The nozzle ring adapter has to be angled correctly for the boost and water lines to clear everything.
My goal for the water injection is to reduce ignition timing correction and intake air temperatures that occur due to heat soak. If the car is relatively cold, it will make full power. Continuous hard driving will cause heat soak leading to reduced power. Heat soak takes the form of hotter intake air temperatures (I’ve seen 60F-70F above ambient) and the whole engine running hotter. I am not trying to make more power which would require a specific tune for the water injection. If I were trying to make more power, I’d do a tune for 50/50 water and meth and I’d inject after the intercooler with a much bigger nozzle, but I want the car to be perfectly safe if there’s a failure of the injection system.
With my goal defined, I needed to determine how much water to inject. I figured a safe level of water to inject without causing any drivability issues is to only spray enough water to increase the relative humidity of the air around 30%-40%RH. Knock resistance doesn’t come from only the reduction of intake air temperature, but also from the water mass added to the air and fuel mixture in the combustion chamber. It’s the same principle as exhaust gas recirculation; extra mass is added into the combustion chamber to absorb heat to reduce the peak combustion temperatures and therefore reduce the likelihood of knock. And like EGR, water injection should reduce NOx emissions too because NOx is formed during high combustion temperatures. Back to the topic of relative humidity, modern GM ECUs have ignition timing adjustment tables based on relative humidity. The humidity where I live typically does not exceed 60%RH during the day and 85%RH at night. Factor in the IAT will be a chunk above ambient, and the water should stay a vapor in the airstream and not condense into water droplets.
Because my goal is to not have a dedicated tune for the water injection, I’m going against all the guidance out there for nozzle placement on a turbocharged engine. Every guide out there will say to spray after the intercooler. Typically, the goal is to have a fine mist with the water in liquid form enter the cylinder. Once in the cylinder, the water vaporizes absorbing a ton of heat. A dedicated tune can run more boost and ignition timing to take advantage of the cooling effect. BMW’s OEM water injection setup sprays straight into the intake manifold and their tune automatically adjusts to make more power. A dedicated tune for injection involves spraying a lot more water than I’m planning to spray. I would note that I’m only spraying water and not any meth; so, I don’t have any concerns about meth corroding the aluminum of the intercooler.
With the Stage 2.5 kit controller, I can control how much water I inject by playing around with the start and end boost pressures that define the range for the pump speed. I got a 2 gallon per hour nozzle (GPH) which is about 125 mL/min; this is a tiny nozzle compared to what most people run for 50/50 water/meth injection setups; for the boost I’m running and approximate power level, the generic nozzle chart recommends an 8 GPH nozzle. My boost runs around 1.4-1.5 bar in the midrange and 1.6 bar peak at redline at sea level. If I set the start pressure at 1.0 bar and the end pressure at 1.5 bar, the system will give me the full 125 mL/min at 1.5 bar of boost. Well, this is approximate as it depends on the how much pressure drop there is in the water flow path and also how much boost the water has to push against. Anyway, if I set my start pressure to 0.8 bar and end pressure to 2.5 bar, I will only inject about 59 mL/min at my peak boost of 1.6 bar. So, depending on how much water I want to inject up to the limit of the nozzle, I can play with these settings instead of changing the physical nozzle size.
I plugged in values for water flow for my street settings which will spray ~60 mL/min at peak power. You can see the reduction of the intake air temperature isn’t a ton and the values assume 100% of the water vaporizes. Even though I’m only spraying about 1 gram per second of water, the cooling power is about 2500 watts. That’s roughly 3.3hp of cooling power from a single gram of water getting vaporized which is why water injection is effective.
Back to my strategy of spraying the water into the airstream right at the compressor outlet. The air is hottest coming out of the turbo of course. Doing some basic calculations based on my pressure ratios and a few assumptions, you can see how much the temperature of the air increases after going through the compressor; at peak power, I’m estimating the air temperature is increased by 320F. Considering water boils at 212F and I’m increasing the air temperature far beyond that, the little bit of water I’m spraying into the hot air should vaporize completely. You can see my calculated temperature reduction is only a modest 5%-9% when spraying 1 g/s of water at my air flow rates and pressure ratios.
With the intake air temperature being 30C, the mass of water I’m spraying in will increase the relative humidity by around 30%-40% points RH. That should be little enough in most cases of my daily driving to keep the water from condensing as it goes through the intercooler. Even if a few droplets form, I’m spraying so little, I’m not concerned. Keep in mind that 60 mL/min is 1 mL/second which is 1 gram per second. Hotter air holds more water of course and during sustained track use, I’m expecting upwards of 60C intake air temperatures. Or more if I’m tracking in warmer weather. At those temperatures, 60 mL/min will only increase the relative humidity by about 10% points. For track use, I’ll increase the water injection flowrate to the full rating of the 2 GPH nozzle, or about 125 mL/min. I do need to be mindful of my water consumption rate on the track because the reservoir is only three quarters of a gallon. So, I may only have about 25 minutes worth of water if I’m flowing the full amount that the nozzle can do. I’ll start off at a lower flow rate and work my way up to make sure I don’t run out of water during a track session.
13 comments
I know you achieved your goal with just water, but wouldn’t adding a bit of meth to the mix improve things even further?
If Kheim wanted to make more power, he’d tune for 50/50 water and meth and inject after the intercooler with a much bigger nozzle, but Kheim want the car to be perfectly safe if there’s a failure of the injection system.
It could. A few reasons I’m avoiding any meth; I can buy distilled water at any local grocery store or CVS or Walgreens. At my street injection rate, the water flow rate is about 4%-5% of fuel flow rate as I calculated in the table on page 7. When I double that for track use, the water flow rate will be close to 10% of the fuel flow rate. If that were 50/50 meth, that should certainly get its own tune. I do have the reservoir inside the vehicle cabin. I don’t know how prone it is for methanol to evaporate out, but… I’d rather not find out. And lastly, I used regular plumbers tape on the water line fittings and that’s not compatible with meth.
you can also buy meth at any local hardware store… well you can in most of the country, not sure maybe california banned it or something.
I should have been more clear with my question, I meant add meth but keep it to an amount that still doesn’t require a retune… instead of a 50/50 maybe a 20/70? or even less if thats too much for no retune.
strong point on the tank being inside the cabin… I didn’t look too closely and just assumed it was in the frunk. I wouldn’t risk running any mix of meth inside the cabin either. but being inside the cabin… is the pump loud? I’m sure you don’t hear it at full tilt with the windows down at the track, but do you hear it on the street? or do you only turn it on at the track?
This project car has become one of my favorites. Thanks
Thanks! I’m putting in all of the lessons from my previous cars. I have one more round of easy mods. I’ll probably get around to doing the Numeric shifter cables per the recommendations of a reader. I would really like to do an equal length manifold (just coming to market) with a Garrett G25-550 in 0.72 A/R housing (needs custom fab/water lines/oil lines/connections, not bolt on…). The crazy dream would be the DeMan shorter gearing. I really want to do the equal length header because the stock header is really poor design. And I want the G25-550 for that ball bearing goodness. The current Raggdoll Motorsport RENNEN turbo is great for a bolt-on and I expect their equal length manifold alone would improve spool-up 300-500rpm. But gunning for what I consider the ultimate.
Yeah, we love what you are doing here, I can’t find anything to criticize! Lol.
it’s rare where someone only talks about pure water injection, and not a 50/50 (water/meth) setup being used. I love it! It should be advocated more. There are so many cases where if we were not blinded in chasing for more power, but rather improve reliability and safe power; water injection is a great way to go.
Then we also don’t need to worry about special “meth safe,” sealants for the tank or sourcing the proper meth additive that is not a de-icer that slowly gums up screens and nozzles…
I build to be able to drive as hard as I want and not worry about anything breaking! I tried searching for info on pure water injection setups and very few people do it. Of course, there’s the BMW application on their M Competition where they tune for more boost with the water. There’s a bit of chatter about the percentage of water injection vs. air which is why I calculated my values; the implication being that too much water hurts power output due to quenching the combustion. There’s also various research papers on water injection in various engines and relative humidity on the amount of timing advance that can be run.
It’s when I read about the relative humidity impact that I decided to go the direction of injecting water at the turbo discharge. I make sure all the water is vaporized which also gets me the maximum air temperature reduction. And it all made sense, the impact of the mass of water vapor (relative humidity) to reduce peak combustion temperature to reduce knock; similar mechanism to exhaust gas recirculation.
Yeah, overall goal, build the car so that I can drive hard with the least amount of worries.
you can also buy meth at any local hardware store… well you can in most of the country, not sure maybe california banned it or something.
I should have been more clear with my question, I meant add meth but keep it to an amount that still doesn’t require a retune… instead of a 50/50 maybe a 20/70? or even less if thats too much for no retune.
strong point on the tank being inside the cabin… I didn’t look too closely and just assumed it was in the frunk. I wouldn’t risk running any mix of meth inside the cabin either. but being inside the cabin… is the pump loud? I’m sure you don’t hear it at full tilt with the windows down at the track, but do you hear it on the street? or do you only turn it on at the track?
With my pump essentially set to half speed, I don’t hear it at all even with the windows all the way up. I’ll hear the water sloshing in the reservoir once in a while. If the pump is running full speed, then I can hear the pump if the windows are completely up. With the windows even cracked a bit and getting some wind noise, I can’t hear the pump at all, even at full pump speed.
This is a great read. I have an ’02 Audi TT which is no longer street legal, just autox and track with this exact same setup. I only run water and have been playing around with injection times. At the last event, intake temps were spiking to 140F in paddock and getting down to 125F on track. I played around with injection timing and landed on starting injection at 4psi and max injection at 18psi (I run up to 22psi, stock K04 turbo per SP class rules). I found that I was able to get intake temps down to 110F without any noticeable breaking up. I run race fuel and have my ignition table tuned for MBT without knock at 95F intake temps. Without water, I still do not get any knock at 145F but I am sure I am losing power due tot he air being less dense hence my reasoning for going with water injection (no meth).
One thing I’ll point out though, if you run a tiny nozzle, I keep having issues where the nozzle plugs up. I used distilled water so not exactly sure why this occurs but to prevent it I put a cap full of CLR in with the water which seems to be working. Note that CLR is really bad for aluminum so I am taking a risk but it is really hard to get to my injection location (post intercooler) but I am hopeful that at such low concentrations it is evaporating and flowing out the tailpipe.
Anyway, I just wanted to vouch for this .. it really works and when the water injection hits, it is noticeable compared to not having it on. I have a standalone ECU (MaxxECU) and run a separate VE table for when injection is on and a separate Lambda table to slightly lean out the fuel to make room for the water. My ignition table is the same with out without injection since I don’t get any knock. Interestingly, my “injection on” VE table has settled at a slightly higher number before peak torque which I think suggests a slightly more dense charge.
Thank you so much for sharing your experience! That is interesting about the nozzle still getting clogged with distilled. I don’t know if there are any regulated purity standards for calling water ‘distilled’, so, I wonder if there’s still very minute levels of impurities in the particular distilled water you’re using. You could try putting the distilled water through another heavy duty filter?
A bit of an update, the plastic reservoir that came with my kit sprung another leak. I think about half a gallon leaked out into my trunk and it took 3 days to dry out after many cycles of towels and a blow dryer. I’m done messing around with this plastic junk and I ordered an aluminum 1 gallon fuel tank that I’ll use as the coolant reservoir. Not the intended purpose of the tank, so the drain location isn’t great, but I’ll make it work. I’ll post in the next update.