The 6V pump worked with a diode on it so it would not get the inductive kickback from the injector. It seemed to pump about the right amount of water. But I had one major problem with this setup. It somehow made the car run very rich. I am thinking the motor being in parallel with the injector kept the injector open longer than usual. Either that or it confused the computer somehow. I am not sure what was causing it but using the water I was only able to get about 53MPG even if I wasn't actually pumping water. For this car that is absolutely terrible for summer driving. After unhooking the motor the mileage went back over 60 so it was doing something wrong.
I need to build a small transistor circuit to feed the injector signal to the motor without actually loading the injector circuit. A quick look around the web found this:
https://web.archive.org/web/20120320...g/diy_aid.html It looks like it would be a pretty simple thing to build. I will probably build it and try again.
I think the setup has promise and will work once I get the details figured out. I just have to get time to build that circuit and try again. But I do like the setup described by KingZiptie so I might also try that if I can find all the parts laying around for it.