DIY water injection setup

[sambb]

Well here are the pics of my water injection setup. I'll get explanations of the system up in a bit. At the moment I just want to get them posted before this post times out and I scream.

[sambb]

So the pump is a $28 "Singflow" ebay shurflo rip off. Its a diaphragm pump 5L/min, 100psi and 4.5A rated current. It has viton seals so its meth friendly but I'm just running water. The main thing is to get the highest pressure you can with the lowest flow. High flow pumps won't have the same start up punch and/or if they do they'll draw a lot more current than they are designed to when they try to pump all their flow capability through a spec.
I got this one because its metal bodied and had external pressure switch wiring. I didn't want the pump pressure drop triggered (I wanted it boost triggered) so I needed the pump to have external pressure switch wires so that I could bypass them ie I just cut the actives path through the switch and wired directly into the motor. If you don't do this, when you trigger the motor with your boost switch, the motor will start to pump but because its trying to pump through a tiny nozzle the pressure will suddenly rise switching the pump off again. But your boost switch is still telling it to run so it tries again and the same cycling will just happen over and over. The answer for this (if you buy a motor with internal pressure switch wiring that you can't get to) is to run a water bypass. Now I ended up having to run one anyway (more on that later), but it would look exactly like what I've labelled in green on the motor pic. It takes excessively jacking pressures (the kind that will switch off a pressure triggered pump) and feeds it back to the low pressure side of the pump so that the pump can run continuously. My tip - just try to buy a pump with external wiring.
As you can see the motor is mounted in the boot wheel well and is fed by the clear lines from the tank and pumps out via the pneumatic hardlines under the car, up through the guts of the driver side fender and out into the engine bay in the gap between the fender and headlight, and then directly into the solenoid.

[sambb]

So the 8mm hardlines bring pumped water (in red) into the solenoid. The solenoid is a quality $45 LPG fuel lock off valve. You can buy them off ebay. It had a rated working max pressure of 100psi (which I thought would be ok) and I chose it because its designed to lock off liquids and is also meth safe if I choose to run meth down the track. The solenoid controlled water leaves in blue, coils around a bit and then feeds through an industrial high pressure filter in the next pic. Filtered water exits in 4mm hardline into the nozzle.
The solenoid is wired in parallel with the pump so that they will come on simultaneously. But a problem I soon discovered was that you would hear the control relay click, the pump start but the actual nozzle wouldn't fire until up to a half second delay. In half a second I reckon my turbo can just about double boost so that wasn't going to cut it. What I found, was that basically the solenoid piston was being overcome by pumped backpressure. The first test fire was perfect - the second test fire was delayed. When the solenoid plunger dropped to cut flow the pumped pressure behind it would spike perhaps due to pump run on but also due to the flex in the pumps diaphrams. Then when you went to trigger the pump/solenoid again the solenoids plunger was struggling to retract as the water pressure was torqueing it laterally. I found this by disconnecting the pump and when I triggered just the solenoid it would control the water perfectly.
The ideal solution is to just get a proprietry water/meth injection solenoid but $$$ and the fact that I'm running out of time before an upcoming hillclimb made me just go for the pressure bypass method that I mentioned in posts above. I ran the pump till I knew the solenoid was bound, then cracked the bypass until I saw the hoses suddenly relax. That way I knew any residual pressure couldn't build behind the solenoid when the solenoid turned off. Nozzle works perfectly now with no perceptible change in spray start up and the pump will no doubt be less stressed. Like I said this was just a way around a slightly over powered solenoid. If you have a proprietry solenoid eg some are rated up to 300psi or are pumping at lower pressures eg 80psi pumps are common, then this will not be an issue.

[sambb]

So the nozzle is fitted into the inlet pipe about 2 or 3 inches up from the MAP sensor. I made my own bung fitting for it which is really neat on the inside of the inlet. The aquamist nozzles use a M8 X 0.75mm thread. In hindsight I should have used snow performance/ devils own/ AEM nozzles as they are 1/8th NPT thread that you can easily buy taps for. My china ebay M8 X 0.75mm tap just arrived for $8 they just take a month. The inlet is fully insulated as you can see.
The nozzle itself is an aquamist 0.4 "A" nozzle which is their smallest one. It flows 50cc at 12v at 100psi and cost $40 bucks delivered. I chose the pre throttle body position because from what I could gather it'll result in better fog distribution throughout the plenum, gives the water more time to perform its intercooling and runs no risk of plenum vacuum siphoning water out of the nozzle. The risk to the throttle body apparently comes mostly when you use meth as it eats the seals and allows the fluids into the circuitry. I've spoken to plenty of people who have run pre throttle body water for many years on many cars and had no such dramas, so we'll see I suppose.

[sambb]

The other pic just show my relay bank - intercooler water spray, water injection and intercooler fan. And the other shows the manifold that I've fitted to mount my boost pressure switches. The far one is for the water spray. It turns on at 14psi and off at 12psi. The closer of the two is the water injection boost trigger. I'm still experimenting with the best switch on point but aquamists recommendation to set it at or just below the max standard boost level that it had from factory seems to be close to best. ie if VAG engineers thought it safe below 11psi then I'm taking care of all the temps and cylinder pressure that I've added beyond that.
Logs to come...

[sambb]

So the set up issue I was referring to earlier springs from the fact that this is only a boost triggered full pressure system, being DIY and all. When the boost switch exceeds a certain threshold it will trigger the pump that will run at full pressure for the whole time thereafter. This works very well when you are hammering it. During boost build up/max torque = lower rpms, when cylinder pressures are at their peak and the injectors are at maximum duty cycle, this full flow is exactly when and where you want it. As the engine moves up its power curve the injector duty cycle will drop off but although you still have full water pressure/flow it will not oversaturated with water because your rpms are so high and the rpms can digest it. What I've found is that where this system falls down is when you are street driving and accelerate moderately from lower rpm in higher gears than you would when you are hammering it eg to overtake. In 4th or 5th it will flick to near peak boost in an instant but the throttle will be only partially open and injectors will only just be stepping out of lean cruise mode. The injection will trigger though because of the boost level and oversaturate the mixture causing bogging. My problem is that it sux when this happens, but at the same time I'm getting my biggest temp drops when I trigger early. So I'm going to trial slightly tighter plug gaps to make it a bit more quench resistant and play more with the switch point. I suspect though that finding a compromise trigger point is going to be nearly impossible and that with this kind of setup will it will basically be reserved for events, national parks and V8's next to me.
The more expensive (read stupidly expensive) systems that look at a combination of boost, fuel injector duty cycle or maybe MAF too don't have to make this trade off. They'll either know not to trigger in these lighter load/boosting situations and run full pressure only at the right time or they may even have infinitely variable control of pump pressure so that they can trigger very early yet meter the water with load much as your ecu would with fuel injectors.
But on a bang for buck basis this system which is just a DIY copy of the proprietry base systems and that has cost me $160 is working great. Its fine as an intercooling/knock prevention insurance policy that is an add on auxillary. That is, I'm not tuned for the water. If I was I'd be pulling fuel out and adding timing, but in doing so I'd get great results but also be making myself dependent on the water. If the water failed then motor goes ping ping bang. With cheap auxillary sytems like this, you can actually loose power if you spray excessively into safe riche mixtures so that's why I'm on the smallest possible jet used for auxillary intercooling purposes only.

[sambb]

can't seem to post the excel VCDS logs. anyone know how?

[sambb]

Just an update on my WI kit. I'd ordered a check valved version of my existing nozzle which just arrived today. It has a cracking pressure of 15 psi meaning that it will shut the nozzle off sooner after the solenoid closes. On the other hand it saps 15 psi of pump pressure just to get/keep the nozzle flowing, but I have plenty of pressure so that shouldn't be a problem. The non check valved nozzles spray would trail down into a mini stream and then dribble after switch off which I wasn't happy with.
So I fit the check valved version of the same nozzle size and for some reason it has a nice broad ball shaped spray pattern whereas the non check valved one was more direct, which looked better to me. It definitely cut off the trail down of the spray, it arrests it much more suddenly but there are still some loose drips that form after turn off - nothing that 'll be a problem though. Anyway, peace of mind that the MAP sensor isn't being wet down or the intercooler back filling. The new aquamist nozzles come with screw down compression fittings now which are pretty solid.
I know a few of you other guys are running direct port nozzles with Snow/USRT nozzles. Do your check valved nozzles still allow a bit of dribble after solenoid shut off?