DIYEFI.org Forum

Again, PWM is your answer, perhaps MotoFab's correct, but inflammatory comments shed a little light on it for you?

im not sure i understand how pwm would prevent from hulk smashing a valve into a piston as it begins to fall or comes to tdc. also pwm would create a lot of noise and stress on the valve seats that they were not designed for.

At ten thousand RPM you're talking 83 engine cycles per second, for a four stroke. A valve opening event is approximately sinusoidal for, worse case, a quarter of that. So multiply by four. That's 320Hz. Now, PWM at 100kHz and you can shape that wave form in real time with your valve opening desires and easily let it come to a gentle rest on the seat. Of course, to do a good job of that you'll likely need a fairly powerful processor, though I can think of a way it could be done well enough with FreeEMS. I'm not certain that I'd recommend it, though.

ok. what would be your suggestion?

I don't have one, but look to a high end ARM or something.

i ment you said,

Of course, to do a good job of that you'll likely need a fairly powerful processor, though I can think of a way it could be done well enough with FreeEMS. I'm not certain that I'd recommend it, though.

i was wondering if you would specify or pm with what your talking about since you wouldn't recommend it.

Just by firing strategic interrupts and adjusting the pulse width to limit movement and seat rate, etc. Really, you need a powerful CPU to do this.

would it be possible to have a single s12x pass the decoded wheel data to another s12x that would have the sole purpose of just driving the pwm frq for the valve train. or would it be easier to have both see the timing wheel?

im asking because on a small engine it might be possible up to 4 cyl to use a single chip. 6 may or may not be possible based on what testing determines. a v8 would be impossible.

question on the s12x, can you set clock cycle int like you can on the x86. so that a desired int happens every so many clock cycles?

and thanks for the education on pwm.

NinjaRider wrote:and thanks for the education on pwm.

You're welcome, glad we got there in the end! :-)

You can set periodic interrupts from a stack of sources.

It's hard to say without experimentation how it would function, as the obvious benefit, no throttle, would both require and hugely benefit from pre-emptive control. There are a lot of ways to accomplish this, though. I suspect that as challenging as the software would be, the hardware will be your biggest difficulty.

I'm going to have to exit this thread now, as I've got a lot of stuff to achieve both for myself, personally, and the rapidly growing and in-need user base.

Good luck! :-)

Fred.