Gallons Per Hour Error?

[dcb]

nice scope,
to cut through the noise/pwming I am thinking of using those nice spikes at the start and end of the injection event with a toggle flip flop, to make it look like a square pulse to the mpguino. something like: Digital Logic - Flip-Flops

mebbe this works with the right bias:
http://www.nteinc.com/specs/700to799/pdf/nte754.pdf

[dennyt]

That's an interesting idea... just curious, how would you make sure you're measuring the injector "on" period instead of the injector "off" period?

[dcb]

good question I guess by trying it and seeing if it is a problem.

[dcb]

Hey denny, can you get a couple pulses on one screen, and note the times and voltage peaks?

Also I expect there is a pre-pulse? can you see if that section is constant with rpm/load and how long it is?

[dcb]

Here is some of the other waveforms in the wiki, if I flip yours then it is a pretty good match. Yours looks like it is positively switched (or the leads are backwards).

I'm trying to figure out if the area in red, the "peak", is a constant. As well as get some landmarks on voltage and time.

[dennyt]

Very interesting... I'll try to take more measurements tonight.

[dennyt]

OK, I got some more traces after work. Things look a bit different when we include the left side of the pulse.

Here's one injector cycle with some load on the engine. There is a ~1000 microsecond pre-firing at 14V, then a big negative spike, then the PWM cycles to keep the injector open, then a long discharge.


Here's another single cycle with no load on the engine, showing the same 1000 microsecond pre-fire, and a shorter PWM duration.


This one shows the voltages at different points of the trace. +12V for the pre-fire and the top of the PWM, and -3V (maybe zero) in the noise between PWM pulses.


Here are two firings on the same screen, showing the behavior between firings.


I've also had problems with my kill switch freezing or restarting the MPGuino once every ~10 times I use the switch. It is a big 20A toggle switch wired directly into the injector ground line. I'm not an EE (can you tell yet?), but I suspect it's sending nasty voltage spikes to the 'guino. Sometimes my tach will jump when I kill it, too. Maybe it's time for a relay, or a protection circuit, or a 9v battery power backup circuit like someone here did...

I'm enjoying this learning experience, thanks for your help!

[dcb]

maybe a small capacitor across the switch leads would help with the electro-noise. You have a prebuilt or custom jobbie?

I'll have to think about how to trigger the flipflop on that negative signal a bit, without getting too fancy. It would be much simpler if it were positive switched, any chance you got the leads reversed? The scope ground lead goes to the chassis?

[dennyt]

Hmm, it's possible I've tapped into the inj+ lead. I just looked at my wiring diagram, and there's no way to tell which injector lead is positive or negative. They both teinate at the ecu. I'll try to check the other lead with the scope tomorrow.

[dcb]

I think the waveform looks quite different on the other side of the injector.

To ensure the scope polarity is right, or isn't being inverted in software, always clip the black lead to battery negative or the chassis, ensure the red lead goes into the red hole on the meter, and tap the red lead to battery plus. The line should move up when 12 volts is applied.