[ar5boosted]

Originally Posted by toc

Simple approach is to add an injector into the intake after the airflow meter and calculate a pulse width based on current fuel injection pulse width (be it 0.2 times or less)?
Tune the injection time based on real world findings..

Is that a dumb idea?

Sounds ok to me. I've been told by the background neighsayers that the injectors will rust. However, Injectors are cheap and plentiful at the wreckers and I bought one for $3 to test how long it would last but got distracted.

Using spray-nozzles is problematic for various reasons. 'Limiting' is a better description, however that's what I used and could get it to work in a somewhat approximate system.

I based water input at a constant rpm of about 1,100 rpm on my Mitsubishi. Slightly above idle and slightly below accelerating values, but I was satisfied with that.