Replacing IC with water - Experiments
 

Well my slovenly self got off it's ass and got the car sorted for some water injection experiments.

I have a Flyin' Miata turbo and MoTeC M4 ecu installed on my Miata. I have bypassed the intercooler with a crossover pipe. The normal air temp sensor is at the throttle body and a second one is in the compressor inlet pipe to allow trimming of the water injection.

A water jet is installed at the turbo compressor outletand is driven off the boost control function of my MoTeC M4 ecu. The M4 boost control allows a direct PWM output as opposed to a PID algorithm if desired, so the jet will be directly driven off a MAP x RPM table. In addition, the M4 allows the output to be trimmed based on various factors including aux t - in this case the compressor inlet temperature. This gives the ability to map the water delivery fairly accurately to achieve a desired cooling effect across a wide range of temperatures.

The boost control moves to another auxiliary output and will be based on tps and water flow rate, allowing boost to be contrained to the mechanical base if a jet is blocked. I have a float switch to install in the water tank as well to complete the safety net.

I hope to be able to keep a fairly predictable intake temperature ceiling under boost this way, and ultimately run the full 9 psi with no intercooler. I measured my current jet's flow rate at just a bit under 300cc/min which should be enough to reduce the intake temps at 9 psi from a hairy (Vegas in August) 110F underhood to 85F. My injector duty on RC Eng 550s is about 60% max allowing a theoretical flow rate of 330 cc/min of water, so the jet is well sized.

Looking forward to seeing what can be done! 8)

Sounds like you have everything in hand. Keep us up to date. Many have been happy with dumping IC for water injection, most especially on IC systems that are inefficient either in cooling or pressure drop.

OK folks, I have been testing and made some observations. Firstly, water injection does indeed do a good job of cooling - I am seeing on the order of 50C on a given day. However, the effect is not precisely what I thought it would be, so the effect leaves something to be desired.

Anyway, first some log graphs:

http://www.laurenson.com/users/robbi/wigraph.jpg
http://www.laurenson.com/users/robbi/wigraph2.jpg

The basic issue is that I assumed that the temperature change from water injection would be felt immediately at the throttle body and then disappear at the end of injection. Kinda stands to reason, don't it?

But the thermal behavior is far more laggy than I suspected. As can be seen from the first graph I can fairly successfully keep the general AT around where the turbo intake temp is, which is good for general driving. But if you take a look at the second graph you'll see why this doesn't really work under boost.

The thermal change is laggy - as we go into boost the AT starts to rise with the first pull but despite the fact I'm injecting about 25% of fuel in water the cooling effect is not restrained to the duration of injection but rather starts a general cooling trend which itself is laggy. This means that the AT increase continues as we change gear and get back on the throttle so by the time we're in full boost after the gearchange the AT is now higher.

This explains why my initial run into boost sees no detonation but subsequent ones do, because the heat is rising all the time.

What I was hoping to see was a sharp drop in AT dueing injection and a sharp rise as soon as injection stops. This would allow me to get fine control over AT under boost which would be perfect. The current situation may be fine for on-off spurts into boost but multiple runs into boost seem problematic due to the slow response.

Now it has occurred to me that the AT sensor might be slow - I have a call in to MoTeC about that. If not, it looks like a regular IC and a det jet is about the best option for intercooling.

Any suggestions?

I'm sorry, what does "AT" stand for? Also, I couldn't get the graphs to work. I'd be interested in seeing those.

Also, I went to the track yesterday and found what you are describing. I have a 1g DSM with a 16g boosting 21psi on the stock sidemount IC. On the highway when tuning, I see no knock. Last night at the track I would have no knock in 1st, 2nd, and 3rd gears, but then when I shifted to 4th, I would see a huge spike in knock, as if I had turned off the water injection.

I think that the water injection was keeping knock at bay as long as the intercooler was still fairly efficient (fast highway tuning where lots of air was going through the IC). At the track, however, I think the IC heat-soaked quickly and then was less efficient and caused and increase in knock when it had had enough (4th gear).

I have a two stage WI with one nozzle just after the IC and one nozzle at the TB. Maybe I could keep the first nozzle spraying during drag passes so that in between shifts I would have still have some cooling.


Thanks for the info!

-Craig

Sorry - AT is Air Temperature at the throttle.

Could'nt get the jpgs to download either.., get requested url cannot be retrieved , would like to see them though.

Cheers
Ryan

Weird
 

I have the basic 1s and a single 1.0mm and run 12 psi on a compression of 10.5:1 no detonation and a/f under wot is 12.2-12.4 from 6000-7000rpm and from 4000-6000 11.8-12.1. On the 1/4 the car would pull gears hard all the way to 4th near redline 125mph(7000rpm) where I end 117mph at the trap.

There is a drop and no increase in temp from 1st gear pull to 5th gear 7000 on the highway thats 160mph :D

Another Bmw owner Boost M3 has a super charge setup running 11 psi on stock compression and with every gear pull to redline intake temp in the manifold dropped on average 50-60'F vs no water using a obd II scan tool.

There is no reason why you are getting increase temp on the higher gears. It does not make sense unless the water injection pump is losing pressure on the long gear pulls. ??

Check the voltage at the pump under load and could your pump be weak?

Another question are you guys spraying enough water at what psi are you triggering the water to come on maybe you can reduce the trigger point.

If you trigger late there wont be enough time to drop the temp steadily.

Water injection works that is a fact maybe its not setup right try and mix of methanol up to 50% to help evap the heat from the intake.

Interesting. I'm using a 0.7mm jet so that should be ok. The pump is in the trunk so that might be a factor here. Note that the battery is in the trunk as well:

http://www.laurenson.com/users/robbi/wi/wi.html

Also, the charts should be working now. Let me know if they aren't.

Very nice clean setup you have there. :)

What I have found is that on the mappable system if your duty cycle is low then you wont be spraying enough water.

At what point are you triggering the water to come on I would reduce the trigger point and test the temp again or go with a larger nozzle.

The water is there to reduce the intake temp and prevent detonation if its not doing that you are not spraying enough water.

There is another board member that did a test and got very small drop he went to a larger nozzle and got 70'F drop in temp.

http://waterinjection.info/phpBB2/viewtopic.php?t=25

Go to a larger nozzle you will know when you have too mcuh water the motor will bog down and not pull as hard and sharp.

In the summer I use the 1.0mm and in the winter I dont really drive the car but go down to .7mm-.8mm

Flyin Miata measured the cooling effects in some situations. You can see the results here http://www.flyinmiata.com/tech/water_injection.php

Generally, they say that temps drop about 90*F on their Miata with a 1s set up.

-- Chuck