When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
Auxiliary InjectionThe place to discuss topics of water injection, alky/meth injection, mixing water/alky and all of the various systems and tuning methods for it. Aux Injection is a great way to have a reliable high power rotary.
I finally purchased all necessary parts for water injection. I bought AEM pump, devil's own 2.25gal baffled tank, nylon lines, Devils own level sensor, and snow performance flow sensor. I'm injecting straight water into the cold side of the intercooler end tank via 100cc Holley solenoid nozzle. I don't want to use methanol because I can't get it reliably.
All this will be controlled by Haltech Elite 2500.
My engine (20b w/ gt42) currently has been tuned with 91 octane to 12psi without AI. She currently makes 650whp.
This is my first time going AI. I am posting because I would appreciate it if an experienced member can evaluate my initial WI plan.
I will have pump turn on 100% at anything higher than 1-psi MAP. Water injection flow will be controlled by the solenoid nozzle and/or flow sensor. My initial plan is to start opening solenoid nozzle at 4psi at 10% duty.
Here's my question. I plan on tuning the water injection by increasing solenoid duty say to 50% at 6psi or whenever it starts to misfire then pull back about 20% and build my curve that way. or is there an established curve I can use for pure WI?
I read around and it looks like I need about 500cc/min WI for 650whp.
What do you think of creating a WI delivery curve starting ~100cc/min at 4 psi and maxing out at 500cc/min of WI flow at 650whp based on MAP and RPM? and tune each cell with EGT reading?
I'm a fan of turning the pump on first to charge the solenoid. I use a Shurflo pump that has an internal bypass. It turns on a 2psi and charges the line.
I use my Haltech PS1000 to control the solenoid valve with a PWM map like you are proposing.
Injection happens really fast and I initially over shoot the desired flow then flow settles down to my desired rate. I'm still trying to tweak the PWM map to compensate.
I can track this using my AEM 30-3020 gauge and software that shows an injection flow rate data plot.
Without the software to capture the data, I don't know how could read a gauge under boost flying down the highway. I've tried it. All I see is the needle sweep and that's about it.
Does the Snow flow sensor come with data collection software?
You'll have to experiment with the initial duty cycle of the solenoid to get a desired flow ramp. I try to match flow rate to boost based on percentage of fuel at various boost levels. I try to stay around 10-15% water. Trial and error.
Do you mean 1000cc Holley solenoid nozzle not 100cc?
I run two smaller nozzles instead of one large nozzle. I think the smaller nozzles have better control.
I have a on/off/auto switch that primes the pump after sitting. I get in the car and hit that for a few seconds.
The same switch is wired to my boost controller. If WI switch is off, the boost controller is off and I'm running boost on the 7.5 psi waste gates springs. If on in "auto" mode then the boost controller is active and WI is active.
I think I can log the flow sensor data with the haltech. yes I meant 1000cc solenoid nozzle.
For safety I was going to use level sensor and flow sensor. I'm running closed loop boost control, so I can set the boost solenoid duty cycle to 0% whenever I see no flow or low level.
well I try this setting for now: 100cc/min @ 4psi ramping up to 500cc/min @ 13psi based on rpm & MAP. I still need to mount the water tank and run the lines. still waiting on a bulkhead fitting for the tank which I just realized I need to buy separately! thanks again for your input.
Last edited by stickmantijuana; 05-06-17 at 08:56 AM.
I think what TRRAPLN suggested would work well. You ideally want full pressure in the line before spraying into the intake path, and I could think of no better solution. I also read plenty of stories online where the intake sucks up the water in the line enough to empty meth/water tank after a short drive due to leaky check valve. I think any solenoid inline with the nozzle (as close as possible) will work. I bet some of those evap solenoid you can pick up for a few bucks on ebay will also work if you want to be a pioneer at finding an economical solution.
I think what TRRAPLN suggested would work well. You ideally want full pressure in the line before spraying into the intake path, and I could think of no better solution. I also read plenty of stories online where the intake sucks up the water in the line enough to empty meth/water tank after a short drive due to leaky check valve. I think any solenoid inline with the nozzle (as close as possible) will work. I bet some of those evap solenoid you can pick up for a few bucks on ebay will also work if you want to be a pioneer at finding an economical solution.
Unfortunately an evap solenoid would not work. They are designed to open during vacuum and fail when fuel runs through them. The plunger has a rubber seal, so they fail from liquids. They also are not designed for 150psi. Only 30" hg max.
Unfortunately an evap solenoid would not work. They are designed to open during vacuum and fail when fuel runs through them. The plunger has a rubber seal, so they fail from liquids. They also are not designed for 150psi. Only 30" hg max.
Both have Viton Seals and should be happy with Distilled Water. The way I'm thinking though, the lubrication of the Diesel Fuel Solenoids isn't going to be there with water (same as why a fuel pump doesn't do well with water), and it'll result in premature death due to overheating.
In fact, now that I think about it, what's stopping us from using a standard 3 Port MAC Solenoid? Doesn't it already have stainless internals and viton seals?
The diesel valve draws 2A at 12V, so it would probably burn out a driver over time. It probably is not capable of rapid duty cycle, so it would be an ON/OFF valve.
I think the reason that WI valves are so darn expensive is because of their performance under high pressure and duty cycle. If a standard fuel injector would work, that would be the way to go for cost, but they don't like water.
I looked up MAC valves and there has been seemingly no attempts on that front, probably because some MAC valves are not internally sealed and allow the air to pass over the coil for cooling.
The diesel valve draws 2A at 12V, so it would probably burn out a driver over time. It probably is not capable of rapid duty cycle, so it would be an ON/OFF valve.
I think the reason that WI valves are so darn expensive is because of their performance under high pressure and duty cycle. If a standard fuel injector would work, that would be the way to go for cost, but they don't like water.
I looked up MAC valves and there has been seemingly no attempts on that front, probably because some MAC valves are not internally sealed and allow the air to pass over the coil for cooling.
Yes, but we must brainstorm ourselves to failure to find results we seek. Better to approach this at all angles with a fresh perspective and let the results fall in place.
The diesel valve draws 2A at 12V, so it would probably burn out a driver over time. It probably is not capable of rapid duty cycle, so it would be an ON/OFF valve.
I think the reason that WI valves are so darn expensive is because of their performance under high pressure and duty cycle. If a standard fuel injector would work, that would be the way to go for cost, but they don't like water.
I looked up MAC valves and there has been seemingly no attempts on that front, probably because some MAC valves are not internally sealed and allow the air to pass over the coil for cooling.
You know, now that you say that, the Elite 2500 has 8 injector drivers designed for 2A Hold/Constant Current and 8A Peak Current. Just treating the controls as a 7th injector seems really stupid simple... Solenoid = Injector #7, and Pump is PWM based on MAP Sensor.
Maybe, but you are relying on the pump to provide linear flow. Most new ones internally bypass, since they are designed to run at 150psi continuously. I haven't had a pump hooked up to measure amperage, so I'm unsure how much they draw. I know that an average fuel pump will draw 6-8 amps depending on fuel pressure. I get that a WI pump is not the same as a fuel pump, but I just looked up a Devil's Own 300psi and it draws 10 amps continuously. That will nuke a driver in short order. You would need a driver module to the pump, or a solid state PWM relay. Some older Jeep Grand Cherokees had PWM fan relays that could work.
Maybe, but you are relying on the pump to provide linear flow. Most new ones internally bypass, since they are designed to run at 150psi continuously. I haven't had a pump hooked up to measure amperage, so I'm unsure how much they draw. I know that an average fuel pump will draw 6-8 amps depending on fuel pressure. I get that a WI pump is not the same as a fuel pump, but I just looked up a Devil's Own 300psi and it draws 10 amps continuously. That will nuke a driver in short order. You would need a driver module to the pump, or a solid state PWM relay. Some older Jeep Grand Cherokees had PWM fan relays that could work.
This link may do exactly what you are talking about.
1) The Pump has it's own 12ga Power and Ground with a 40A Relay on it.
2) Signal Controlling The Pump comes from the ECU Digital Pulsed Outlet Wire of your choosing
3) Signal Controlling The Solenoid comes from Injector #7
4) You map Injector #7 as you would like and SEPARATELY map the Pump's PWM control based on MAP Signal.
Not sure what little fuel pumps you use, but even the very efficient Walbro 416 (e85) uses 14-18 Amps and can easily shoot over 20 Amps. The Bosch 044 is the same. I'm not sure how that amperage draw is an issue at all considering you're using a signal to a relay to the pump with it's own power source.
1) The Pump has it's own 12ga Power and Ground with a 40A Relay on it.
2) Signal Controlling The Pump comes from the ECU Digital Pulsed Outlet Wire of your choosing
3) Signal Controlling The Solenoid comes from Injector #7
4) You map Injector #7 as you would like and SEPARATELY map the Pump's PWM control based on MAP Signal.
Not sure what little fuel pumps you use, but even the very efficient Walbro 416 (e85) uses 14-18 Amps and can easily shoot over 20 Amps. The Bosch 044 is the same. I'm not sure how that amperage draw is an issue at all considering you're using a signal to a relay to the pump with it's own power source.
I work in an automotive shop on regular cars. Regular fuel pumps don't pull much over 10 amps, if ever unless failed. I use a labscope to diagnose cars almost daily, so I am familiar with how a factory fuel pump waveform should look.
For some reason I misread that you wanted to control the pump on PWM, and I totally missed the solenoid= inj #7 driver. My mistake.
I work in an automotive shop on regular cars. Regular fuel pumps don't pull much over 10 amps, if ever unless failed. I use a labscope to diagnose cars almost daily, so I am familiar with how a factory fuel pump waveform should look.
For some reason I misread that you wanted to control the pump on PWM, and I totally missed the solenoid= inj #7 driver. My mistake.
This makes your insight that much more valuable and I appreciate some realism added to this discussion.
For simplicity sake, we could completely eliminate the Water Injection Pump Digital Signal Output wire from the ECU by using a Hobb's Switch routed off our boost/vacuum block, but then you're talking vacuum reliability vs. electrical reliability. If you have the spare ECU output, it only makes sense to drive the pump's signal by that vs. a steady ON/OFF switch triggered via boost reference sensor (Hobb's Switch).
I priced out an entire build from Aquamist and here's what I came up with using my design parameters:
Reservoir (1/8" Nipple Exit) to Pump w/ 6mm Compression Fittings on IN/OUT of Pump.
6mm Line to 6mm Compression Fitting Attached to 130 micron, high pressure filter (Inlet Side).
4mm Compression Fitting on 130 micron filter Outlet to 4mm Line
4mm Line to 1300cc/min Solenoid w/ 4mm Compression Fittings for IN/OUT
4mm 600cc/min Jet with Integrated Check-Valve so it can't possibly operate under vacuum
Last edited by RGHTBrainDesign; 05-11-17 at 03:36 PM.
"For simplicity sake, we could completely eliminate the Water Injection Pump Digital Signal Output wire from the ECU by using a Hobb's Switch routed off our boost/vacuum block, but then you're talking vacuum reliability vs. electrical reliability. If you have the spare ECU output, it only makes sense to drive the pump's signal by that vs. a steady ON/OFF switch triggered via boost reference sensor (Hobb's Switch)."
This is exactly what I did to save the use of the PS1000 output. The pump turns on at 2psi via pressure switch while the PWM map starts the solenoid at 6psi.
This is exactly what I did to save the use of the PS1000 output. The pump turns on at 2psi via pressure switch while the PWM map starts the solenoid at 6psi.
That's actually a damn good idea. Do you have a separate lockout for the pump, in case you run low on water?
It has a flow meter, hi & lo Safe guard and low level tank switch plus other stuff. It shuts off boost controller as I remember if you get an alarm.
This does all the monitoring that I can keep up with.
The flow meter is sensitive to heat, but I put it in a box and insulated it. No problems.
Maybe this is just the 1st Gen (FB), but I have level sensors in nearly EVERY reservoir from "Cold Climate Starting Fluid" to BMC to Coolant Reservoir, etc. It's really easy to find an OEM solution with a low level reservoir that fits your engine bay and water injection capacity needs. It's not fancy, but it's REAL easy to wire in a dummy light (if it didn't already have one to the gauge cluster) for low fluid.
I want my cake and eat it too. Much like my girlfriend...
I feel like this might have been an issue on earlier Haltechs, but the Elite 2500 shouldn't have a problem setting it up exactly how I previously stated. As far as being responsive enough for each rotary cycle, that's asking a lot...so that's why he continues on about sequential solenoids run on a block.
I was definitely right before and Richard just misinterpreted the design intention of using an injector driver for the solenoid control. Let's be honest. Setting up that solenoid as an injector with a very simple table due to MAP and RPM to squirt for X amount of time is really all we'd be using it for. Has nothing to do with matching the injector pulse duration
or pace. It's just an elongated squirt during certain loads/rpm.
I mean, sure, if we could run it like an injector, the 10-15% water content would be stupid easy to figure out based on fuel mass injected, but the whole point of water injection is it's a dummy's game. You just size the system correctly from the get-go and run rough calculations based on system design and current injection duty.
I'll do a bit of local shopping and see if I can get the majority of the expensive AquaMist parts in the US for us cheaper, then set up a Group Buy through my company to whoever's interested.
Stickman, are you ready to make some magic over there? :P
I have yet to see a water-meth flow gauge work correctly. I own the AEM 30-3020 Water Meth Failsafe and it has never worked correctly for me. I bought a new sensor and it still isn’t reliable. It doesn’t weird **** like read 200cc/min when the pump is off and the car is idling. My friend has the $$ Aquamist flow gauge and it is just as dogshit. You’d be better off measuring the flow at different duty cycles and just commanding the duty cycle you need for the flow you want. If you want a failsafe, monitor line pressure to the nozzle with a transducer.
It’s also worth noting that if you PWM the pump directly, it is very non-linear. See attached. This is AEM’s data using their controller which PWM’s the pump with no solenoid. That’s how I controlled my AEM pump and nozzle, but using the Megasquirt ECU’s water injection features and outputs. I was able to build a 2D map for pump duty cycle vs RPM and MAP and it worked great. I wired the AEM failsafe’s 0-5V to the ECU to datalog and setup failsafes, but since it never worked right, I just converted all the injection lines to 4AN PTFE and crossed my fingers that my new AEM pump wouldn’t **** out.
05-05-17, 03:42 PM
We WILL figure this out.
