. Knowing the harsh operating conditions the high pressure turbo operates under such as blistering inlet temps, having to ingest and compress a second stage of charge hovering in the 90 lb/min range, there is definite room for improvement in the high pressure turbo efficiency. Because of of the heat present in the inlet in excess of 350*F, my goal is to drop this temperature as drastically as possible. For lack of room for a second charge air cooler (CAC), in addition to the reluctance of adding that much more intake tract in an already dizzying amount of intake tract volume, precomp WI is my choice. Now for the nuts and bolts. I had previously experimented with WI at the inlet of the secondary charger with varying results and crude fabricating. I now wonder, after reading all 25 pages over again, if my purposes wouldn't be better suited by injecting pre-primary turbo? Two things come to mind...firstly, injecting at ambient pressure, with the minimal depression at the primary's inlet, vs injecting inter-stage, where the environment pressure is 37 psi and 350*F. Which would be more beneficial I wonder...injecting in a hot, pressurized environment, or the adverse, pre-primary turbo?

Additionally, I am talking about large quantities of water/meth, and am going outside convention regarding injection pressures and nozzle types. In my first attempt, I utilized furnace nozzles. High flow rates, and a hollow, spiral, 45* cone, placed an inch in front of the secondary's compressor wheel. I raised the nozzle's spec operating pressure from 100 to 200, hoping for more flow, and a higher atomization rate. I plan on going the same route, but with a different pump. I plan on belt driving a 3 stage high pressure pump on a clutch in excess of 1000 psi, or as much as 3000 psi. These pressures, of course, will be dictated by what the nozzles can handle, and some pretty crucial testing will take place on the bench before any mounting will occur. My initial thought is if I can find the nozzle will hold 1000+ psi, atomization rates should be quite high. The other option , of course, is a lower pressure system with narrow angle hollow nozzles that will provide high spray quality. I wouldn't mind going narrower than 45*, as this would allow me a little more space between the wheel and nozzle. All my findings in the past showed me that a hollow cone that avoids the wheel nut, yet is narrow enough and close enough to enter the wheel in its first 1/4' inch of radius showed no signs of erosion of impingment in the 10,000 or so miles i had it operating....at maybe a 30% to 50% duty cycle (I have a lead foot and see high boost frequently hehe).

All I need to do now is affect an injection system that will yield an efficient nozzle arrangement, and maintain that efficiency at say 20 gal/hr!?!? Is it feasable?

Chris

[accre]

Okay, after looking at all of the fantastic responses on here I am convinced of the effectiveness of pre-compressor injection as a way to maintain the spool up of a small turbo while getting good top end performance. That being said, I am concerned about impingement.

I've just spent the last couple of hours on the Spraying Systems Co. website, and am very interested in their flomax air atomizing nozzles. Autospeed has an article about water injetion using one of this companies air atomizing nozzles using boost pressure as the air source. The Flomax nozzles work on pressures as low as 5psi, which I'm assuming is a pressure differential, so it means less than 5lbs of manifold pressure since the turbo is creating a vacuum on the intake side. They also are available in pretty small diameter spray patterns, and they make a 90 degree lance. I'm thinking that one could place the lance not far from the compressor, with the nozzle directed right at the center of the wheel. The fine spray (around 30 micron max from what I gather) combined with a targeded spray at the center of the wheel should make impingement negligable. They also produce nozzles that can be pulse width modulated like an injector.

how would you control such a system? I was thinking a small pre-compressor nozzle at high boost and rpm, and a larger nozzle (not air atomized) just after the intercooler that would be on more than the pre-comp nozzle.

[keithmac]

I`ve had a good look at my setup and think if I go directly before the turbos the jet holders will cause too much of a restriction to the turbo, plus if I need to service the jets it`s a big job to get to the rear one.

My other option is to inject after the MAS meter (draw through type) with 2 jets on a T to feed both intake pipes to the turbos, the pipes are aprox 2 foot long each. It won`t be as good as direct spray at the compressor centre but 0.3mm jets will hopefully atomise well before they hit the compressor. This setup would be simiar to the wetted airfilters that`a been talked about already on here.

May have to increase the jet sizes to get a noticable effect but that would just mean pulling out the MAS which is a 2 minute job.

Will keep you all updated as I go.

[cheekychimp]

accre:

I too looked at the air atomizing nozzles. If they utilize boost pressure air that has already been metered then they look like a good bet. I was reluctant to try them in a MAS metered car because any air injected post MAS is unmetered and could damage the tune significantly. If you are running SD you can of course ignore this.

keithmac:

My only concern ith what you are suggesting is that whilst injecting further from the turbo allows time for atomization, there is a very high risk that those tiny droplets will het the sides of the intake where surface tension will lead to many attaching themselves to each other forming larger droplets that will damage an impeller. Safety really does rely on a sufficiently well atomized spray injected staright into the impeller.

I also seriously like the idea of the hollow cone spray avoiding the nut on the impeller shaft altogether and going straight into the blades. That sir is genius!

[masterp2]

Where did the EGT reductions come into play?
I've yet to find a documented case of EGT reduction anywhere in the (sparse) literature.

EGTs not increasing at even higher boost -- yes
EGTs decreasing ---> haven't seen that yet.

http://members.cox.net/td-eoc/INDUCTION-THERMAL%20FEEDBACK%20PRIMER.doc

[JohnA]

Looks like an interesting document :smile:

However it mainly tackles airflow restrictions upstream of the compressor - not water injection.

My reference to EGTs was in relation to WI :smile:

[nutron]

A few small points,

1) by adding the water/methanol you displace some air and so don't get as much of an increase in efficiency as you might through cooling alone during compression.

2) water and oil get drawn into the turbo all the time, methanol does not. The seals in a turbo are not designed to interact with methanol and over time (I know someone said they had been doing this for a year) the seals could degrade and the turbo might fail at a fraction of it's intended life.

Erosion of the blades is not a major concern as I have said previously, as the breather for the engine feeds back into the intake before the turbo; any oil mist will be going into the turbo in the same way water would and so turbos are designed to withstand this type of particulate flow.

A lot of turbos are water cooled to deal with the high exhaust temps of petrol engines, you could seperate the cooling system from the engine cooling and cool the turbo more effectively. This would allow you to decrease charge temps by not imparting so much heat in the first place.

More small nozzles are better than one large one, as they cause less turbulance.

[masterp2]

John, there is a section in the article that addresses turbo shaft HP and drive pressure. There is also an energy balance that shows that reducing compressor heat (either be reducing restriction flaws, OR by misting) would reduce EGT. How much, remains to be determined.

[JohnA]

John, there is a section in the article that addresses turbo shaft HP and drive pressure. There is also an energy balance that shows that reducing compressor heat (either be reducing restriction flaws, OR by misting) would reduce EGT. How much, remains to be determined.
Yeah, well the pressure ratio is not exactly a new concept, is it.
One could argue that misting actally increases airflow restriction because it changes the viscocity of the fluid :? ...
Someone else could argue that there is a decrease in restriction because temperatures are not allowed to to grow (with precompressor injection)
Maybe there are more factors affecting airflow and who knows what the overall result is (may it's variable too!)

We really need proper scientific experimentation don't we
Controlled environment, repeatable results, the lot...

[masterp2]

Actually, I have struggled with the reality, that evaporating water, while it absorbs energy, does not necessarily relieve the exhaust of the job of supplying it. It is a thermo dilemna that I have argued back and forth.

[Richard L]

Here are the pictures of larry's turbo. The picture has restored on the 13th August 2009


http://www.aquamist.co.uk/forum/compressor9eb-s.jpg

[masterp2]

Has that compressor been isolated to water damage only? IOW, was filter in place entire time? Nice detailed pic, but it does not have the characteristic water damage. That looks like sand.

[Richard L]

The turbo picture was takem by Larry, Hotrod. He explained the damage is due to many other things as well, read the fist few pages of this thread. I hav e also heard report from PuntoRex but his image link is gone. I have emails him and hope he may find them so that I can re-post them here.

[masterp2]

perhaps this pic can be layed along side an earlier image from, say a year ago, to determine wear progress.

[Richard L]

The picture I posted yesterday is the same as the picture psted a few years ago. I have some server problem and lost some of the pictures in the original post. I didn't realise until I looked at the link a few days ago.

[snoroost]

I read a fair amount of this thread but I have a question is injecting pre turbo very good at stoping the deto???

[Richard L]

Not sure on the effect on det.

People are more interested on extend the useful range of a small turbo charger.

[RICE RACING]

You really need a fine atomized spray (~5 microns) to do the job 100% without damage to the wheel and even better if your intake system can be configured to shoot this into or near the center of the compressor wheel. Here is what I have fitted to my own car (my own WI system) but I have seen others use other vendors product with much less atomization and if its targeted to the right spot no wheel damage will result ever.


http://img216.imageshack.us/img216/6252/mounted1p.jpg

http://img36.imageshack.us/img36/8656/mounted2.jpg

http://img269.imageshack.us/img269/9718/mounted3.jpg

http://img504.imageshack.us/img504/4939/mounted4.jpg

[masterp2]

Another good use for pre-compressor injection, is the sustained high load, trailer towing. The CAC for these vehicles is in front of the radiator. The heat of compression is so overwhelming, T2 can easily be over 500 F (250 C) at higher altitudes. This heat load rejected in front of the radiator is enormous, calculated at times, over 300,000 BTU/HR. In short, the radiator can no longer function well enough to serve the cooling needs. It is possible to absorb 150,000 BTU/HR in latent evaporation across the compressor, and this is a large burden removed from the CAC, and hence radiator can perform adequately and keep coolant under the BP.

[masterp2]

Here is a one application. An 8 nozzle unit. I did this article to show how diesel application differs from a gas application


http://www.maxxtorque.com/summer-2009

[Richard L]

You really need a fine atomized spray (~5 microns) to do the job 100% without damage to the wheel and even better if your intake system can be configured to shoot this into or near the center of the compressor wheel. Here is what I have fitted to my own car (my own WI system) but I have seen others use other vendors product with much less atomization and if its targeted to the right spot no wheel damage will result ever.



I kept trying to tell people to use air assisted nozzle or direct the spray towards the centre of the compressor wheel. This simple message never seen to get through.

[Richard L]

Here is a one application. An 8 nozzle unit. I did this article to show how diesel application differs from a gas application


http://www.maxxtorque.com/summer-2009

Excellent article. The advantage of pre-turbo injection idea will eventually come throught, but not soon enough.

[Dust]

Rice, interested in more pics of your airwater nozzle setup

[RICE RACING]

Rice, interested in more pics of your airwater nozzle setup

Here is the completed RR air box adapter with RR WI nozzle mounted in it. < Before I made the retainer to hold the air/water atomizer in place. To these I fit my bespoke flow control valves and do the machining to the delivery side of the jet so they have the flow rate I require and also a bit better atomization. My one jet works is variable from 100cc/min to 850cc/min @ 1.5bar pressure.

http://img199.imageshack.us/img199/3213/ricebox1.jpg

http://img35.imageshack.us/img35/6223/ricebox2.jpg

http://img196.imageshack.us/img196/8729/ricebox3.jpg

http://img20.imageshack.us/img20/1792/ricebox4.jpg

[RICE RACING]

The key is a large area to discharge into, and also to have a bit of pre swirl of inlet air, this combined with low particle size for your droplets and you will have excellent performance (at very high water flow rates as I and my customers use) and non visible long term effects on the compressor wheel.

You can see the RX2 in this video http://www.riceracing.com.au/Videos/500_bHp_rx2_water-injection.wmv and you will see one of my jets located on the back of the compressor cover pointing into the air filter void and against the oncoming air stream. This 13B engined car makes 600+rwhp (over 750bhp) on Optimax 98 Ron pump petrol and uses around 600cc/minute water flow rate. It has consumed well over 200lt of water and the compressor wheel looks just like a new one :)

[rodknox302]

Here are the pictures of larry's turbo. The picture has restored on the 13th August 2009


http://www.aquamist.co.uk/forum/compressor9eb-s.jpg

I can tell you from my experience that this compressor didn't have any water methanol injection on it. If it had it would have been squeaky clean. The water methanol would have washed the impeller clean. This unit was ran with out an air filter in a dirty environment which caused the leading edges to wear. Anybody who knows water methanol injection can tell you there impeller is super clean.

I've ran 10 and 14 gph nozzles before my turbo with no problems. These are huge nozzles!

Rodney

[mrx79]

Hallo,

i've tried to read the last 27 pages more or less but i'am curious about if there was something like a conclusion about this discussion?

open questions would be:
- is there a difference between pre-compressor and post-intercooler injection?
- is there a difference (i guess it is) in in-cylinder cooling effects on the injection position?
- would it be maybe a benefit to use both types of injections to improve turbo efficiency and have a in-cylinder-cooling effect?

[RICE RACING]

Hallo,

i've tried to read the last 27 pages more or less but i'am curious about if there was something like a conclusion about this discussion?

open questions would be:
- is there a difference between pre-compressor and post-intercooler injection?
- is there a difference (i guess it is) in in-cylinder cooling effects on the injection position?
- would it be maybe a benefit to use both types of injections to improve turbo efficiency and have a in-cylinder-cooling effect?

I am no rocket scientist (only mech eng qualified)

But in cylinder effect (end gas pressure and temperature) will not vary a lot regardless of where you are injecting the fluid for the simple reason that its the conversion from liquid to gas that had the highest heat conversion rate.
In my tests over the years I have always found a net vehicle increase in speed (over numerous test conditions) with pre turbo V's any post turbo inj location. *all of the other locations don't yield a performance increase when other variables are held constant*
and lastly I am sure if you have a very advanced system like made here you can use multiple points of injection and get the best theoretical performance from this.

This is but one response and view though ;)

[Richard L]

I have often been asked about compressors wheel wear, is this is myth or probable? You are the one of the long term users of pre-turbo users. However I like to point out that you use a more superior technology by using air-assisted nozzle with super-fine droplets.

Please post some of your experiences on this myth.

[RICE RACING]

I have often been asked about compressors wheel wear, is this is myth or probable? You are the one of the long term users of pre-turbo users. However I like to point out that you use a more superior technology by using air-assisted nozzle with super-fine droplets.

Please post some of your experiences on this myth.

I can honestly say this (some will be surprised), in some applications wheel wear will happen and in others it will be basically no existent. In about 2 out of 10 cases I will get feedback of wheel wear the others say it does not happen or is so small you cant notice it after injection of 100lt or more.

I had one customer/friend who bought a USB bore scope camera and took a video to show the water going into the front of the turbo (in his case he had noticeable wheel wear with my system), it seems to be related on where exactly the unit is placed and also if the compressor housing has a bell mouth like found on GT Garrett turbo's (GT35R and TO4Z for example). I have many other people (in stated ratio) who say its just not so and you cant see any signs of wear at all on the leading edge of the compressor wheel.

My personal recommendation is regardless of system (yours mine someone else) to place the water jet against the flow of air if possible or if its following the path of air or mounted on the out pipe wall to injected it a long way down stream of the compressor wheel and not to mount it on the inner radius of a pipe bend. It is more complex than what it seems and it looks like it can be done the wrong way and mark the turbo compressor.

On the weekend I was at the local drag track and there is a friend of mine who runs a very old school turbo set up with suck through tripple SU carbs (in front of turbo), T04, 36psi boost, methanol fueled, and water injection, the water "jet" is nothing but a nozzel and a solid stream goes in front the middle carb! no compressor wheel wear has ever been seen on this set up *go figure*

[RICE RACING]

My personal feeling:

I am not concerned with a small amount of compressor wheel wear given the many benefits of running this technology. Many years ago one person I know changed his compressor wheel over from one that looked like the above picture to a brand new one and there was no measurable difference in performance of the car on the dyno or the track...... he wasted $1300 on the rebuild, but I can understand why he did it.

After I go though at least 50lt of injection fluid I will post up my own comp wheel picture irrespective of what it may look like :lol:

[Richard L]

I don't think your system wouldn't do any damages to the wheel, the droplet is so amazingly fine. I do recommend your system to many people. I also don't think a pitted wheel makes too much difference to the flow, compared to a new wheel. The cooling effect far out weighs the cost of the wheel.

[forcefed86]

My personal feeling:

I am not concerned with a small amount of compressor wheel wear given the many benefits of running this technology. Many years ago one person I know changed his compressor wheel over from one that looked like the above picture to a brand new one and there was no measurable difference in performance of the car on the dyno or the track...... he wasted $1300 on the rebuild, but I can understand why he did it.

After I go though at least 50lt of injection fluid I will post up my own comp wheel picture irrespective of what it may look like :lol:

Been following your results for some time now. How much pressure (water and air) do you use to atomize the water so well?

So all the talk of injecting water straight into the nut of the turbo is not the ideal setup in your opinion? Wouldn't spraying the water towards the filter cause saturatuion of the air filter and excessive puddling on the inside of the pipe?


all video and testing provided by 99gst_racer on dsmtuners.com

Like so...

http://i68.photobucket.com/albums/i17/paulvolk/IMG00104-20100221-1309.jpg

http://i68.photobucket.com/albums/i17/paulvolk/IMG00109-20100221-1717.jpg


Here is a video of the setup in action.

http://www.youtube.com/watch?v=F51xymg33-w


And here's a video of one of the pulls:

YouTube - Engine Dyno Pull @ Street & Strip (http://www.youtube.com/watch?v=lYTMEwPfU2E)

TEST RESULTS

Set-up specifics: The engine on the dyno was a relatively stock 2.3L Ford engine out of a late 80's Thunderbird. The turbo was a Holset HX-35 running at 22-23 psi. The engine was producing roughly 450HP and 389 ft/lb torque. The nozzles were all mounted in the same location throughout the tests - about 4" away from the compressor nut. The pump used was the Devil's Own 250 psi pump.

Here are the results that we gathered:

No injection;
Before IC temp increase - 29 deg/sec
After IC temp increase - 4 deg/sec
Peak boost 22.09 PSI
------------------------------------------------------
1GPH 50/50 mix;
Before IC temp increase - 19.1 deg/sec
After IC temp increase - 3.0 deg/sec
Peak boost 22.88 PSI

1GPH H20;
Before ic temp increase - 24.2 deg/sec
After IC temp increase - 3.4 deg/sec
Peak boost 23.21 PSI

1GPH Washer Solvent;
Before IC temp increase - 24 deg/sec
After IC temp increase - 3 deg/sec
Peak boost 22.96 PSI

---------------------------------------------------
7GPH 50/50 mix;
Before IC temp increase - 16.9 deg/sec
After IC temp increase - 4.5 deg/sec
Peak boost 22.96 PSI

7GPH Water;
Before IC temp increase - 13 deg/sec
After IC temp increase - 4.7 deg/sec <---no IC fan - OOPS!
Peak boost 22.92 PSI

7GPH Methanol;
Before IC temp increase - 19.3 deg/sec
After IC temp increase - 3.5 deg/sec
Peak boost 23.0 PSI

7GPH Washer Solvent;
Before IC temp increase - 11 deg/sec
After IC temp increase - 3.8 deg/sec
Peak boost 23.3 PSI
-----------------------------------------------------------
10GPH Methanol;
Before IC temp increase - 14.3 deg/sec
After IC temp increase - 3.2 deg/sec
Peak boost 23.05 PSI

10GPH Washer Solvent
Before IC temp increase - 11.2 deg/sec
After IC temp increase - 3.4 deg/sec
Peak boost 22.71 PSI

[RICE RACING]

ff86, I have video with pressure transducers mounted in air and water feed through telemetry showing target droplet size happening at 7psi, I generally tell customers to switch it on at about 10psi+ and some drag C16 users operating at 36psi boost pressure tend to turn on the system at 20 to 22psi level.

Got a few users who inject into the end cap of the air filter and they don't report water loading of the element, probably to do with air flow rates, seems to be not an issue, but I am sure if you did just one pull and shut off everything off from an open throttle then you would see some evidence of water there.

So far as ideal spray position? my feeling is to inject into a chamber, this is what I tried to create on my own car, this so far seems to work best.

[forcefed86]

ff86, I have video with pressure transducers mounted in air and water feed through telemetry showing target droplet size happening at 7psi, I generally tell customers to switch it on at about 10psi+ and some drag C16 users operating at 36psi boost pressure tend to turn on the system at 20 to 22psi level.

Got a few users who inject into the end cap of the air filter and they don't report water loading of the element, probably to do with air flow rates, seems to be not an issue, but I am sure if you did just one pull and shut off everything off from an open throttle then you would see some evidence of water there.

So far as ideal spray position? my feeling is to inject into a chamber, this is what I tried to create on my own car, this so far seems to work best.

I'd love to check out that video. Can U email it to me? So your saying the air side of your atomizing nozzle is only seeing 7psi? How much water pressure are you using?

Thanks!

[RICE RACING]

ff86, I have video with pressure transducers mounted in air and water feed through telemetry showing target droplet size happening at 7psi, I generally tell customers to switch it on at about 10psi+ and some drag C16 users operating at 36psi boost pressure tend to turn on the system at 20 to 22psi level.

Got a few users who inject into the end cap of the air filter and they don't report water loading of the element, probably to do with air flow rates, seems to be not an issue, but I am sure if you did just one pull and shut off everything off from an open throttle then you would see some evidence of water there.

So far as ideal spray position? my feeling is to inject into a chamber, this is what I tried to create on my own car, this so far seems to work best.

I'd love to check out that video. Can U email it to me? So your saying the air side of your atomizing nozzle is only seeing 7psi? How much water pressure are you using?

Thanks!

If you look under the Gallery section I just did a post in it with like to video and pressures at the end of my car thread (both around 7psi, you can see them on the lap top).

[stevieturbo]

What a read !!!! All based on turbos though.

What about a centrifugal supercharger being spun hard ?

I currently spin my Vortech V7 YSi to a little over 65k, which is above their reccomended limit ( assuming I see no belt slip )

Oddly the compressor map they post for this unit...they only go to 50k.

http://www.vortechsuperchargers.com/maps/YSi-Trim_Compressor_Map.jpg

Currently makes around 1 bar at 5000rpm, 1.2 bar at 6000rpm, and 1.4bar at 7000rpm.

Although generally shift at around 6700.


I do run a 50/50 after the IC, but could easily fit a nozzle before the Compressor too.

Given the overspeeding and apparent innefficient use of the compressor...would injecting pre compressor be of much benefit ? If it makes life easier for the compressor...would that mean easier on the belt drive ?

I have a pretty large IC, and temps do remain within say 30degC of ambient at most times now. Where I live is generally pretty cool too. Rarely over 20degC

Ive never measured pre IC temps though. Must try and hook something up to do that.

[gluis]

Wow, I was wondering myself what would be the effects of pre turbo WMI. It is more interesting than I expected!

On the other hand, would a system like that benefit from having pre and post WMI? is that possible with our current system using similar pulses but different injector sizes? How can we calculate how much WM mixture needs to be injected before the turbine? Using MAF measurements perhaps? Why it would be better not to have an intercooler? (I don't want to make custom plumbing besides the WMI itself, that's why I'm mounting an AMR TMIC instead of a FMIC). So many questions arise from this post...

Thanks for such an interesting topic,

Guillermo

[Rotard]

I just got the water injection finished up on my RX7 last week, the nozzle is an external mixing nozzle utilizing *air* and *water* for atomization, it is a pumpless system as well. It's rated 600cc @ 20psi. My main purpose for it is a knock deterrent as I'm not really focused on extending my turbos map. Here are a few pictures, I will post any helpful information I come across.

http://i9.photobucket.com/albums/a63/HybriDsm/Seven/DSC01429.jpg

http://i9.photobucket.com/albums/a63/HybriDsm/Seven/DSC01425.jpg

http://i9.photobucket.com/albums/a63/HybriDsm/Seven/DSC01426.jpg

[JoseCabra]

Hi,

after reading a lot about pre turbo water injection I decided to give it a go.
I have a 08 STi with a HFS-3 system from aquamist.
I did some CFD analysis on the intake and decided to put the nozzle about 10 inches before the turbo after a bend on the inside. I found this was the area that provided high turbulence and low pressure so the injection would atomize effectively and quickly before the turbo.
I used a 0.3 nozzle pre turbo and a 0.5 nozzle at the intercooler pre throttle body.

I have to say response at around 4500 rpms is excellent. The car just goes; although I have noticed a little compressor surge but I still have to do some datalogs to verify this.

Before using a single nozzle I used to have bogging down even after tuning ignition to compensate for the water injection.
Now even without tuning the car feels more linear.

I will be inspecting turbo blades later on for some damage but don't think there will be any.

In my opinion injecting pre turbo is more about getting the air to absorb the mixture quickly and not about doing it into the eye of the turbo.

[RICE RACING]

WARNING!

Be extremely careful about trusting general figures for flow rates on cheap copies of my mechanical WI system.

On a full price proper set up like I supply each and every single atomizer is hand modified and tested then stamped to give the proven flow rating capacity (not an "about" setting). Those atomizers as pictured vary wildly as they come from the supplier in both the quality of the air cap and the water side and each and every component needs to be modified & "correctly specified/modified" NOT JUST COPIED FROM ME and hope for the best!.

The precision metering in flow control valves again wildly copied by some are nothing like the quality of what I use and supply in RRWEP110 atomizers. Each and every single atomizer has its own flow sheet corresponding the the flow rate for a turn setting of the precision adjuster. Each and every jet achieves the atomization specification I state... there is a reason I charge what I do just for the atomizer as what some sell a whole WI kit! cheapo Rice Racing copy for.

Your mounting method for the atomizer is not recommended practice either, it is cheap and nasty to have the locking ring inside the air stream, its far harder to mount it properly but I suggest you go to the effort.

****To get the best out of this system you need an adjustable atomizer along with a flow rating that is legitimate and can be verified by anyone testing the unit**** Otherwise you are just guessing at the end of the day and you will not end up with the ultimate results WI can and does deliver.

The widely imitated system! original and the best below!


http://img216.imageshack.us/img216/6252/mounted1p.jpg

http://img36.imageshack.us/img36/8656/mounted2.jpg

http://img269.imageshack.us/img269/9718/mounted3.jpg

http://img504.imageshack.us/img504/4939/mounted4.jpg


Tank

http://img269.imageshack.us/img269/2248/witank1.jpg

This is the thread you need to read about what goes into tuning WI and the recorded results as well, all measured to high precision and no guess work. http://www.aquamist.co.uk/vbulletin/showthread.php?t=1590
You WILL NEED a fully adjustable atomizer to get best results and you will NEED to know how much you are injecting as well. otherwise you will get disappointing results and be chasing your tail.

When done right this stuff is magic :)

[Howerton Engineering]

Hi,

after reading a lot about pre turbo water injection I decided to give it a go.
I have a 08 STi with a HFS-3 system from aquamist.
I did some CFD analysis on the intake and decided to put the nozzle about 10 inches before the turbo after a bend on the inside. I found this was the area that provided high turbulence and low pressure so the injection would atomize effectively and quickly before the turbo.
I used a 0.3 nozzle pre turbo and a 0.5 nozzle at the intercooler pre throttle body.

I have to say response at around 4500 rpms is excellent. The car just goes; although I have noticed a little compressor surge but I still have to do some datalogs to verify this.

Before using a single nozzle I used to have bogging down even after tuning ignition to compensate for the water injection.
Now even without tuning the car feels more linear.

I will be inspecting turbo blades later on for some damage but don't think there will be any.

In my opinion injecting pre turbo is more about getting the air to absorb the mixture quickly and not about doing it into the eye of the turbo.

Great for giving it a go. Keep us updated to any data logging or results you come up with.

[RICE RACING]

In my own testing examples I am coming to the point where I am making too much power (for a road car) and it's getting stupidly quick! I did not think I would get to that point, but the more I research followed by long term tests in my own fully instrumented vehicle, the more I am finding the same conclusions Sir Harry Ricardo found in the early 1900's.............. there is indeed no real limit to how much power you can make on Water Injection, and you will reach the practical limits of 2ndary things before you will find the limit of stable combustion

Obviously Harry Ricardo and Frank Walker were on the money, I too have after running my own bank of tests am finding exactly the same. When you follow their guidance (spelling out how to do this correctly) it just opens up another whole world of power and best of all durability that far exceeds what you had when running "less power" :) AMAZING STUFF!

[ricekikr]

Great thread. A couple of questions though. (I only got to page 15 + a couple other pages)

Which is better? Directly in front of the turbo vs on the filter head vs anywhere pre-turbo?

Which is better a 80* cone spray or a 160*cone spray?

I have a Garrett 50trim (46lb/min) what size jets should I use? Mainly I want to shift compressor map. I will be boosting around 30psi (which I think is the limit of my turbo), will the water injection pre-turbo help?

Will a separate pump/controller be needed or can it be T-connectored to the post turbo nozzle?

[Grant M]

I have read all 41 pages of this thread and feel I have increased my knowledge of this subject ten fold!

Is there an ideal distance to mount a pre turbo WI?

I am currently in the middle of building a Mazda rx4 with a bridge ported 13B, BorgWarner s300sx3 turbo. Admittedly it isn't a small turbo but it isn't a big one in my eyes.

Now from what I gather most of the I put says that pre comp WI sees the greater benefits when used on the factory turbos which are to small and are operating outside of their efficiencies? Would I still gain the benefits of pre comp WI on my S300 turbo?

Before I read the info on this thread my major concern from other people was that using pre comp WI was that the water may drop out of suspension in the air when passing through the intercooler, but from what I have read this suggestion has only been mentioned once or twice out of the whole 41 pages!

Am I right in saying that the water will not drop out of the air due to the fact that when the water air mix reaches the IC it will have evaporated in the heat, cooling the charged air, making it more dense (more oxygen) and not actually have the time to form into larger droplets and pool in the IC?

From water I gather spraying WI into the centre of a conduit and not in from the side ( spray hitting the other wall essentially pooling the water)

Would I gain a benefit to run WI after the IC aswell to further cool the air?

I am hoping to run a the WI system that Rice Racing uses as I have discussed a lot of things with him on other forums and he has a rotary application not to dissimilar from mine.

[mr2turbored]

i to am seriously thinking about doing this after a good read. my only issue is that i'm using a tiny turbo and very short air intake

the compressor intake is only 48mm so fitting a jet right infront will obstruct most of the air intake. Right before the compressor there is a 90 degree bend at 48mm diameter which the outside is to close to the oil filter to fit the jet there pointing at the compressor. But after that bend its 12" of straight 2.5" diameter pipe with the air fitler at the end. My idea was to fit the jet into the end of the air filter pointing straight down that longish intake.

Would that be ok?

also with using the pre turbo setup, i would still like to run a normal setup pre TB, would i be able to just run a T section off the pump to power 2 jets or do i have to run a twin pump configuration? I'm using the Aquamist 2D system and would like if possible to run 50/50 meth/water

Any issues do you think?

thanks

[aquanew]

Has anyone tried using a mist maker to vaporize water mix pre-turbo?

[mr2turbored]

anyone??? :)

[stevieturbo]

2 jets off a single high pressure pump will work just fine.

[Dust]

MR2,I have seen several people with cone filters tapped at the top for a nozzle, so I did it myself, and my filter is right at the turbo. See my warm air intake thread.

[zakshaker]

Hello all,

I'm actualy thinking of installing a preturbo jet to my HFS-3 system.
I actualy run one single 1mm jet just before throttle.
now, where to fit the preturbo jet is the real question.
I understand it is recomended to install it just in front of impeller. But this is a rally car with a 34mm restrictor just in front of inducer. so I don't realy want to restric this area with any extra item (jet).
I'm wondering what is the downside to have the jet close to airfilter ?

http://i394.photobucket.com/albums/pp24/zakshaker/glanza/DSC00978fb.jpg

I did this pic before instaling the HFS3 (there is no more BOV now, air filter has been relocated in front of engine bay, but stainless piping is still the same), I've been thinking to install a 0.4mm preturbo jet on the stainless air intake piping you can see on top of exhaust manifold (just next to black rubber piping which goes to air filter) what do you guys think of it ?

[Howerton Engineering]

We have many customers inject pre-compressor now. If it is not possible to inject right before the turbo, then as far away from the turbo will work if it is a smaller jet. You want to try to have the spray evaporate and break down as much as possible before it enters the turbo. Right after the air filter should work. Good luck.

[zakshaker]

thanks for your answer, the stainless piping which goes to the turbo inducer is about one meter long. Does it sound far enough ? the jet would be just on the right side of the "TOYOTA" writing. I was wondering if there is any risk of water reforming on piping wall ? what is the max jet size which could be used in this typical area ? could I just plug it on the side of the piping ? of should it be installed in the center of the piping ?

[Howerton Engineering]

There will always be some water that gets on the piping wall, what you need to determine is if this will be an issue. If the air velocity is high and hot and dry, then it should still evaporate off. If it's a cold ambient temp and very humid, then it will have a harder time evaporating. Adding some methanol would help.

You can stick it in the pipe wall this far away from the turbo. The jet size would be by HP and engine size, but I would start with a .3mm or .4mm and see how it goes.

[zakshaker]

Thank again for your help

Air velocity should be quite high as I start spraying at 0.6 bar boost (TD04l turbo) which happen around 3000 rpm
I live in tropics island, So it is quite Hot most time of the year. average weather temp is 22-30 ?C. It never goes less than 15?C. and it can go up to 35?C in summer peak. Now it is never dry, it is pretty much humid (around 80%) . I plan to use WM50:50

[T.F.S.]

Hi Guys

Interesting thread..

I have an MR2 turbo...it runs no intercooler for one reason or another..

http://i42.photobucket.com/albums/e327/TEFUS/2012-05-13-207.jpg

At present it has a red 0.7 and a green 0.5 jet on the old style aquamist pump...50/50 ethanol/H20 triggered at 6psi.

Engine is bone stock rev2, it runs 20psi boost on a GT3071R with a link G4 ECU on 880cc injectors at around 60% DC, present power is around 350bhp/350lbft

It has been mapped with the methanol/H20 and also runs 95 octane fuel..very happy with the amount of timing advance we could get from the motor with it setup like this, when we mapped the car I still had the small sidemount intercooler and temps with a single 0.7mm jet was fine and even on an ambient 25c day after 1 hour of repeated mapping runs on the motorway we never saw over 35c degrees

As mentioned the intercooler has now been removed and even with the addition of another 0.5mm injector temps are very high and have been recorded at 90c on my datalog, i hope to get this on video but for now i just have a still image


http://i42.photobucket.com/albums/e327/TEFUS/G4.jpg
http://www.youtube.com/watch?v=s6OX4vBXBpA&feature=youtu.be

I have recently ordered a devils own kit and hope to inject pre turbo as well as using the present kit in its present location, I will update this thread with the findings on the datalog as i go, i am only recording temp just before the TB and I am only interested in reduction of inlet temps at this time.

I thought some of you guys might be interested to see what happens, any ideas as to how much extra H20 I will be able to inject?, will I be able to get down to 30c with H20 alone?

[T.F.S.]

Second kit added.

Twin tanks in the boot, separate one for each kit
http://i42.photobucket.com/albums/e327/TEFUS/2012-06-02-242.jpg


The Devils-Own injector will be placed looking into the "eye" of the turbo as close to the impeller as possible doing this is not easy....the best position on offer is firing the injector from the wall of the hose so I adapted some mesh to give me the optimum position, lets hope it stays in place and does not offer too much restriction :0

http://i42.photobucket.com/albums/e327/TEFUS/2012-06-01-229.jpg

http://i42.photobucket.com/albums/e327/TEFUS/2012-06-01-228.jpg



the injector is just spraying pure water at the moment and hopefully will help lower temps a bit closer to 30c, this injector should flow circa 400cc per min
http://i42.photobucket.com/albums/e327/TEFUS/2012-06-01-230.jpg

http://i42.photobucket.com/albums/e327/TEFUS/2012-06-01-234.jpg


Air filter in position
http://i42.photobucket.com/albums/e327/TEFUS/2012-06-01-238.jpg

http://i42.photobucket.com/albums/e327/TEFUS/2012-06-02-243.jpg



After a test drive in cooler conditions that the last test the inlet temp has dropped by circa 40c
http://i42.photobucket.com/albums/e327/TEFUS/bothsystems07054060h20andD04100h2015camb ientwetroad2psitrigger19psimax.png
http://www.youtube.com/watch?v=dEX08Fos0O8&feature=youtu.be


Ambient temp was 16c, injection trigger point for the Aquamist system (post turbo) was dropped to 4psi, Devils system (pre turbo) was triggered at 2psi and ramps up to max flow at 19psi (boost sensative progressive system)

Boost temps were seen to be around 40c, I never saw 50c on the test but I could not pull max RPM on a wet road and ambient temps were quite low.

I will try for a slightly larger injector soon!

[Grant M]

A 40 celcius drop in temp by just adding water is brilliant. Very interested to see what results you will net by playing with different nozzle sizes and if you run WM50/50 pre turbo. The results you are showing are indicating that temps can be run low enough without the need for intercooling at all.

Has the elimination of the IC reduce turbo lag? If you had any that is?

Are you running this car on the road?

[T.F.S.]

A 40 celcius drop in temp by just adding water is brilliant. Very interested to see what results you will net by playing with different nozzle sizes and if you run WM50/50 pre turbo. The results you are showing are indicating that temps can be run low enough without the need for intercooling at all.

Has the elimination of the IC reduce turbo lag? If you had any that is?

Are you running this car on the road?

Yes I intend to play with Ethanol mixture soon...I am going to rely on the Devils kit alone due to the pressure it runs at and add an injector just after the compressor where the present Aquamist injectors are and run a 50/50 mixture....or maybe I should run everything pre turbo?

The car is a daily drive and runs very well like this..it seems to have less lag than before too but I have no data on this until I get to the dyno.

For more info on this car please see.
http://www.imoc.co.uk/forums/viewtopic.php?t=124264&highlight=

[Grant M]

You will be able to see whether or not if you need the post turbo injectors by disconnecting it and going for a drive and logging everything.

Rice racing on here will be able to give any advice about pre turbo injection as he's been running that for a while on his rx7. He uses a water/methanol blend 50/50 by weight iirc injected preturbo. He is running up to 35psi on his T04Z which is prety much the turbos limit without overspeeding it.

You may also want to try moving the pre turbo injector further away from the turbo so that the atomised water can help cool the air being sucked in even more. That is If you have the space.

[Grant M]

You will be able to see whether or not if you need the post turbo injectors by disconnecting it and going for a drive and logging everything.

Rice racing on here will be able to give any advice about pre turbo injection as he's been running that for a while on his rx7. He uses a water/methanol blend 50/50 by weight iirc injected preturbo. He is running up to 35psi on his T04Z which is prety much the turbos limit without overspeeding it.

You may also want to try moving the pre turbo injector further away from the turbo so that the atomised water can help cool the air being sucked in even more. That is If you have the space.

[stevieturbo]

@ TFS.

You do realise that mesh screen will be posing a huge restriction to airflow. Get rid of it !

[Dust]

TFS, is the end of the filter metal or filter? I tapped the end of a motorcycle filter and used it for summer preturbo work. My filter wasn't as large or long as yours, but with the airflow that should be incoming, and not through-flowing, fluid contact with the side should be less likely, unless your hobbs switch gets set to open the solenoid before the spray starts :(

[T.F.S.]

Guys I am unsure of how much of a restriction that actually represents...it the mesh is more corse than say an AFM?, air filter has a rubber end..

[Howerton Engineering]

Just some general info; Sometimes the open area of a screen cannot be equated to the airflow through that screen.

If you look at a normal household window(insect) screen, typical in the US, they will generally cut airflow by 50% through an opening. All screens are a bit different, but the numbers are sometimes surprising.

The screen in the pic appears to be a bit more coarse, but it could be a significant restriction.

[T.F.S.]

OK I removed the screen and fitted the injector at the filter end....not happy with the distance from the compressor but until I have something custom made thats not going to change.

Thanks for the advice.

[Grant M]

dont take what i say as gospel but from what i understand by reading the posts made by people on here, having the injector further away from the turbo should further reduce the intake temps, i think its to do with the atomisation of the water, water droplet size taking heat out of the water air mix. so long as the droplets size is small you shouldnt get any problems with compressor wheel damage

[T.F.S.]

Ok so the next stage I took was the removal of the Aquamist system and I added twin injectors to the Devils system, 600cc pre turbo, 400cc post turbo (D04/D05)

I relocated the pre turbo injector to the end of the air filter, this we know is not ideal, I hope to change this soon.

Water injection supply lines were rerouted away from heat source.

Post turbo injector 400cc (D04)
http://i42.photobucket.com/albums/e327/TEFUS/2012-06-14-276.jpg

Devils progressive controller pic to be added.

Pre turbo injection position to be added.#






Datalog..
http://www.screencast.com/users/monosodiumphosphate/folders/Jing/media/86543bc4-03ce-440d-b950-ab44e89ddcc0


Ambient temps were back up to 22c and this is the first daytime datalog after my one last night was lost due to incompetence lol

I used a 38/62 ethanol/water mixture and I was surprised that if anything temperatures were higher than the previous run...yes ambient temps were 4c higher than the previous test but I saw almost 70c at one point at the very top of the RPM range at circa 130mph?...in general temps were around the 50c mark on boost.

I thought a total of 250cc extra liquid was injected on this run but I had incorrect advice as to the flow rates of the injectors

I planned to inject a mixture of 78% alcohol* rather than the 38% we had in the last run to test this theory.



The temp did not drop and after a close look at the injector flow rates
https://www.alcohol-injection.com/images/flowchart.gif
and it seems I am simply not injecting much more liquid than my previous tests...only 750cc per min total, another injector is planned for the next test to bring injected amount closer to 1150cc per min




*due to a lack of ethanol I am using a small amount of isopropanol in the ethanol/water mix to bump up the alcohol content.

[Dust]

I think the D05 is your problem. DVLDOC has said to stay D02 or lower pre-turbo

[T.F.S.]

Interesting, I did wonder if there is a limit to how much I can add pre compressor but I thought if looking to reduce inlet temps the more mix I put through the turbo the better?

Also I have been told by Devils own to double the flow rating on the injectors if using their 250psi pump but I have also been told by a user that this is not true and I should go with the original ratings...if the latter is true it explains a lot because I wouldn't be injecting any more than my last test and TBH I have had very similar results....

Sure I have had a higher inlet temp at 7000rpm but due to a wet road I was unable to test above 5000rpm on the previous test so we cannot compare at over 5000rpm.

The D02 would work out at only 120cc in that case...a very small amount for pre compressor?

[Dust]

Also I have been told by Devils own to double the flow rating on the injectors if using their 250psi pump but I have also been told by a user that this is not true and I should go with the original ratings...if the latter is true it explains a lot because I wouldn't be injecting any more than my last test and TBH I have had very similar results....


The 2 gph is at 100 psi. The flow will be greater at 250 psi. it's not double, but it is greater.

https://www.alcohol-injection.com/images/flowchart.gif

[T.F.S.]

So in total about 750cc going by that graph eh.

[T.F.S.]

The D04/D05 injectors are moved back to the injection points I was using with the Aquamist system...I adapted the injectors for straight fitment at this time.

http://i42.photobucket.com/albums/e327/TEFUS/2012-06-22-285.jpg








Rear view of injection point/s

http://i42.photobucket.com/albums/e327/TEFUS/2012-06-22-286.jpg







Fitted.
http://i42.photobucket.com/albums/e327/TEFUS/2012-06-22-289.jpg








http://i42.photobucket.com/albums/e327/TEFUS/2012-06-22-300.jpg







D05 pre injector moved back to the turbo entry point..I set it back about 1/2 inch from the turbo
http://i42.photobucket.com/albums/e327/TEFUS/2012-06-22-304.jpg







Dump valve added
http://i42.photobucket.com/albums/e327/TEFUS/2012-06-22-293.jpg





http://i42.photobucket.com/albums/e327/TEFUS/run4.png



The engine bogged and misfired on this run with well over 1000cc per min being injected, temps were closer to 40c on this run and hit a max of 46c on an extended high speed run at near 130mph, I would have liked to have added a full datalog for this run but I am having trouble using the screen recording software.

I played with different mixture settigns but the misfire would not clear...it seems that either the mixture was simply too rich (dipping under 10.0/1) or had too much water content to ignite.

With the coilpack dwell time being too high as it is I will need to get the car remapped to take the extra fuel into account...for the moment I will drop a jet size for around 900c per min and take some more measurements.

I may also use more smaller jets amounting to the same injected amount to offer a more atomised mixture to the inlet to help drop temps even more...I dont know how much of a drop this can provide but right now anything is worth a shot!

[T.F.S.]

I fitted two more injection sites in the turbo scroll and fitted much smaller D075 injectors in those areas...the idea was to give a more even distribution of cooling mix within the turbo and reduce the chances of dropout on the walls of the turbo whilst also taking advantage of the very slightly better atomisation that a smaller injector/s provides.

http://i42.photobucket.com/albums/e327/TEFUS/2012-06-27-348.jpg

http://i42.photobucket.com/albums/e327/TEFUS/2012-06-27-351.jpg



The pre turbo injector has now been reduced to a D04.


I took the car for a test run last night and although the WI system was providing something like 25% less coolant/mixture and ambient temps were higher than previous tests at around 21c the inlet never went over 50c even on an extended high speed run....something is working a LOT better here!

The mixture was set at around 50/50 (water/ethanol) and another factor I should mention is that the dump valve was also leaking quite a bit...I am sure this only would have made the turbo work harder and increased inlet temps so now that has been repaired I may get some datalog tonight.

Methanol has not arrived yet..

[stevieturbo]

Anyone any interest in this ?

http://www.evo-host.co.uk/getimg/29348.jpg

After a little grinding and filing

http://www.evo-host.co.uk/getimg/29349.jpg

I'm not saying it's perfect, but its cheap.

I had it cut from 4mm thick aluminium and all legs are 4mm thick also.. This one is 102mm OD so it's a snug fit inside my air intake pipe.
I made the external legs 10mm, these can be cut down. Idea being drill a hole into the silicone tube and these legs will secure the fitting from going anywhere.
Again cheap and easy to use.

Just trim and adapt as necessary. Could get them made in other diameters if anyone interested ?
Or could have legs reduced to a smaller size, I was just paranoid about strength so opted for 4mm. 2-3mm would probably do fine.

This is my first attempt at pre-compressor, although I'm also spraying a lot pre throttle as well. Centrifugal supercharger application

[T.F.S.]

Yea I would be interested to hear how much they would cost.

[T.F.S.]

Well moar methanol arrived today so I filled up a 15/75 water meth mix and went for a datalog run...I managed to change the software so the inlet temps are much easier to see too!

After the last run on a 50/50 mix I was expecting temps to be quite low but this was not the case!!!, the temps smashed into the 80c region pretty dam quick

http://i42.photobucket.com/albums/e327/TEFUS/15-75h20-methrun1700cc.png

I am not sure why that is...its pretty much the worst temps I have seen yet and I had to check the tank volume to assure myself that the WI system had activated at all on this run

AFR was hitting 10.0/1 which is the lowest my AEM will read (so who knows what the real AFR was)

If this is going to work we are going to need a LOT more liquid injected at this mixture level


Another two runs on different mixtures show the same thing...massive spike on first application

Run 2 50/50 methanol/H20
http://i42.photobucket.com/albums/e327/TEFUS/run250meth.png


Run 3 50/50 ethanol/H20
http://i42.photobucket.com/albums/e327/TEFUS/50ethanol700cc.png


It was obvious something wasnt right here...thats not what you expect from a WI system, a leak was suspected but upon checking could not be found :(

[T.F.S.]

So it seemed that rather than a leak it was air that was pushing through the post turbo injectors and managing to empty the lines through the pre compressor injector. This gave a delay in injection delivery so temps were seen spiking upon first application in the logs above.

This was rectified by fitting a check valve just before the pre compressor injector.

http://i42.photobucket.com/albums/e327/TEFUS/2012-07-05-364.jpg

A D05 injector was also fitted between turbo and dump valve at this point.

http://i42.photobucket.com/albums/e327/TEFUS/2012-07-07-371.jpg



http://i42.photobucket.com/albums/e327/TEFUS/2012-07-07-372.jpg

The next run went very well..helped along by low ambient temps (16c) and a nice dry road allowing me to hit max RPM and hold the throttle wide open for some time, temps were back in the low 40's at max RPM and stayed in the 30c region at lower RPM's which is almost exactly what I aimed to get at the beginning of this project :)

The initial spike in inlet temps we saw in the last log was absent from the data, the check valve doing its job!

The car didn't bog or misfire, nice clean power although fueling is (as expected) still pig rich and requires some remapping to clear up.

I thought I had saved the format from the last log I done but for some reason it reverted back to normal so I have done several screenshots.


Initial application of throttle from 2k through to approx 6750rpm (engine limiter), 42c max inlet temp recorded
http://i42.photobucket.com/albums/e327/TEFUS/2012-07-06_1.png



Into third, max inlet temp 43c at 6750rpm
http://i42.photobucket.com/albums/e327/TEFUS/2012-07-06_2.png



6750rpm in 4th gear shows 44c
http://i42.photobucket.com/albums/e327/TEFUS/2012-07-07_4th_gear_pull.png

The other two high RPM peaks further over to the right again max out at 41/42c


I may add a touch more WM mix for the next log.

[RICE RACING]

Good to see someone taking leaf out of my book of testing :)

NOTE:
I'd like you to refine your Link G4 records and simply crop the screen shots down to the applicable test (be it 3rd gear pull or whatever) and give us a start marker and and end marker with the cursor on each point. *see my tests of like 90kmh to 140kmh* 3rd gear or whatever you choose to do.

100kmh to 200kmh is nice, though if somewhat impractical to test each and every time.

It will make it easier to follow and compare as you keep adding lots and lots of info on your path to discovery.

It will also help you quantify in a basic way performance tests if you pick a standard and stick to it, then you can easily measure time plots to speed increments etc, its never 100% accurate using ECU data but its a great guide if you keep the variables in check and dont cheat yourself running under weight or down hill lol.

Keep up the good work, its one man testing like this that make this site worth reading and inspires people of all backgrounds, resource levels and skill levels.

Peter

[T.F.S.]

Most if not all of the tests are beginning of 2nd to end of 4th on the same stretch of road, sometimes I cant pull top RPM due to wet roads (we have a LOT of that in the UK at the moment)

I don't think I have seen any of your testing threads, are they on here buddy?

[RICE RACING]

Most if not all of the tests are beginning of 2nd to end of 4th on the same stretch of road, sometimes I cant pull top RPM due to wet roads (we have a LOT of that in the UK at the moment)

I don't think I have seen any of your testing threads, are they on here buddy?

http://www.aquamist.co.uk/vbulletin/showthread.php?t=1590&page=25

There you go.

[RICE RACING]

SNAP SHOT:

I am running another 100kpa more boost than you are, with lower charge temp. Very small factory IC being used for historic reason (1 of 25 race special made cars I test this stuff on).

Have a read, there is lots and lots of good info, may save you some time and grief ?

I recommend an inter cooler and WM50 injection and spraying the IC@! but this is at very high levels of power, as proven by vehicle performance linked above.

Anyhow, keep up your testing :) its people like you that make this site worth while.

[T.F.S.]

Another D05/5gph injector added..hope to shave a few more degrees off the inlet temps on the next midnight run

http://i42.photobucket.com/albums/e327/TEFUS/2012-07-09-376.jpg

I would guess that the temp just before the cylinder is a good 5 degrees lower than the temp i have recorded pre throttle body...IIRC atomisation suffers at under 30/40c so I have pretty much suceeded in what I aimed to do at this point and the extra 5gph injector is not needed really.

[stevieturbo]

When dumping so much water/meth in, it could become hard to actually measure intake temps.

How much is actual charge temp and how much is simply the fluid temperature on the sensor ?

Has anyone done any testing before and after the nozzles using a few sensors ?

Maybe stick a thermocouple into the intake manifold close to the intake valve, so hopefully any state changes have already occurred by then so may give a better picture of actual charge temp ?

I did try my own car a few days back with 2 nozzles pre TB and a single 1.1mm in front of the compressor.
There was no measurable performance gain over no WI despite running slightly leaner with some 5deg more timing.
Although I suspect it's down to me fitting a new 200psi pump as well. I'm probably just drowning the engine with water/meth 40/60

need to resolve a few other issues before can try anything more.

I'd previously been using an 5 year old 150psi pump and two of the largest AEM jets. the new 200psi pump must be flowing a huge amount more !

[Howerton Engineering]

Water does not do a great deal for IAT cooling after the compressor. The boiling and flash point of methanol make it much more effective for IAT cooling. What happens in the chamber is another matter and argument, but for purely IAT reduction, pure methanol is the most effective.

We have measured a 140F drop in IAT to 40F below ambient at times with pure methanol.

[T.F.S.]

Interestingly the extra injector didn't make a lot of difference...I expected in the region of 5-10c drop now we are injecting around 1350cc per min but it was more like 0c-1c. The system is not leaking and I am still wondering why we did not see a more significant decrease , could it be that no more methanol can evaporate into the airstream and I either need to increase water content or add another chemical into the mix?

I believe that 250cc is about the maximum amount of water I can inject into a 2.0 engine at this boost pressure..whenever I have tried to inject beyond that figure on this engine it has bogged or misfired so with a 10-15% water content in the tank on this run it would seem I am almost at that limit already

On a much longer run than normal on a 20c daytime blast we saw temps in the low 40's at high rpm and at one point just as I came off of the throttle a split second before I changed gear temps hit over 70c which again is not something I expected!


http://i42.photobucket.com/albums/e327/TEFUS/2012-07-13_0955_1350.png




Maybe stick a thermocouple into the intake manifold close to the intake valve, so hopefully any state changes have already occurred by then so may give a better picture of actual charge temp ?


Yes I may try this!

[FlyingGrape]

Maybe try something like a methanol/ethanol/water mix? Would be good to see some data on adding another chemical once you have possibly reached near saturation with the others

[T.F.S.]

I have been told that I would need to inject far more than I am to hit saturation on the meth but I may introduce another chemical anyhow and see what happens...it may just be that the 4c increase in ambient from the last run accounts for ther missing temp decrease

I plan to do another midnight run later tonight to test that theory.

[T.F.S.]

No real updates this week due to the progressive controller failing again...
https://www.alcohol-injection.com/forum/technical-questions/progressive-controllers-3568.html

I had some time to fit the new pre compressor mount

http://i42.photobucket.com/albums/e327/TEFUS/2012-07-20-394.jpg


http://i42.photobucket.com/albums/e327/TEFUS/2012-07-20-395.jpg


Also the blades on the compressor are still suffering from wear..
http://i42.photobucket.com/albums/e327/TEFUS/2012-07-20-393.jpg


Just waiting for a Coolingmist progressive controller to arrive so we can continue with the thread

[T.F.S.]

Another change to the pre compressor injection point and I have a feeling it may not be the last, the nozzle has been brought forward just over 1.5 inches..it is now within 5mm of the compressor nut, maybe this is too close for comfort?

http://i42.photobucket.com/albums/e327/TEFUS/2012-07-24-412.jpg




http://i42.photobucket.com/albums/e327/TEFUS/2012-07-24-414.jpg






Another change to the pre compressor injection point after noticing the wear on the compressor in the previous post and I have a feeling it may not be the last, the nozzle has been brought forward just over 1.5 inches..it is now within 5mm of the compressor nut, maybe this is too close for comfort?


http://i42.photobucket.com/albums/e327/TEFUS/2012-07-24-412.jpg




http://i42.photobucket.com/albums/e327/TEFUS/2012-07-24-414.jpg



It's not easy to strike a good medium with the injector placement...I think the previous nozzle position may have allowed the WM mix to hit the sides of the turbo inlet and pool into larger droplets which have damaged the blades as they have impacted, I dont know what is the best distance from the compressor but after rigging a very basic test I believe it to be 8-12mm away from the compressor thread/nut


Well I have a 3071 core sitting around due to a rough bearing, it's not easy to hold the injector and work out an exact distance for the best placement so I took a video to watch back afterwards.
Now as crap as this vid is it would seem that the best placement would be 8-12mm away from the compressor nut would you agree?

http://www.youtube.com/watch?v=OH9wbHRJxbA&feature=player_embedded

[stevieturbo]

A good 25mm or so away looks better. But then what it looks like static, and what will be happening with the compressor spinning will be very different
I'd still recommend using the Aquamist jets for this. They are just so much neater, and would be impossible for it to unscrew.

Was there no damage at all to your compressor before you started this ?

[T.F.S.]

The turbo was new a few months ago so no prior damage...I have another picture to add to an earlier post so you can see the damage progress.

I have wondered what is happening to the spray pattern on boost...does it try to narrow the beam so it hits the nut or does it direct it to the beginnings of the blades?...looking at impact tests on jet turbines it would *seem that it is directed towards the beginning of the blades, with this in mind a shorter distance may be better.

We can only try to find out, I'll take a pic of the compressor every month and update, lets get some hard data on this.

[Grant M]

It may be worth you exchanging a few PMs with riceracing, he may be able to help you with pre turbo nozzle selection and positioning. He has done a lot of custom work to his nozzles and setup and hadn't experienced any wear.

[Richard L]

Peter has been very clear about how he achieved his set up, we have regular pm exchanges, discussing the WI applications.

I may need to drop a few other projects so I can spend more time on the air-assisted pre-turbo nozzles. I have a few ideas how to achieve it but haven't had time to build prototype nozzles and the associated flow control mechanism.

This weekend may be good to further the project.

[Richard L]

TFS,

Depending on the spray angle of the jet, you may not need to have the jet too close to the compressor nut. If it hits the compressor nut, the spray pattern can be very unpredictable.

You really need to find out the spray angle before deciding on the nozzle distance.

[T.F.S.]

Thanks for the advice guys.

The spray angle can be seen in the video posted at the end of my second to last post.

The video tells us nothing though really...the goalposts change when the compressor spins up, I expect that the spray pattern will widen though so a closer arrangement may be better..in the 8-12mm range IMO.

IIRC rice said to use a D02 injector pre turbo..I am trialling a D04 so using double the amount, I do have a spare wheel in the garage though so I am not overly concerned about wear for now.

[Grant M]

Can the impeller wear on pre turbo injection be the result of the speed at which the Injectant is Injected into the compressor? Due to some of the systems I have seen the pumps are at a very high pressure to gain atomization of the injectant?

Looking between a few of the system available the Rice racing kit doesn't seem to force the mixture out at a great velocity, due to him using turbo pressure to drive the system whereas other pump driven systems do.

Just a thought iv been having to explain why people experience wear and others don't.

[stevieturbo]

It's simply down to atomization.

Big drops can hurt it, very very fine mist will not. Rice doesnt need high pressure to achieve this due to nozzle design.

[reid-o]

I've had an air atomizing pre turbo kit (500cc at 20 psi fluid pressure)
for some time that I've been running mechanically along with a Aquamist HSF 2 system

The nozzle fired into a gtx3071 with a 63 hotside on a Subaru EJ257, which if you're familiar with Subarus, is a rather restrictive setup on the exhaust side at high rpm and boost pressures. My thinking was to create a very broad torque by using wet injection to increase compressor efficiency and thereby reducing the exhaust energy required for a given flow rate. I was hoping that this would help to reduce EGBP a tad and give me a great responding street setup that was fun to drive but without the high egts typically associated with running small turbines (I know from experience that these combos are sensitive to timing increases)

While the setup appeared to run fine, it didn't seem logical to have invested in a PWM system only to have half of the system running 2 dimensionally. I've been bothering Richard and Jeff at Howerton for ideas to plumb the air atomizing nozzle into the Aquamist system. While it's easy to use restrictors to control flow and linearity, it seemed as though lowering fluid pressure to the atomization air pressure was the issue.

I did the following tests:
1) plumbed into the HSF2, with the flow valve on the air atomizing nozzle bottomed, I was able to get 300cc/min of flow with 100% duty cycle (HSF2 in test mode)
2) with air pressure on the line at 25psi, atomization quality appeared to be about the same as a standard aquamist injector (making the nozzle worthless)

I also noticed that, despite what others have argued as a non-isssue, I felt that this particular air atomizing nozzle's air ports were rather large, and when boost leak testing my intercooler pipes, it was clear that this bleeding would in effect require higher than average shaft speeds to at least reach the turn on point of the injection system.

After some thinking, I've come up with the following solutions.
1) After disassembling the Air Atomiznig Nozzle (AAN), I noticed that the system didn't use internal mixing. The chambers for fluid and air were separate. The atomization occured at the tip of the nozzle.
2) The separation block used 1/8 NPT ports.
3) Although the block had separate passages for fluid and air, there was a small hollow cavity on the fluid side.

I found an old 1/8 NPT water injection nozzle rated at 300cc 200psi line pressure. I figured what the hell and screwed it in with a leftover check valve i had from a previous kit. Plumbed into the (diagram below) HSF2, I found that:

1) water flow was roughly 200cc per minute (aquamist pump is rated at 160psi)
2) pressure at the outlet of the AAN was significantly reduced
3) atomization quality increased significantly!!! I could get the droplet size to be much finer than the standard aquamist nozzle at max duty cycle even with lower air pressures (down to 10 psi)

Next, since I am running a standalone ECU. I decided to plug in MAC solenoid to interupt the air pressure (I have a 3d map for this) mirroring the activation of the HSF2.
Working together, immediately I noticed that
1) response has picked up significantly (the pressure bleeding although not significant enough to prevent pressurization is significant enough to delay spool by a tad)
2) The loss in torque right at peak boost pressure has disappeared but the top end has not. I could never get the 2 system to work together, even though I had the HSF2 on a separate table. I'm a bit obsessive compulsive about low load areas, and the slight dead spot at peak torque was driving me nuts. I didn't whether to pull fuel, pull PWM from the aquamist, or reduce the flow control valve on the AAN.

Yes I know that the purpose of the pumpless system is reliability and KISS. I'm not intending to argue otherwise. My reasoning was simply that I had a full PWM system, and I wasn't using it. Placing the water injector into the fluid port on the AAN and allows the pressure to dissipate as the fluid fires into the small cavity of the AAN. Fluid flow remains unchanged from the rating of the water injection nozzle. I am now on to serious testing, as I can stay in the car, examine logs and make adjustments without getting out. I just thought I'd share in case anyone wanted to combine theirs into one system. Although I've solved my the issues that I created for myself, I have more variables to play with. The beauty is that now I can add a failsafe to the system, as I can pull fluid flow from one line. I have PWM control over RPM and load (3d table). I don't have the data right now to export, but it's clear that, since I've likely reduced total flow by 20 percent, I have engine noise near peak torque (very bad). That was on my initial pull. I've increased the flow in those cells, and it went away. Now I need to think about the ratio of pre to post turbo flow, but I think I'll see how much torque I can wring out of this setup right now.

(water injection nozzle is on right side)
http://i281.photobucket.com/albums/kk214/rsunahar/IMG_0682_zpsba2e588c.jpg

http://i281.photobucket.com/albums/kk214/rsunahar/2013-02-17_1412_zps9c9b3ff9.png

Here it is with the filter attached directly to the turbo. I plan to fabricate a cold air system soon.
http://i281.photobucket.com/albums/kk214/rsunahar/IMG_0683_zps6f28a80c.jpg

[Dust]

I'd like to know what your total flow is out of the AAN nozzle, and what you are using. I had considered using an inline speed control valve to reduce flow, and had also considered using an outside mount nozzle as a restrictor on a PPS system as a return to tank for keeping the pump and pressure up. Let us know how this works for you, especially when you extend the distance from nozzle to turbo. I would like to try it, but the car I am looking at has a good distance from inlet to turbo.

[reid-o]

I'd like to know what your total flow is out of the AAN nozzle, and what you are using. I had considered using an inline speed control valve to reduce flow, and had also considered using an outside mount nozzle as a restrictor on a PPS system as a return to tank for keeping the pump and pressure up. Let us know how this works for you, especially when you extend the distance from nozzle to turbo. I would like to try it, but the car I am looking at has a good distance from inlet to turbo.

200cc ish with the water injection nozzle plumbed into it. I explain it in the post.


It's designed for 500cc with 20 psi water pressure. It's important that the air pressure is a tad higher than the fluid pressure. This increases the atomization.

[reid-o]

My ideal setup right now would be an air atomizing nozzle that could be run at high liquid pressure to low air pressure.

[Richard L]

I wonder you can put a 1 psi checkvalve on the water line so the air pressure is always a tad over the water pressure.

To achieve the reverse, ie water pressure is higher than air pressure may be more difficult. May require redesigning the nozzle. Perhaps contact the air nozzle maker, they may be able to help.

[reid-o]

I wonder you can put a 1 psi checkvalve on the water line so the air pressure is always a tad over the water pressure.

To achieve the reverse, ie water pressure is higher than air pressure may be more difficult. May require redesigning the nozzle. Perhaps contact the air nozzle maker, they may be able to help.

From what I understand, there's a big difference between an internal mixing atomizing nozzle and external mixing. It seems as though the internal mixing requires higher air pressures while the external mixing requires less.

I was thinking of experimenting with this with a cheap HVLP gun I have at home as those are externally mixed.
http://img.photobucket.com/albums/v425/MississippiMan/Wall%20Prep%20and%20Painting/AWagnerHVLPNozzle.jpg

I'm wondering if one could drill out the center orifice for a standard aquamist nozzle whilst pressurizing the external air ports for increase atomization. Maybe I'll play it with this weekend.

[Richard L]

It appears that you are not going to stop until you have perfected it!

I have modified one a few months ago, it worked. I drilled a through hole (for air) and water enter from the side wall. Worked better without and water pressure. water is drawn up by the venturi effect of the air jetting out of the front orifice. As soon as I lowered the nozzle below the water level, the atomisation got worse. I did have time to go further with this.

Here is a picture:

http://www.aquamist.co.uk/forum/spinners1ss.jpg

http://www.aquamist.co.uk/forum/air-jet2s.jpg

http://www.aquamist.co.uk/forum/air-jet1.jpg

[reid-o]

I think that this would be considered an internal mixing type and wonder if designed like the one above you could just increase the atomization of the standard nozzles by using air to further propel the atomized water. This type wouldn't rely on the siphoning effect of the venturi, and I think this is why the external mixing types don't require high air pressure. My theory could be wrong though. I haven't yet found the parts to start, as I'd have to machine a block that would allow insertion of the nozzle and keep the air separated in a separate chamber.

Without trying, I would think that the variables right now would be the effect the propulsion has on the spray angle, as well as whether the current jets spray angle is too wide and will make contact with the air holes. The concept seems sound though.

[Richard L]

It is definitely internal. I did it for curiosity.

I like to hear how well the external type works if you mind posting some results. If the atomisation is good, the spray angle won't matter too much. hopefully it will be fog like. If it workout, you can PWM a valve to control flow.

[reid-o]

I drilled out the nozzle orifice on my hvlp gun to insert a jet and tried it today.
I didn't have much time, but it seems as though the depth I place the nozzle matters. The atomization is increased slightly but I see the issue at heart. The dual air ports require that the liquid pass as the two air streams meet, and the water nozzle creates a fan. Only the water passing in front of the air focal point receives increased atomization. I think if there were small air ports surrounding the nozzle in a circle would work. I don't have the tools to build this though. Ill play with it more this weekend, but at the very least it has the standard atomization quality with some spots of improved atomization.

[Richard L]

If the jet of water is narrow, there is a good chance the entire stream is atomised?

[reid-o]

An easier way to think about is to design the air dispersing ports to cross the bulk of the water flow. Here are 2 external mixing nozzles

This one requires the water meet at the vertex of the air ports.
http://i281.photobucket.com/albums/kk214/rsunahar/4459716140cdbaeac56c6ae7397b006a_zps90eb 6760.jpg

This one seems like it would fit the existing aquamist nozzle, but it would require some thought into the angle of the air ports.

http://i281.photobucket.com/albums/kk214/rsunahar/cbdc9eb9686db769e84939384aeca4cb_zps5d19 a291.jpg

I haven't tested it further, but I think you could make a modular system whereby your standing nozzle screws into an external mixing base. It would likely be the second type.

[Richard L]

I like the multi-hole one, I think that looks wonderful.

You don't sleep much.

[reid-o]

I'm a grad student, so no :)

Okay, I've been noticing decreased spool time (opposite from what I've read on some boards) and have the nozzle pointing directly at the center nut. This tells me that the integration and atomization are key. I'm injecting a total of 850cc of fluid with the preturbo handling 280cc worth of total flow.

4th gear
Post intercooler only:22 psi by 3600rpm
Pre turbo activated at 4 psi (with post turbo): 22 psi by 3400 rpm

I'm wondering if using stages would work better. I haven't inserted the logs into a dyno simulator yet to compare, but I suspect that having the post intercooler jets activate a bit later would help keep EGT a bit higher till full spool. I don't have curves to compare yet, but seat of the pants suggest that torque increases on activation of the pre turbo jet supporting increased mass flow rate with reduced compressor work.

I started playing with timing, increasing a tad during spool to account for the early turn on time. Since I have full pulsewidth control over the system, I have it reduced on the low end such that the flow rate ramps up pretty quickly but begins at a very low flow rate.

I think that the major factors that affect spool are as what everyone has been saying "atomization" at low flow rates. My thoughts right now are that if using a pump, a constant pressure system is the only way to go pre turbo. I can see how reduced atomization would slow spool with the reduced atomization till of course the pump pressure ramps up. Of course if one runs a mechanical system with an air atomizing nozzle the atomization is not a problem as water pressure and air pressure would rise 1:1 keeping atomization good as pressure rises.

My questions now are about calculating total flow rate between the two locations and what works best. Also what can I do to make a dual stage system out of the HSF2 using only one FAV? Can I place a solenoid right before the post intercooler jets to turn on at a higher point? I'm thinking that the PWM will no longer be linear with only one jet spraying and will have a dip once the solenoid opens.

One other thought would be to run 2 FAVs and have a Hobbs switch on one that completes the circuit at a certain pressure.

[Dust]

Hobbs and solenoid after the FAV should keep you PWM, unless of course you don't want that.

[leman_opc]

Such a nice thread and immense portion of knowledge I'll have to structure somehow in my head.

Anyway, having considered the effect of pre-turbo injection for my setup I decided to give it a go with the primary objective to increase the efficiency of the turbo compressor. To get a feeling on how the pre-turbo injection may work out in my case I'm currently planning to install additional jet pre-turbo (now running single jet post-IC for knock suppression reasons primarily) via a tee to existing line (post PWM valve). Would appreciate a practical advice from the people here.

In my setup I currently use single 0,8mm jet flowing ca. 420cc/min bearing in mind ca. 1,5 bar max boost I have. Is this a good idea to go with a 0,3 mm jet pre turbo, and change the post-IC jet to 0,5 (i.e. basically reallocate what I’m injecting now) or should I rather consider adding another jet on top of what I have?

Also, my system now is referenced to IDC (progressive). Based on my reading here I note that most people reference the pre-turbo injection to boost, which is reasonable given the amount should correlate with air mass rather than fuel. Is there a high chance I face any issues here due to changing AFR across the map?

[leman_opc]

Did some field testing last days.

A couple of words on the car setup: 2.0 lt engine with K04 based stock turbo, boosting 1,5 bar (midrange) and gradually decreasing to 1,2 bar on the redline (presumably due to turbo maxxed out flow-wise). WMI installed some months ago for knock suppression primarily, the map I have on the car (an I'm not able to tweak it unfortunately) appears to be pretty aggressive with 0,9 lambda (13,3 AFR) in midrange (which is also peak torque range obviously) and 0,8 lambda (11,8 AFR) in higher revs. Without WMI I have ignition retards of up to 12 degrees, with a single 0.9mm jet post-IC the retards become sporadic and mostly up to 4 degrees, which I reckon is acceptable. I inject 40% ethanol, 60% water.

The idea is to reallocate some injection to preturbo jet so that the effects on (i) knock suppression effectiveness may be tested (pre-turbo vs. post IC) and (ii) claimed increase of turbo capacity may be estimated.

I made test runs in 3 different setups:
1. Single 0,9 mm jet post-IC. Result: max. air consumption 790 kg/h, ignition retards up to 4 degrees;
2. 0,3 mm jet preturbo, 0,5 mm jet post-IC. Result: max air consumption 789 kg/h, ignition retards up to 8 degrees;
3. 0,5 mm jet preturbo, 0,3 mm jet post-IC. Result: max air consumption 791 kg/h, ignition retards up to 9 degrees in 1st run and up to 12 degrees in the second.

In twin jets setups (2 and 3) the 15psi checkvalve has been attached to the preturbo jet water line, the line to preturbo jet from the tee is also much longer. I verified both jets were flowing normally (controller in the test mode) but would expect in real life preturbo jet starts spraying with a delay compared to post-IC.

Based on above I may say that no turbo efficiency gain have been seen yet, however the detrimental effect on the knock suppression from shifting the injection to preturbo is pretty obvious. It would seem logical to increase the injection post-IC to fight the second (e.g. 0.5mm jet pre-turbo and 0.5mm or 0.6mm post-IC), I will further test this.

I have also encountered that the flow on the twin-jet setup was strange. After a periods of low load (car running but no water injection) when the WMI kicks in I get high flow for a second or so (up to 5 or 6 bars on the gauge), then it gets down to 3 bars and keeps like this to the redline. With a single 0.9mm jet I had generally linear flow increase up to 7 bars with the same SC setting. I will have a look into this as this may obviously affect the results.

[leman_opc]

Just a couple of thoughts why I may not be seing any gains in air consumption:
1. My flow issues and hence simply not enough water injected pre-turbo.
2. I was looking at max air consumption but not at the whole curve. It may have been that the air consuption gains are there but not at the peak point.
3. The whole concept is around the compressor efficiency gain, but it has been correctly noted that the energy consumed by the water evaporation in the compressor is still produced by the turbine part, hence the efficiency gain of the turbo in whole is debatable. Based on above logics it may occur that if the compressor part is a restriction for more flow then preturbo injection may have better results compared to the case where the exhaust mainofld is restrictive, in the latter case the effect may be only minor (if any).

The first 2 points may be remedied. I'll double check.

The last item is hard to verify but I believe the high exhaust manifold backpressure could be my case.

[Liborek]

The whole concept is around the compressor efficiency gain, but it has been correctly noted that the energy consumed by the water evaporation in the compressor is still produced by the turbine part, hence the efficiency gain of the turbo in whole is debatable.This assumption would apply in post turbo setup - if compressor works at certain mass flow and PR with correspoding efficiency, introducing injectant after it can only decrease manifold pressure for given flow rate, or increase mass flow for given pressure.

Pre-turbo injection shifts actual compression process from adiabatic closer to isothermic, and thermodynamics says that later process requires less energy.

To actually increase peak flow, you would have to increase up-stream pressure, or decrease up-stream temperature, and considering usual ambient temperatures, one can't achieve significant temperature drop in pre-turbo location without methanol.

Otherwise you can read-up papers on water injected turbines, IIRC they don't have too much increase in peak flow conditions, but turbine work requirement for given mass flow is decreased by very substantial amount which results in net increase in power and efficiency.

Lotus did development work of pre-compressor fuel injection (gasoline and E85) on supercharged 1.8 4 cylinder and they were able to increase BMEP due to charge cooling effect and also net power, torque and efficiency level were increased as supercharger consumed less crank power.

Thing is, that reduced compressor work on turbocharger, can be observed as reduced backpressure, but if the engine isn't much sensitive to it, or compressor doesn't work in some boundary condition to begin with, or both, not much gains in actual power can be observed.

There is actually test of pre-turbo vs. post-IC injection over rx7club, pre-turbo setup made about 25 rwhp more at peak power (500HP level), without significant change in AFR or injection pulse width - essentially same air mass flow. So whole power increase has to come from increase in overall efficiency through reduced pumping losses against turbine.

[leman_opc]

Pre-turbo injection shifts actual compression process from adiabatic closer to isothermic, and thermodynamics says that later process requires less energy.

The isothermic compression process requires less energy, but the energy spent on water evaporation is on top of that, correct? At least that's what I got from the last pages of this thread...

Anyway, I believe you are saying the actual case is opposite to what was my guess (i.e. there should be no major peak flow increase, but less energy spent on the compression process -> less backpressure, and hence gains if the backpressure was an issue, and maybe no gains if the backpressure was low to begin with)?

[Liborek]

The isothermic compression process requires less energy, but the energy spent on water evaporation is on top of that, correct? At least that's what I got from the last pages of this thread...I didn't thoroughly read last pages but no, you got it other way around.
Energy required for evaporation of water, or any other liquid for that matter, is drawn in form of heat from the actual process. That heat is free. Water simply absorbs heat of compression and according to ideal gas law, keeping volume constant (lets consider one running condition) while decreasing temperature will decrease pressure, so compressor will internally work with lower pressure ratio and hence, lower power requirement. One has to account for increased mass of water introduced into compressor and certain change in specific heat ratio for mixture of air and injectant vs. air alone. Some of these factors increase turbine power requirement, some decrease, but in every case I have seen, sum of these factors always decreases turbine power requirement.

If it wasn't the case, all the proved examples from the real world wouldn't work in the way they do :D

Anyway, I believe you are saying the actual case is opposite to what was my guess (i.e. there should be no major peak flow increase, but less energy spent on the compression process -> less backpressure, and hence gains if the backpressure was an issue, and maybe no gains if the backpressure was low to begin with)?It depends if the backpressure is high due to restrictive hotside, or if it "has to be" high to supply compressor power requirement. In later case, we can imagine overspeeding compressor at peak flow working with poor efficiency. Introduction of water would decrease work requirement so backpressure would be lower.

If compressor works with good efficiency but turbine is restrictive, introduction of water won't make much difference - even if it would decrease temperature of exhaust gas, this decrease in thermal energy must be made up by mass flow or pressure, and since we would be limited by air+fuel+water mass, pressure would again increase to reach equilibrium.

Problem of proving the concept in automotive turbocharged engine lies in fact, that even if you would decrease backpressure by large amount, say 20% in absolute value, it would increase engine power by 5% and only in case of engine which is very sensitive to backpressure, like wankel rotary, and if it was high in the first place. And such increase in power can easily be in the range of error of measuring device etc. And of course, water flow rate to reach such theoretical values is quite substantial...

Nice thing about RX-7 example is the fact, that pre-turbo injected car actually runs faster on track and then, all the theory comes to an end :-)

[leman_opc]

Thanks for the detailed explanation on this.

Based on what you say I assume with preturbo injection the result may be close to the following options:
1. Boost decrease but virtually no gains in air consumption (in case of restrictive hotside being an issue), or
2. Unchanged boost but increased air consumption (in case the hotside may still flow the increased air mass)?

[Liborek]

Thanks for the detailed explanation on this.

Based on what you say I assume with preturbo injection the result may be close to the following options:
1. Boost decrease but virtually no gains in air consumption (in case of restrictive hotside being an issue), or
2. Unchanged boost but increased air consumption (in case the hotside may still flow the increased air mass)?

I assume that your setup is intercooled, correct?

Intercooler is very efficient device and it largely diminishes difference in temperature of charge air which would be caused by different compressor efficiency, so when you measure some absolute pressure and temperature in intake manifold, you know it has certain density and it doesn't matter how this state has been achieved.

So don't dwell that much on airflow number. Setup without intercooler would produce more obvious difference in power in regards to position of water injection.

Engine net brake power isn't just about combustion of air and fuel, but also about minimizing losses. Very high back-pressure not only impedes airflow, but its also direct negative work for engine. Unfortunately, I really don't know how much power could be "freed" in reducing one, some complex engine simulation could do it, we must live up with just performance testing :-)

[leman_opc]

Yes, this is an intercooled setup. True - I've been more than happy with my intake temps when injecting only post-IC, so hardly any gains here.

It's still puzzling for me that no gains in aiflow are seen due to expected backpressure drop. I follow your logics above re. minimized engine losses, so probably there are gains in power output even despite unchanged air consumption, but I still would expect at least some (probably faint, but still) air consumption increase if there is any change in engine flow characteristics when injecting pre-turbo.

Anyway, I have now my car back from the workshop, so more testing in the next days. The photos of the compressor wheel have been made also to see if there will be any wear. I will also have my car custom mapped by the end of the month so will see if any gains are seen on the dyno.

[leman_opc]

Some more testing today after I sorted my flow issues (it appears that a dual nozzle setup with a checkvalve on one nozzle only is a no-no, now I put in a checkvalve for each nozzle). Had two different runs:
1. 0,5 mm nozzle preturbo, 0,3 mm nozzle post-IC. 805 kg/h air consumption, knock retards up to 9 degrees.
2. 0,5 mm nozzle preturbo, 0,5 mm nozzle post-IC. 805 kg/h air consumption, knock retards peaking 8 degrees in 1 cylinder, but generally up to 5 degrees.

So, the major outcome is that the concept finally works, 2% increase in air consumption is not much but the pure fact is important. Furthermore it has been noted above that the gains may be higher power-wise.

Previously I obviously was not injecting enough preturbo due to flow distribution problems. Will have a more detailed look on the logs next days.

[Liborek]

So, the major outcome is that the concept finally works, 2% increase in air consumption is not much but the pure fact is important. Furthermore it has been noted above that the gains may be higher power-wise.

Previously I obviously was not injecting enough preturbo due to flow distribution problems. Will have a more detailed look on the logs next days.

I assume that airflow numbers are STP corrected?

In your case, airflow increase is not high since you already pointed out boost drop in high rpms which indicate that compressor is out of flow. Any increase is possible only via decreased temperature pre-turbo. Bigger difference would be seen with water-methanol mixture.

[leman_opc]

Since its mass air flow I'm not sure STP correction is relevant here... Anyway it's just the data I log from the MAF sensor and I'm not really aware which adjustments are there.

On the last point - I was actually expecting to see the flow capacity of the compressor expanded, this was the core idea behind the test. I'm now thinking whether a bigger jet pre-turbo is worth trying. It makes sense but on the other hand should be probably done with more advanced atomisation than a regular jet...

I'm currently injecting 40% ethanol 60% water and would really like to keep this mixture while the system is in test mode so to say (just don't want to deal with meth while I'm changing jets etc...).

[reid-o]

Instead of looking at it as percent airflow increase, you may want to also plot it as ve increase. You can compare curves rather than peak values.

[parmas]

Here is my opinion of PRE-Turbo Water injection after reading all posts:

1. Nozzle Position :

I am really surprised how in 9 years time, there is not a definitive answer on where to position the nozzle in a pre-turbo application. Some are saying directly to impeller will prevent blade damage but would result in abnormal airflow while others a few inches away or exactly on the filter will have the maximum time to atmoize and mix with Air.

What if we position the nozzle infront of the compressor wheel but instead injecting exactly the opposite way? Let's say we are activating at 10psi, the high intake flow coming from the filter will push the mist back into the compressor forming an umbrella type of mist. What do you think about it? Check attachment

2. Nozzle Size vs Head Flow:

Again after reading all posts, it seems that many of you are going insane flow levels and above recommended nozzle size. Some seen also wheel damage because of that. When we are talking pre-turbo, I agree with some of you that nozzle size should be between 10-50micron or 0.01-0.05mm. A 130cc/min nozzle is a 0.4mm while a 50micron is 16cc/min at the same pressure. The flow seems low compared to others flowing 200+ cc/min. True FOG is 30micron.

Quote : "If you are injecting pre-compressor at 2%-3% rate compared to air flow you will have about all the WI you need.

Suppose you spray at 2% air flow rate, and you have a 11.5:1 fuel air ratio at red line.

If you are flowing 65 lb/min, that is 29510 grams of air per min.
At 2% flow that means you would be injecting 590 grams / min of water.

If you have an 11.5:1 air fuel ratio than your fuel is 29510/11.5 = 2566 grams / min
Since most gasoline has a density of about .78 then that is 3289 cc/min
10% water to fuel would be 329 cc/min
15% water to fuel would be 493 cc/min

You would be spraying about 18% per min to fuel if you sprayed 2% of air by weight with a max power AFR of 11.5:1.

The beauty of using the air flow is is self corrects as you lean out the fuel.
the higher your AFR (say 12.5:1) the higher your percentage of water/fuel at a fixed water to air ratio.


3. Injection Pressure :

This is no taboo. The more the injection pressure is the finer the mist will be. A finer mist will cool the intake charge more in less time. Nowadays injection pressures above 150psi till 250psi are common.

4. Injection Activation :

Some are saying to inject at high levels of boost due to :

- Compressor wheel damage at low boost levels (under 9psi)
- Spooling is retarded due to colder EGTs

If nozzle is small (measured in microns) and pressure is high (150+psi), wheel damage should never occur.

Spooling should come earlier with pre-water injection due to thermodynamics pushing from adiabatic to isothermal having More Air per Less Turbo RPM equals to a better efficient compression.


5. PRE-Turbo vs Ambient Temps:

This was rarely mentioned but surely is something to understand that at lower ambient temperatures, Mist is more difficult to mix with Air in time.

6. Check Valve Importance

Through out the entire thread, this was an important factor to eliminate pooling. A check valve in the most vicinity of the nozzle with a high pressure rating works the best!

7. Injecting Hot/Boiling water

Hot water is easier to make it Mist while Boiling Water makes a lot of pressure on it's own. What if we can use the "FREE" heat energy from the exhaust manifold itself. This could be done by fixing a copper tube right on the exhaust runners and press the hot water into the check valve than to the nozzle.

[Richard L]

It is way beyond the technical/mechanical ability of an average user to create an near ideal system to meet the above criteria.

[HYDE16]

I've read all 50 pages, legitimately, condensing the main points to determine my own post-MAF setup and testing.

I've read here that for pre-turbo jets which can only be installed further upstream (post-MAF) and not at the compressor based on OEM turbo housings should be installed to spray against the direction of oncoming air post MAF to atomize most effectively (basically spraying at the back of the MAF housings the incoming air breaks up the fluid better).

Based on my VW GTI's engine layout and more specifically the intake layout, I've been studying the air velocity within a 90 degree bend (the bend leading to the horizontal intake pipe running across the back of the engine (from the black MAF housing to the beginning of the long straight).

VW GTI Intake layout:
https://farm6.staticflickr.com/5602/15637437570_a88c615c5c_z.jpg

This video shows in yellow the highest velocity stream which would help atomize any water meth spraying inside of this intake pipe:
https://www.youtube.com/watch?v=OvprEu-yH-A

Based on the video example alone, should I install the jet at the top of the horizontal pipe to spray against (into) the yellow stream? Or at the bottom of the pipe to spray into (with) the yellow stream?

When people place jets within the intake tract, are they accounting for flow diagrams or streams?

[parmas]

I've read all 50 pages, legitimately, condensing the main points to determine my own post-MAF setup and testing.

I've read here that for pre-turbo jets which can only be installed further upstream (post-MAF) and not at the compressor based on OEM turbo housings should be installed to spray against the direction of oncoming air post MAF to atomize most effectively (basically spraying at the back of the MAF housings the incoming air breaks up the fluid better).

*** This is a theory I thought of myself. But that depends the nozzle size / injection pressure vs water injection activation @ ? boost / Intake tube width / Compressor size.

Basically the theory will work if the nozzle size is proportional to the injection pressure and activated at compressor high speed.

The thing that you should take care is that if the water is pressurized more than the rush of air pushing it forward, water will form around the side of the charge pipe forming droplets. These are very dangerous for the compressor blades !

In few words begin with the least possible nozzle size and activate it at least above 10psi.

Based on my VW GTI's engine layout and more specifically the intake layout, I've been studying the air velocity within a 90 degree bend (the bend leading to the horizontal intake pipe running across the back of the engine (from the black MAF housing to the beginning of the long straight).

VW GTI Intake layout:
https://farm6.staticflickr.com/5602/15637437570_a88c615c5c_z.jpg

This video shows in yellow the highest velocity stream which would help atomize any water meth spraying inside of this intake pipe:
https://www.youtube.com/watch?v=OvprEu-yH-A

Based on the video example alone, should I install the jet at the top of the horizontal pipe to spray against (into) the yellow stream? Or at the bottom of the pipe to spray into (with) the yellow stream?

When people place jets within the intake tract, are they accounting for flow diagrams or streams?

Water acts like fuel in terms of velocity. It does not like cornering ! Avoid spraying into a stream that is coming into a corner as water would actually condense in that part instead of going with the family. Is the video pipe actually your intake ?

[HYDE16]

Water acts like fuel in terms of velocity. It does not like cornering ! Avoid spraying into a stream that is coming into a corner as water would actually condense in that part instead of going with the family. Is the video pipe actually your intake ?

It is technically the same, just re-oriented based on the the photo of my engine bay. The video shows airflow rising straight up then bending 90 degrees to the right.

In the photo above the air flow runs towards the back of the engine/firewall then bends 90 degree left. I would like to install a small .3mm jet on the inside of the bend, right after the bend at the beginning of the long straight away running to the 2nd shorter 90 degree that runs down to the turbo inlet on passenger side of the car.

This location should allow me to spray into the air velocity exiting the 90 degree bend and flowing through the straight away for atomization.

[parmas]

It is technically the same, just re-oriented based on the the photo of my engine bay. The video shows airflow rising straight up then bending 90 degrees to the right.

In the photo above the air flow runs towards the back of the engine/firewall then bends 90 degree left. I would like to install a small .3mm jet on the inside of the bend, right after the bend at the beginning of the long straight away running to the 2nd shorter 90 degree that runs down to the turbo inlet on passenger side of the car.

This location should allow me to spray into the air velocity exiting the 90 degree bend and flowing through the straight away for atomization.

I do not recommend sparying pre-turbo with the nozzle head pointing to the other side of the pipe even if it is at a 90 degree location. The reason is that due to high pressure of the water againist air, the water will eventually be more pressurized condensing the other side of the intake wall forming droplets.

I did this bracket (see attachment), thanking Richard for his suggestion, as I think it is the best option for water/air mix. Although some airflow will eventually be lost due to the bracket.

The Pre-turbo setup is still not tried yet due to other work but will eventually within the next few months. Nozzle size is 150cc + another 150cc post turbo

[HYDE16]

I'm stuck since the compressor is buried down within the compressor housing of the turbo inlet. It's either test the smallest jet in some location or I can't test pre-turbo at all.

[parmas]

I'm stuck since the compressor is buried down within the compressor housing of the turbo inlet. It's either test the smallest jet in some location or I can't test pre-turbo at all.

You could try to make a bracket similar to mine where the arrow is marking. I do not suggest to go further near the filter as the reaction of water injection could be lost eventually.

You can connect the bracket by using flexible hose pipes from one end to the other

[HYDE16]

You could try to make a bracket similar to mine where the arrow is marking. I do not suggest to go further near the filter as the reaction of water injection could be lost eventually.

You can connect the bracket by using flexible hose pipes from one end to the other

Thank you so much, that would be a great option, the one small issue is that where the gold pipe ends and meets the turbo inlet, there is another 90 degree bend within the turbo's compressor housing.