Water Injection and EMU?
 

Hey guys, I am thinking about getting water injection for my 1991 Greddy Turboed Miata with an Emanage Ultimate and Custom Intercooler. How would I hook it up? Would I just run a single from my greddy presure sensor to the water injection kit? Where are you guys putting your tanks at or are you using the stock windshield washer fluid tank? Thanks.

put the tank in the trunck so you have more range you can either switch via the greddy or just use a pressure sensor if you want variable DC then you will need a HSV. Wich would take a modified injector signal to be effective still i wouldn't spray until atleast 6 psi depending on your IC etc. And depending on your power gaols i would go 10-12% of your totall fuel potential with WI, with about a 30% mix of methanol.

Mine is autocross only, so I am using the windshield washer tank. In a hundred mile freeway criuse, not boosting much, I used three quarters of it. For track or daily driver a bigger tank is better.

I am running the Snow Performance (colorado company) Stage 1 boost cooler set to come on at 4.5 PSI using a 60ml/min nozzle. No intercooler yet. Spraying just before the dummy throttle body on my Supercharged setup.

No dyno numbers since the install, and I will have other changes as well before I get back to the dyno, but ... the exterior of the intake manifold is much cooler FWIW.

You will want to get a high speed valve and run it off the EMU's additional injector output. Then you have fully programmable progressive WI on a 2D map. Easy breezy lemon squeazy.

And perhaps there's a way to run WI off a 3D map that I'm unaware of.

sure all you need is pressure input then you can do DC vs rpm and pressure the same as we do on MS. I am not sure the emu can do that but it would be worth looking in to.

I don't have the EMU software on this computer, but I don't recall being able to choose IAT as the load axis. However, obviously you can choose MAP, and a 2D table of MAP vs RPM is pretty damn good, and better than nothing. Start at like 7 psi or so, and use your fuel map as a base, and basically just multiply each cell by 10% to get a 1:10 ratio of water to fuel.

The EMU can do a variable DC output on only two things:
1- It can rescale a variable DC input, but not by RPM or MAP, only a fixed scale (worthless)
2- It can generate a variable AFM output based on RPM & MAP. This would work if you retain the stock AFM.

The solution might be the sub-injector map. This gives you a full RPM/MAP table and is independent of the main injector maps. The only possible hitch would be that the period of this output varies with RPM, so you'd need to verify that whatever device you're driving will operate over a fairly large range of input cycle times. The "high speed" valve might do.

that is exactly what we are talking about.Joe

Then it's a no-brainer. Connect the valve to one of the un-used injector outputs (remember, the injector output driver is a switched ground, not an active high) and in the Parameters -> I/J tab, set the first "Type" parameter for that channel to "Sub I/J" and set the second "Type" parameter to "RPM" which will time it to the crank, rather than one of the main injector channels. (I believe that the firing order on the main injectors switches to full batch fire at some point in the load tables- I've got the diagram at home.) If you chose output A or B, ensure that JP 19 or 20 are set correctly (1-2).

Then on the "Map Select" tab, enable the "Sub I/J" map.

Now open up the "Sub I/J" map and select either PSI or VE, depending on what you're comfortable with. PSI is a more commonly used unit, but the MS crowd have been trained to think in VE. Populating values into the map is going to involve a knowledge of system pressure at the pump, nozzle size, and guesswork.

You can now go to the Logger and enable Sub I/J as a logged parameter to watch it go.

If you need a switch closure to turn on the pump when approaching boost, use the Aux output map. (page 42 of the EMU operating manual) This map drives the "VTEC" output wire (pin 14) to +12 when on. Use this to operate a relay that supplies power to the pump. Set it to turn on a few PSI below boost, so it's primed and ready when you cross over and start squirting.

Sounds fairly easy. Anyone have a link to a "high speed valve" that would be suitable for this application?

dont start to early esp if you have an IC spraying below 3 psi is not good for plugs.

aquamist 806-244
expensive :mad:

WOT says DevilsOwn will have one soon.

I sort of grasp the concept of this water injection thing you guys keep talking about. Help me out on something. Do you need to regulate the flow bassed on rpm or load? Could this be done by regulating the motor rpm? Or are you PWM modulating a small solinoid on the pump's outlet? If you want to control the motor's RPM there's a few inexpencive ways to do it I can walk you thru.

With the standard progressive systems, water is injected at a fixed rate based on manifold pressure, without regard for RPM. Technically, it would be better to have the water injected at a rate based on both PSI and RPM, and the technique described above accomplishes that.

Assuming the use of a high-speed water valve, you're not really modulating anything, you're just opening the valve for a certain amount of time every revolution, just like you do with the fuel injectors.

The Devil's controller, lacking the valve, attempts to vary the RPM of the pump motor by using PWM. This works, but it's not the most elegant solution. I think the "killer app" would be to use the valve and control it with the same box (the EMU) that's controlling the injectors. The amount of water injected can be very accurately controlled relative to the fuel in this case.

One piece of data that we would need to discover is an accurate understanding of the valve's response time for the off-on and on-off transitions. What Greddy refers to as "lag time" in their EMU software.

That's what modulation is. Pulsewidth modulation (PWM). on-off-on-off... That's what injector drive is. Would the system need a dampener like a fuel pulsation dampener? I guess the lines themselfs would absorb some of it. The reason I ask is there always a better/less expencive way to design something. If it was worth my while I might look into designing something so I'm just getting a feel for what it is your after in water injection and alot of you are using it so it might be worth my while developing a better system.

From Aquamist:
Basically it's a stainless steel high z fuel injector

PWM generally refers to the use of a switched voltage to simulate an analog signal, using variable duty cycle on a fixed period. Class-D audio amps, switching power supplies, solid-state motor speed controllers and lighting dimmers, that sort of thing.

Fuel injectors drivers aren't attempting to simulate a variable voltage, they're just turning the injector on and off. So I don't really consider injector drivers to be PWM circuits any more than my annoying nephew who loves to toggle the lights in the room off and on.


Hmm.... Assuming one cycle per crankshaft revolution, that works out to 4800 RPM maximum. No good.

I guess on a 1.8 engine you could time the Sub I/J to one of the main injector channels and that would give you one cycle per two revolutions, which would drive the HSV at roughly 60Hz at redline, which would work. I really need to find that damn document that shows the injector timing on the 1.6...

Fuel injection is width bassed on load and frequency bassed on rpm. I'm not trying to argue I'm trying to help you understand what PWM is. You are correct in saying class-d amps, dc motor controllers, smps, use PWM. Fuel injection is another. If you've looked at an injector signal on a scope you can see what I mean.

http://en.wikipedia.org/wiki/Pulse-width_modulation


On the waterspray... Are you trying to find a solinoid valve that will open/close like a fuel injector once per revolution? Does the water need to be controlled as accuratly as the fuel injectors need be? Why not just use a fuel injector for the water?

bringing this back into action as id love to read more on it......

More about Lazzer & I debating the finer points of digital modulation theory? :gay: