RPM trigger noise w/ GPIO vr conditioner

bobnova · Post by **bobnova** » Sat Jul 22, 2006 5:29 pm

First off, i have a V2.2 board with MS1 24s9(or s11/14/somewhere in there), on a honda prelude 2.0SI (two liter).

I built the GPIO VR conditioner to read the four point cam driven VR sensor in the honda dist.

To get it to work, i had to remove C11, R13, and the opto chip, they (the opto mainly) were sucking too much voltage out of the vr signal, and bringing it below the triggering voltage for the CPU. The VR out goes directly to the cpu trigger pin.

I also had to reverse the suggested VR polarity to the vr conditioner to get it to trigger megasquirt at the correct point.

Now, my problem!

Below 2500rpm things are perfect, however above it i get enough false triggers to bounce MS's tach between 4k and 5.5k rapidly and randomly.

I put a scope on the VR input to the board, as well as it's conditioned output to the MS box (external VR board), in both places it is a nice clean signal both below and above 2500.

I'm guessing that this means i'm getting noise from inside the megasquirt box somewhere, i have one idea as to where (my wire to the CPU trigger pin goes under the jumped bypassing D8), but if i'm not right, how would i go about filtering the noise out of the signal?

Any and all bright ideas are welcomed!

EDIT:
A second issue i found thats only sort of an issue, is that the megatune guess for req_fuel is WAY off, it guesses 14ms, and the best idle is up in the 20ms range, a bit much for a 1955cc engine and 235cc injectors i'd think.

Also, it flat out will NOT run on simultanious injection, it dies imediatly if i switch off 4/alternating.
2/alt works ok, but 4 is much smoother.

snillet · Post by **snillet** » Mon Jul 24, 2006 4:02 am

I have built the VR conditioning circuit from the V3 board on a separate proto board and wired that into my V2.2 MS board...followed the schematics of the V3 board for it all the way to the processor.

It works like a charm for me.

You could compare the V3 board construction with the GPIO construction and change any differences maybe ?, i do not have it in my HEAD right now but i compared them when i built my setup and ther ARE big differences in the components values, especially the filtering.

Just a tip.

Fastest95PGT · Post by **Fastest95PGT** » Mon Jul 24, 2006 5:04 am

Maybe this would have been better posted on the GPIO forum? I've used the GPIO VR Conditioner and had no issues with spiking.

Are you using shielded wire to your VR sensor?

bobnova · Post by **bobnova** » Mon Jul 24, 2006 12:09 pm

I'm using all stock wiring, up to the ECU itself (i built a honda obd0-MS conversion harness), the stock wiring in the engine bay is shielded, the bits down by MS are not.

One thing i did realize is that C1 has a different value in the corrected easyVR circuit, mine (origional easyvr) is a .002uf, whereas it's now listed as a 470pf.

I think i have a 470pf hanging around somewhere, so i'm going to try that.

Thing is, as far as i can tell with my scope the square wave out of the easyvr board is perfect and clean, which says to me that the interfearance is coming from inside the MS case.

The V3 vr conditioner and the easyvr/gpio circuit are totally different. The V3 one is much more advanced, and contains a chip that has a bit of brain, where the easyvr one is basicly just three transistors, some resistors, and a couple caps. No brains in there.

If the cap doesn't fix it and nothing pops up here i'll probably repost in the gpio section.

Fastest95PGT · Post by **Fastest95PGT** » Mon Jul 24, 2006 12:15 pm

Be sure to let us know how it ends up. If you do repost in the GPIO forum, link us here so we can follow along.

bobnova · Post by **bobnova** » Mon Jul 24, 2006 1:22 pm

I tried it connected as labeled, and it triggers (from +4.97v to 0v or close to it) when the wave goes high, then de-triggers (back to +4.97) volts at the steep zero crossing.

Nice square wave, but megasquirt seems to trigger when the +4.97 goes to zero, giving me 51 degrees of advance at idle, and changing widely at differing rpm.

With the inputs reversed (gentle to negative, then steep to positive, then gentle back to zero), the easyvr goes low right at the steep zero crossing as it goes positive (triggering MS), then goes high again when the wave comes back to zero.

This gave me a trigger angle within 5* of where i expected, and didn't change more then maybe 1*(at the most, i'd say less really) from 1k idle to 2500rpm, when the ms started triggering like crazy.

I haven't re-routed my jumper out from under the wing diode jumper yet, i'm hoping i'll have time after work today, but who knows.

bobnova · Post by **bobnova** » Mon Jul 24, 2006 2:29 pm

Ok cool, it is working correctly then.

When the vr signal goes from - voltage to + voltage is when it pulls it low, right dead on the zero crossing heading towards positive.

Post by **jsmcortina** » Tue Jul 25, 2006 2:14 am

Lance wrote:I'm not sure how you set the trigger edge in MSnS-E, in MS-II you use the ignition capture setting.

The MS1 chip IRQ pin triggers on a falling edge only - it is not settable in software like on MS2, so it is important to build the circuit accordingly.

James

bobnova · Post by **bobnova** » Tue Jul 25, 2006 11:41 am

I'm pretty sure i figured out where the noise is coming from!

The wires are shielded nicely by honda, BUT inside the same shielding shell are the crank position sensor wires(24 teeth, cam speed), and cam sensor (1 tooth, cam speed). Remembering the slight variations in the TDC(4 point cam speed, my trigger) sensor waveform, i think it's the CKP broadcasting it's waveform.

So another question, what would be the best way to get rid of that? Right now the two wire ends of the CKP aren't connected to anything, am i likely to be better off with them connected to eachother forming a loop? Or with a cap between 'em? I really don't want to muck with the factory harness, as the ability to plug the stock ecu back in if MS isn't working right is key, so running new wires is out.

Is there any sort of simple filter i could build for the VR input to remove the 12 tooth mini-signal?

I'm experimenting with adding caps here and there on the vr board, and i've had partial success so far putting one from signal out to ground, it removed a good chunk of the small false triggers without affecting the timing of the true trigger.

Whats happening is the board triggers correctly and cleanly, going low.
At the end of the signal when it's about to go high again, it sometimes blips high-low-high-low (jagged one pixel blips, even at a 50ms scope length) before going solid high again.
The cap i put on took away most of the little blips, so that only strong ones get through.
I'm going to try increasing the size of the cap a bit, but i don't think i'll be able to get rid of all the blips that way, not when the whole circuit is triggering.