With Direct Injection systems, no spark is necessary, because the propane gets injected through the hole that the spark plug would normally have been in, and propane is more easily combustible than diesel fuel. It's not an easy thing to work out, though. I just threw it in there for the sake of pointing out the options. The power/RPM is modulated by controlling the injector pulse duration, thus the amount of fuel injected. The system would basically mock a diesel engine in operation. See Diesel Cycle for more clarification.
A boost reference is a "second opinion" to your feedback loop, an MAF would probably be better for a primary reference, but if you have any leaks, the MAF will not register them, where the boost reference will be able to tell the ECM that there is a boost leak (compared to the flow meter), and will be able to compensate for it.
A simple suck through carb setup would probably be sufficient for starters, but you'd need a feedback carb if you wanted ECM control of the system. This is why I recommended injection. You can more precisely control the injection event to precisely meter fuel based on the sensory inputs.
__________________
"¿ʞɐǝɹɟ ɐ ǝɹ,noʎ uǝɥʍ 'ʇı ʇ,usı 'ʎlǝuol s,ʇı"
|