Hello!

I may have mentioned before I'm working on an ECU for a small 1-cylinder 30cc injection engine.

Currently the ECU (8052 based ;-) ) is mainly responsible for adjusting the delay between the beginning of an engine cycle, and when the injector fires through the (hopefully) open valve.

Our team was kindly donated a box of tiny fuel injectors (I'm affraid the supplier and make of the injectors must remain a secret because of our competition!). These units are basically extreamly high speed spring-loaded solenoid valves. When there is no power the injector is closed, and when 12Vdc is supplied the injector opens letting the fuel (which is preasurised) flow through it.

As you cam imagine, getting the microcontroller interfaced to the injector is fairly simple. There is a problem with the injector in that it's opening time is fairly sluggish (we're still talking milliseconds here, but apparently it could be made faster). Our supplier suggested the best way to drive the injectors is to apply around 60V momentarily to slam the valve open, drop to 12V to hold it open for the required length of time, and then remove the power to close it. For some reason our supplier was a bit vague about how we could go about making a circuit to do this, and we are having real problems working out how to generate the 60V pulse from a 12V supply, and how we could generate the required waveform under control of the microcontroller. One member of the team had the idea of storing the back e.m.f. of the injector coil when the power to it is released, we could then perhaps use this as our high opening voltage. A DC-to-DC converted generating 60V from the 12V battery supply was also another idea. With this we could use two transistors to switch the coil on either at 12V or 60V, and then just get the microcontroller to the turn the transistors on at the right time. I think this is probably the way we're going to have to go, unless we can figure out how to implement the back e.m.f. method, or find a better idea.

I was just wondering if anyone has tried to do something similar to what we are doing, or if anyone has any other ideas on how we could drive the injector.

I know this is a slightly off subject, and is perhaps not that suitable for this forum, but as I said, our ECU is 8052-based, and I think being able to interface hardware to the microcontroller is just as important a subject as to how to write programs for it!

Cheers for any help you may be able to give me!

Matt.