<... a temp sensor in the intake could provide input to the controller which could use a lookup table to determine how to bias the injectors ...>
I wasn't willing to automate things that much right out the gate, but down the road it might be feasible.

<... The biggest problem would be in selecting injectors ...>
What's wrong with using the injectors we have right now?  They currently provide all of the fuel for a given runner quite comfortably.  I wouldn't guess there would be a problem with them delivering little enough fuel.  There might be a problem with non linearity of pulse width - like the pulse width to deliver X amount of fuel through a given injector is 57 "ticks", but due to open/close times, the A and B times might be 12 and 49 for one condition and 31 and 29 for another.  But that could be worked out with the internal maps I suppose.  In any event, the system should fail over to one injector or the other.

Sounds more doable every day ... Jim S.

echristley@nc.rr.com wrote:

 
This shouldn't be extremely difficult to incorporate into a injector contoller, assuming the processor has enough bandwidth.  .... The biggest problem would be in selecting injectors.  On a very hot day, the computer might determine that all the fuel needed to come from the throttle body injectors. ....Either way, it may mean a sluggish plane when you're not expecting it.