Replacing the Stock WRX intercooler?
 

Would replacing my standard my00 wrx intercooler with a custom made turbo to Throttle body pipe to remove any flow resriction and pressure drops as associated with the stock intercooler and using my current Aquamist 2C to do the intercooling be a step backwards? At the moment my 2c system is only set using boost activation only which is set to come on at 10psi and im using the .8 jet and the HSV is not utilized(Maybe later when i have to more funds). I run a vf24 with a P20 exhaust housing and it's runnning 18psi and the Unichip has been tuned accordingly with water injection factored in.

Cheers Andy

If you could try injecting both prior to compressor & before throttle, that would be very insteresting.

I believe that can bring you some performence gain & also more aggressive throttle response.

What i was thinking of doing was with putting the jet close as possible the the compressor outlet i'm not keen on injecting b4 the compressor as im worried about knocking the blades around to much.So by putting the jet right next to the outlet there should be enough pipe length to fully atomize with air stream before it gets into the cyclinders so its cooled down as much as possible. But i do agree if i do it right there should be a performance gain because of more unrestricted inlet path as you say BUT if im not careful it could be dangerous and have a damaged motor due to det :mad: .. Would 2 jets be needed?

I have done some measuring using a k-type thermo couple and high accuracy fluke multimeter with a record function and measuring temps on the alloy part of the y-pipe underneath the intercooler i got as follows:
With injecting water post-compressor ( injecting before TB 0.9mm jet) and normally driving the temp sat on 46.4'C AVG (thats with a few full boost squirts, stop at traffic lights ect) then to start things i went WOT 4th gear till 6500rpm then to 5th for a short while i got a measurement of 94.1'C max value then injecting pre-compressor with 0.7mm jet again going WOT in 4th till 6500rpm the temp went to 70.1'C max value which is a good drop from the previous run but i really want no more than 55'C to negate the use of the intercooler.
The turbo also dropped right out of boost and was slow to regain full 18psi too when changing to 5th and going WOT again injecting pre compressor :? ...

I may try the 0.9 jet pre turbo and see the temp results, i could try runing both jets at the same time(0.7 pre-compressor jet & 0.9 pre-TB jet) but i dont think the Aquamist pump will have enough flow...

:Turbo is VF24 with P20 turbine housing still.
:Avg Ambient temp for the time of day during testing was 23'C
:Another fact i noticed when i stopped to refuel and got going the y-pipe temp was in the mid 50's and took at least 8-10mins of 100kmph driving to get back down to the 46'C AVG..

Not injecting pre turbo anymore i found damage to the comp blades not very good at all...

Wow, that sounds awfull!!

I was considering pre turbo injection, but reading you post makes me re-consider.

Any idea what caused it, could there have been drops instead of mist?

Janis

Well it's an aquamist jet and i got the impression that they atomised the water enough for it to be ok but when i pulled the turbo out to change to a bigger unit the blades on my vf24 looked like they were sand blasted and the very outter edges looked like they were corroded :mad: ill get pic and show u what i mean... :shock: Maybe i just ran a to bigger jet i don't know but im never injecting pre-comp again to find out...

That trully sucks.

I can sure imagine that you're not too keen on ever trying it again.

Janis

details
 

It would be interesting to hear some more details, like how far the spray nozzle was up stream and what your turn on point was. Were you spraying pre-turbo all the time or only at high RPM/boost?

larry

I'm highly interested, too.

I've been using pre-compressor injection for almost 3 months. Nothing wrong is detected from the driver's seat.

My pre-comp jet is very close to the compressor blade, maybe less than 8 inches. It's a 0.5mm jet most of the time, swapped to a 0.4 recently, turned on only at high boost.

Maybe I should take some time to check my turbo, too. It's hard to reach, though :sad:

Ill post some pics up tommorrow.. I have a new Avo silcon intake on my rex and where it comes from underneath the manifold and does the bend to goto the AFM they have a 3/4in breather port thats bunged off which is id say now at a guess 9in away from the turbo,so what i did was machine a fitting to place the jet low enough so it flush mounted the face of the jet with surrounding side wall of the pipe so the water would inject into the airstream correctly.(i'll get a pic of that 2) I was injecting at 13psi boost which on my car is 3300rpm approx depending on gear and that was with a 0.7 jet.It has only been for 3 weeks i think if it wasnt for the other turbo going on i would have never known. :shock: ..

Compressor Blades After WI

Jet placement

location of jet from the turbo

Vf24 compressor wheel before water inj

your WI
 

Your injection rate appears to be pretty close to my 4 gallon/hr nozzle that I ended up down sizing to a 3 gph ( may go to a 2 gph in the future depending on temp readings when I get my intake air temp stuff in).

Your injection location is just about identical to mine. I am wondering if in both our cases part of the problem is water dropplets forming on the intake tract wall and then skinning down the surface of the intake tube ( like water dropplets run up a wind screen at speed ).

If that is the case -- aside from reducing the injection rate one possibility would be to put a small step in the intake tract just before it reaches the turbo. This would force the large surface dropplet to break off and re-suspend in the air stream.

I would love to do some experiments with a plexiglass tube to see if that is what is happening, and what step shape would pickup the dropplet and re-introduce it into the air stream.


Just a thought.

Larry

Larry, i totally agree with you in saying that water droplets are forming on the side walls of the intake tract and that's whats causing the comp blade damage, but as too where the droplets are forming or how far away from the injection point its happening would be good to know?

Im wondering weather on gear changes and wide open throttle the water injection has a bit of run on effect and injects the mist on to the opposite side walls of the inlet tract thus causing formation of droplets for fraction of time your off the throttle between the changes?

It would be good to figure out how to stop the compressor wheel damage as i believe that injecting pre-compressor has good benifits..
Andy..

What if the water jet is placed at the center of the compressor wheel? the radial speed is at it lowest and also give the the water droplets to fan outwards rather than hitting the tips randomly.

Finding a way to inject pre-comp without blade damage will be hard but it will raise the compressors efficiency to a degree!!!

injection point
 

That has been used successfully. I've mentioned in the past a guy that injects directly onto the impeller hub. When the liquid hits the rapidly spinning impeller shaft it is instantly atomized.

NACA also did some experiments along these lines developing a drilled compressor hub that channeled the ADI fluid directly into the impeller it self.

(see NACA Memorandum Report E6C25 May 1946 --- "Hydraulic Characteristics of the NACA Injection Impeller" & NACA Memorandum Report E4L23a "Effect of the NACA Injection Impeller on the Mixture Distribution of the Wright R-3350 Engine" Available as NACA TN 1069 1946, also Report 821 "Effect of NACA Injection Impeller on Mixture Distribution of Double-Row Radial Aircraft Engine)

Unfortunately these sort of investigations ground to a halt when the transition to turbojet engines began in the late 40's early 1950's

Larry

"Drilled compressor hub", what a good idea.

If the compressor wheel is anodised or nickel plated, it will resist any corrosion.

I would like to reply to the pre-comp dilemna. You are experiencing exactly what everyone before you experienced. Impingement kills blades. A 50 micron droplet appears to be able to cause impingement. The mass of the drop hitting the blade is too large=erosion. Controlling the particle size is the answer, and I have some ideas on that. Our biggest problem is the practical limit to available pressure.

Maximize:
1. Pressure,

2. number of 'smaller" nozzles, with smaller spray patterns,
temp of fluid,

3. travel distance to the turbo, (debatable)

Minimize:

1. Viscosity, (a 50/50 mix is more viscous than water or meth alone)

2. Base nozzle choice on the SMA (sauder mean average particle diameter) of the nozzle. Get the smallest rated SMA possible. Generally, smaller nozzles have lower SMA

3. mounting nozzles on a conduit wall-big mistake, mount so as to flow out concentric to the conduit.


One way to keep any pooled water from getting to the turbo is to run the mist "uphill". Pooled water will be heavy, and will run down, just give it a place to collect.

Do not use a method of metering fluid that places a reduced duty cycle on the pump, use only the max pump pressure, higher pressure=lower SMA. One shurflo pump can deliver over 150 psi easily into smaller nozzles (flow)

Just some thoughts on how it "should" be done, to stir discussion. Pre comp injection should not be based on the same techniques used for post IC injection, different fruits. i believe it is a good idea, where extreme moderation techniques should be employed.

When you say this you mean to place the jet in the middle of the pipe correct? Going back to post comp injection, if i was to replace my intercooler with a correctly sized pipe from the turbo to the intake manifold and i placed a jet or jets which way would you face it or them? into the oncoming air stream or having face going away with it???..

To answer this, a concept should be explained. A water droplet has a strong affinity to combine with another to form a larger drop. It is seen when you put 2 drops on wax paper, sufficiently close, they will pull together. Surface tension, a science term, is what is doing this. Some covalent bond on the molecular or something...

The problem with a conduit is that it provides an environment ripe for collisions, so as to be counter to what we are trying to achieve. What we are trying to achieve is lowest particle size=maximum surface area of water. The surest way to do this is to spray a lot of water onto a solid surface, like a wall mounted nozzle spraying 90 degrees to air flow. There is a film clip I found that shows just how inneffective this is for evaporation, but very effective for recombining atomized liquid.

There are those that feel that the major cause of blade damage is from large recombined amounts of liquid streaming into the blade. Although government research has shown impingement damage from extensive use in power generating applications using small micron misting, but it must be acknowledged that these applications use injection 100% of operation, with turbines spinning 24 hours a day. We would employ it for less than 1% in a street application, only to enhance top end power and cooling. So long as siphoning is eliminated as a source of potential injury, a good injection method should show no measurable damage over the life of a vehicle. My opinion. That leaves how I would define a good method, your question. Approaching it with these things in mind, if I wanted significant evap cooling prior to the blades, I would line up an infinite number of infinitely small nozzles, spraying into the airstream. All the nozzles would be located in the center of the conduit, yes. Now back to reality, this is not practical, and further, if evaporation prior to the blades is not a goal, then more simple. Spray right onto the blades, with the airstream. There is no chance for water to recombine.

You might consider a combination of techniques. A small nozzle spraying directly into the turbo, and back-to-back nozzles on a T, spraying from the center, located as far upstream as possible. Utilize at least 100 psi, and an effective fogging type nozzle, should get mean particle size down to 50 micron.

If you can get ahold of 500-800 psi, you can have real fog of the 20 micron variety, which will damage nothing. But you have to use nozzles designed for that purpose. Most stuff on the w/i circuit is low pressure, and avg spherical sizes are well over 100 micron in reality.

as for removing an IC, not usually a good idea. You could use the same techniques. I personally think the back-to-back T nozzle arrangement is a practical arrangement that would work very well. Use nozzles that have a spray pattern of 90 degrees or less, to try to keep as much of the spray off the wall as possible.

After a lot of time and thinking and asking questions im going to remove the intercooler and go full WI as my only source of cooling, i have just ordered the right parts from an aquamist supplier to upgrade my 2C to a 2D with a few more bits and pieces as well and sometime soon i'll get the pipework made.

I figure my cars just a pure daily driver as i don't do track work, i only go to drags maybe once or twice a year if i get the time and i don't give it hell when i drive it only the odd burst here and there, and with the turbo i got which is a TD05H/20G running 18psi which is easy for it and it shouldn't be scortching the air to badly so i can remove stock restrictive intercooler and let the turbo flow/move more air and maybe get rid of some exhaust back pressure at the same time and hopefully WI should keep temps under control and the motor det free. :wink:

Andy

You are brave, man!

Good luck to you & I'm really looking forward to see your good result.
Keep updating, please.

:lol: I have faith in Aquamist products i have had the 2C on my car for about 2 years now and not one single issue but this will put it to the test and yes for sure i will update soon as the mods are done and i do some of my own logging and then the final ecu tune.. :D

Andy

Update: Piping is made just have to wait for delivery which be in 2 days time, 2c is now upgraded to a 2D and i will be running 2 * 0.5mm jets.. still have to buy a knocklink though before a full power tune is even considered and i'll keep the boost low around 1 bar...

Is Kel doing the tune Andy?

Yeah Brett he will be, not sure exactly when yet as i have a few more things i have to get sorted.

Some pics of the piping and with it fitted which is all stainless and with the 2 jets in place which are not quite where i wanted them though but it will do, and piping is about as short as it gets. Everythings bit rough but i will tidy it up :oops:

HERE

Subscribe...

I am very interested in this. On a 20g the response should be a lot better with out that huge void to fill. If this works, without knock, or blowing up your engine. I may consider this. I am running a VF22 right now. This is cool though. Good luck, and hope it works. What size nozzles you using? And pump pressure?

Thanks, so far with out tuning and running low boost 13-14psi the actual response isn't any better down low and through the midrange than before, it's only when the tacho goes past 5k that i can notice the real difference and the restriction that the stock IC presented is gone it really shifts through the revs quickly.I'm running 2 x 0.5mm jets , i'm unsure of the pressure, the pressure switch is left set at what ever came from the factory with 2c Kit but i will try and get it set to max pump pressure.. Can't wait to get the Knocklink and couple of other little things sorted and get 19psi flowing. :twisted: i can also here the boxer rumble coming through the induction side too sounds wierd.. And i can't really get a good temp measurement as the stainless piping is holding the engine bay heat so looks like i'll have to get a AIT sensor & gauge.

Andy

This is close up pics of the VF24 Compressor wheel when i had it set up injecting pre-turbo awhile ago too.

http://www.geocities.com/kombiez/yahoo8.JPG
http://www.geocities.com/kombiez/yahoo9.JPG

Andy

Very interesting :cool:

Can you tell us at what boost level your precomp injection was activated?
Was it pure water?
How many litres of water did you go through (roughly!) to get to this stage of tip erosion?


PS
I'll use these pics on my website if you don't mind, there is precious little on precomp injection out there anyway...

I had it it activated on about 10psi, as it was a fair while ago when this hapened and i had it only on for about 3 weeks and it would have only gone through 2 Liters at the most and the mix would have been rain water and Isopropyl alcohol.No probs with the pics :wink

10psi is a bit low for the usual nozzle placement (facing the airstream at 90degrees)
I wouldn't let it activate before 1 bar.
OK, boost pressure is on the other side of the intake, but it is an indication of the sort of airflow passing in front of the nozzle.
thanks

How far is the nozzle away from the compressor wheels? I wonder if things could improve if the nozzle is closer to the wheel.

The jet was placed approx 360mm from the turbo in the silicon intake pipe which is a spare port that normally would be blocked off.



http://www.geocities.com/kombiez/Intake2.JPG
http://www.geocities.com/kombiez/Intake1.JPG

UPDATE: 19/03/06

Installed the Knocklink today and set it at max sensitivity and when for a drive, what it displayed under WOT reading and the boost from the 20G just under 1 bar was no LED's up until 6000rpm, then about 6100 there was 2 greens then at 6500 both oranges then just under 7000 the big red (well a lot of reds and i backed off in a hurry then) .Car is still not tuned and i have removed the Bosch 910 pump and surge tank until i put the 044 pump in so the OEM pump at a guess is probably not up to the task with 660cc injectors and the added airflow up top now there's no restriction from IC anymore, just got to keep it under 6000 and be sensible until the pump is fitted and the car is tuned..


Andy

I was wondering if you can run your car up to 7000 with lower boost setting. The engine get noisy beyond 6000, the knock indication may not be purely knock but other noises. This will give a fairer results.

Richard

I'll try that Richard, i may even just run it too 7000 trying not to use any boost and see what it does and just see if it is just engine noise.