<snip
I have said it before..  I'm still a fan of engine driven (mechanical)
fuel pumps for the pressure part of the system.  Hi pressure electric
fuel pumps need an unreasonable amount of power to operate and demand
you either run two alternators (not a terrible hardship), carry large
batteries, or operationally demand that you land soon after the loss
of your electrical power source.

A mechanical pump system with low-pressure (low power) in-tank (or
near-tank) pumps could be designed to be returnless.  The pump's
output would be proportional to RPM, which is not ideal.  Proportional
to load would be better - a feedback system could either mechanically
or electrically regulate pump pressure by actuating a leakage past the
pump's vanes/rotor.  Use mechanical direct-drive for the brute force,
and low-power actuation for regulation.

If you are concerned about the failure of the mechanical pump,
design-in a fixed orifice "dribbler" system which will run on the
low-power in-tank pumps.  Use a mechanically actuated valve to control
whether this system is delivering fuel to the manifold directly, or
through the engine driven/regulated pump hi-pressure system.

This is if you are determined to ignore the advice/experience of Steve
Beckham on providing fuel to the rotary.


Back to your regular programming...

Matt

Matt -

Does anyone manufacture an EFI mechanical fuel pump for the rotary?
Hmmm.  I didn't know they existed.  Part numbers please, and make/
model I can scavenge them from.

Second point: EFI pumps may draw "unreasonable power", but we'll
quantify that when we test.  Also, don't forget that this system will
be PWM - the current draw will be a function of the load, eliminating
unneeded pump current just to force fuel through a mechanical
regulator and return to tanks.  Bottom line: lower power.

Your whole second paragraph sounds like a design proposal.  Are these
parts OTS?  Cheap?  Reliable?  Proven?  If so, I'd like to check it
out further.  But it still needs that EFI mechanical pump for the
rotary, and we'd need to figure out how to "regulate pump pressure by
actuating a leakage past the pump's vanes/rotor".  It's an idea ...

The dribbler system backup is out of scope, and a separate project.

Remind me: who is Steve Beckham?  I am all for learning from other's
mistakes, so send links to his advice and I'll read up on it.

-Vince Orton