Progressive Water Injection and MS3 without a fast valve.

[prof315]

No questions here but rather shared info on how to use MS3/3X get progressive water injection up and running without an expensive fast water valve. I concede that the fast valve is a better way to go but hey this works!

I am using a coolingmist trunk mount kit with an Aquatec 250psi pump but this _should work_ with a Shurflo pump as well.

So what I did was set up my water injection menu using D14 to control ground on the pump relay, this provides B+ to the pump when WMI is activated. Then I use the mainboard injector driver INJ1 to PWM control ground to the pump. 78Hz works as an operating frequency but I found I actually got more linear control of the flow through the nozzle on the hi freq setting. As a final step and safeguard I installed a pressure switch just before my check valve. Coolingmist calls it a clog detection switch but I am using it to activate tableswitching to my wmi fuel, spark and afr tables.

One thing I have noticed is that open element IATs DO NOT like getting misted. My sensor performance is rapidly degrading, I am going to switch to a closed element sensor and accept the fact that heat soak will be worse.

On the bright side my first real test of the tableswitched agressive spark maps was enough of a success to scare the pee out of me . Nothing like accelerating from 60 to 130 UP a bridge, the suspension actually unloaded at the top!

Jeff

[racingmini_mtl]

That's interesting. What I'd like to know is how much current the pump draws and what sort of duty cycle you use.

Jean

[prof315]

The pump is rated at 10A at 200psi. At present I am starting with 25%DC (that is the minimum I found for good atomization) and max out at 40% DC. My 5GPH nozzle is a little large for my needs. I plan on going to a 3GPH nozzle in the near future. I have tested the system with the nozzle out of the intake tract at 100% DC with no apparent issues. (even running it for over 2 minutes continuous far longer than I would driving)

[racingmini_mtl]

Actually, 100% DC is likely not the worst case. Running at 80-90% at high frequency might be worse because the switching will produce flyback energy that will have to be absorbed by the circuit and the switching will also heat up the FET more than continuous running since the FET is not instantly saturated. So if it were me, I would do some tests at 50%, 80% and 90% DC to see if the FET warms up and if the flyback circuit warms up. There is also the fact that more noise is generated under those conditions so checking that out might be useful.

Having said that, at 10A it should be fine since this is well below the FET rating.

Jean

[prof315]

Thanks for the advice Jean! Actually a friend and I are working on developing a high current (25A+) driver board for PWM controlling WMI pumps, electric water pumps and the like. We already have a working prototype but my friend found some better FETs and Driver ICs so it's getting a redesign.

[VolvoGuy50]

Thanks for the advice Jean! Actually a friend and I are working on developing a high current (25A+) driver board for PWM controlling WMI pumps, electric water pumps and the like. We already have a working prototype but my friend found some better FETs and Driver ICs so it's getting a redesign.

Here's a quick tip for you - old computer CPU heat sink + fan = much better FET life. I noticed when I had a FET board I made to drive a large fan (10") using PWM, the FET would keep roasting itself - added a spare CPU heatsink and an 80mm fan, and it's been working ever since...that was done prob. about a year ago.

[racingmini_mtl]

A fan is a huge load so I'm not surprised about the heat.

Jean

[VolvoGuy50]

A fan is a huge load so I'm not surprised about the heat.

Jean

Well, I say "fan"....It was more-so a fan modified as a propeller. I think the fan alone would have been fine, but the added load of it being under water and trying to propel a 50 pound boat wasn't helping. I'm working on a new jet-pump drive train for it, so I hope that will cut back on the FET issues...and the 10 min. battery life.

[Matt Cramer]

That's good to know. Too bad there's no video of the testing on the bridge.

[sheek]

i was thinking of doing something similar but pulsing it from the pressure switch.

will pulsing the pump reduce the pump life?

[prof315]

No it shouldn't. Most standalone progressive contollers are doing things the same way. Aquamist is the only system I have seen that has a fast valve.

[prof315]

That's good to know. Too bad there's no video of the testing on the bridge.

It actually wasn't really intentional. Just a 4th gear run with lots of load. I was so focussed on the tach and the sensation of acceleration ( the car has always been fast but very smooth no real sensation of acceleration) I wasn't looking at the speedo. Some video would be pretty cool though.

[prof315]

As requested here's my msq with the WMI

[Dookie454]

I setup my AEM Aquamist water/meth injection pump the same way.. using INJ 2 to PWM Ground to the pump, the pump is rated at 10 amps. 12v power to the pump is relayed direct to the battery + terminal, switched on by the turning the key to RUN.

Before connecting this high load pump to the MS3, I've had basically 0% drift in any sensor readings as different loads are applied, but my loads are limited.. things like my low imp. injectors are controlled using a separate injector driver.

Keep in mind this is the first "high current" device directly attached to my MS3. During testing without the engine running, I setup the Water injection settings linked to TPS so I can test without engine running... I see the pump draw creates some sensor drift, most noticable is the TPS sensor.. about 10% drift as the pump runs at a constant ~50% flow rate rate (it seems the pump mechanism itself varies in required torque... therefore creating fluctuation in amp draw, and somehow causing my TPS reading to drift).

with engine running, I'm not sure what would happen, but I assume the same thing would occur. I cant imagine this would cause a huge problem, and may not even be noticeable unless you knew what you were looking for (in this case, fluctuation in TPS reading when TPS was held steady).

Just wondering if there's something I could modify like running another ground directly into the MS3 to control ground loads on the Injector drivers, or something like building a simple injector driver based circuit outside of the MS3 to control the Meth pump so I can try to keep my very accurate sensor readings?

[Token]

Interested in this as well. All my water/meth controllers of past have been standalone controllers that PWM the pumps to ground. I have also used a controller that can translate both an analog 0-5vdc or PWM driver (to ground) signal into a heavy duty signal for the water pump. It would be nice to ditch either controller and have the MS control the pump directly via good MS grounding and possibly modifying the circuit inside the MS, possibly with a better capacitor?

[Reverant]

The Snow Performance Stage 3 controllers actually PWM the 12V supply instead of the ground. The benefit to this is that you can ground the pump somewhere nearby and you don't have to run 2 heavy-duty wires away from the pump.

[Token]

^^ Nice indeed. The 3.0 or 3.57 board's driver can be modified to do the same I imagine. I wonder if a decent enough filter circuit could be made so that adverse noise is not an issue. On a board I had a 2nd party develop, it seemed noise for the most part was addressed simply with a capacitor.

[93white3400z]

Cooling mist is selling this now too: http://www.coolingmist.com/pagedisplay. ... Controller

It's kind of expensive but oh well, I bought one. We will see how it goes.

[Token]

Cool circuit. I'm still going to put some energy into making the MS control the meth directly, vs. an in-between box.

[sjf911]

Cooling mist is selling this now too: http://www.coolingmist.com/pagedisplay. ... Controller

It's kind of expensive but oh well, I bought one. We will see how it goes.

If I'm reading that correctly, it looks like it takes a variable frequency input to convert to duty cycle for progressive injection. Can MS3 provide a variable frequency output between 1 and 1000 hz?