MSNSPNP: 94 S/C Miata reset issue?
 

Friend finally installed his PNP and came over to have me help him tune it. Wired open the actuator for the supercharger to get some base tinkering done. We went for our first drive after getting his idle rock solid, and every once in a while Megasquirt resets. The gauges in megatune go blank, the reset indicator increases, the car stalls for 1/2 a second and the tach drops to 0.

Put the factory ECU back in and drove it for over 2 hours without any problems in drive ability. Doesn't make any difference what you're doing when it happens w/MS installed. Turning, going straight, bump, smooth surface.. Its not a problem with the car, I'm 99% sure. Anyone else ever had this problem with a PNP? Going to try and reflash DIY's firmware and see if that helps. Anyone??

Never had this problem, but perhaps it's a grounding issue? The ECU grounds for the '94 are underneath the throttle body. Check and clean them and see if anything changes.

That was oddly the VERY first thing I checked. Gonna try and clean it though and see what happens. Noticed he had the ground in trunk by the battery disconnected, so we fixed that already. The engine is grounded to chassis with a 4 gauge battery cable.

Hes on his way back over again, we'll clean those grounds and post results. *cross fingers*

Had him clean the crap out of the grounding post and ring connector for the grounds below throttle body, and make sure they were good and snug. After that was complete I went ahead and flashed the firmware into MS again.

Started the car and drove it for 20 minutes before it reset again. After 2 hours of driving it only reset 3 times total. Still kind of bugging me, it shouldn't be doing that.

We've got some other issues to look into; Throttle won't indicate < 10% or > 77% closed/open...

Going to have him make a rather large ground wire to attach from the ECU grounding spot under TB to the chassis and see if that helps the resetting..

The throttle just needs to be recalibrated in software. It's under the rightmost drop-down menu, tuning I think?

Resets can be tricky, and you should be getting none at all. Never. Nada.

The only time I ever got actual hard resets was when I stupidly had a relay coil paralleled with an input to the MS, every time the relay de-energized, the spike caused reset- dumb wiring mistake.

DIY sometimes suggests that a large (3,300 or 4,700uf) cap be placed on the main MS board between +12 and GND. Specifically, +12 is taken at S12C or the striped side of D10.

Beyond that, grounding. I ran two ground wires direct to the head, in addition to the ones on the ECU harness that I re-used.

Yeah I tried recalibrating the TPS three times, still won't close. Since I'm leaving town again and won't be able to help him anymore I left instructions to remove the throttle body inlet elbow and have someone manipulate the throttle, after he backs out the idle screw. Make sure it closes all the way and opens 100%. Since the TB has been moved over to the inlet of the supercharger I thought there might be something sticky or off whack with the throttle cable / linkages. He was planning on taking this off anyway to add a tiny meth injection nozzle to the supercharger inlet in hopes of cooling the bearings / charger / charge off? I Don't think its worth the effort, but we'll see. Also told him to install the IAT sensor at the manifold inlet position before we start tuning for boost.

Other thing that enters my mind.. Just made him replace the 'extension' harness for the TPS which was compromised by the cam gears causing it to run funky.. So maybe the TPS shorted out or is bad? If he can't find anything wrong with mechanical aspect of it, I've got a spare TPS to try.

We're making progress on the reset issues though, thankfully. Once he gets that beefcake ground installed we'll see what happens. I didn't think what we did would've solved the problem.

Should the PNP unit be grounded to the chassis behind the seat well, or is it okay to just mount it on the painted surface? For my boomslang style harness I broke out grounds incoming from the cars harness straight to ground, then coming from megasquirt straight to ground.. Plus added 3 - 5 others under the hood. Had too many problems in the past with cars and grounding issues... Figured I'd cover the bases.

my PNP used to just sit on the carpet and never had a resetting issue.

Mine is just setting in there too. I'm going to velcro it to the floor. I have never had a reset issue.

Just a thought. I had all types of crazy things going on when I was using my Vista computer. I'm sure your not using it. But if so, try a different computer. As soon as I switched all the weird qirks fixed on mine.

Could you post a data log with some of the resets? They may have a couple of clues about what's causing them.

Also, it is best if you don't ground the MSPNP case at all.

Hmm. Even if the TPS is physically jammed in such a way that it won't fully open or close, you should still be able to get it to read 0-100% in MegaTune. That's the whole point of the calibration process- to scale the input appropriately, taking into consideration that a TPS will rarely if ever be built and installed in such a way as to actually provide an exact 0-5v output. They're usually designed such that the throttle motion does not actually cover 100% of the potentiometer's range.
If the "wiper" lead was shorted out, then it is possible the wiper and element could have been damaged by overcurrent when the wiper moved to the +5 side of the element. This would tend to be exemplified by a very jumpy reading, with skips and spikes, when viewed in the RealTime display.

Well I have several logs I captured from last night for some tuning and I know all three resets are captured in the logs.. The problem is, they're all HUGE HUGE logs and I don't know how to identify for sure when it reset aside from watching it playback in Megalogviewer to watch for "RUN" to switch from yes to no.. Hints?

Joe, I'm still stumped on the TPS issue. Think we'll try a new TPS and see what happens.. The realtime and logs both show smooth linear operation.. Highly confused.

Resets appear in MegaLogViewer as vertical red lines so they stand out a bit.

Ok this weekend we will test everything again, IAT will be installed and found my big mamma jamma ground hidden by piping under the heatercore hoses had rust on it from previous water leak, cleaning it off tomorrow and hopefully have a successful venture this weekend.

Cool, Thanks Matt. I'll look for it and post it up when I get home. Guess I've never had a reset logged before that I noticed.

http://www.jugrnot.com/reset.jpg

The cursor is over the reset right there so I could show what was happening at the moment of reset. Reviewed ALL fields before and after reset but I don't see anything in there that would cause it. Battery voltage is pretty steady 13.8 - 14.6 and I see no abnormal spikes or any other indicator. Anyone else see anything strange?

Thats the very second data log, so its untuned grossly at this point, I know. ;)

http://www.jugrnot.com/reset2.jpg

This one is a lot more telling...

Before the reset:

RPM: 2320
MAP: 45
CLT: 184
MAT: 98
AFR: 16.1
PW: 4.1
TP: 15
Spark: 24.8
Boost: -7.4
Battery: 14.1

During the reset:

RPM: 2366
MAP: 113
CLT: 119
MAT: 79
AFR: 15.246
PW: 6.0
TP: 30
Spark: 24.8
Boost: 2.466
Battery: 15.8

I'm puzzled.

Just as an FIY (and probably unrelated to your reset) it looks to me like you've got trouble with your crank/cam sensor input. See that huge spikey section in RPM in the first log, and the fact that max RPM was recorded as 12,xxx in the second? That strongly suggests trigger trouble.

Could you post the whole file? These are often more helpful than a screen shot.

Thanks for pointing that out, Joe. The first time we tried to start the car it wouldn't. No RPM input. Come to find out, one of the wires that were cut for an old timing controller that didn't get patched back together. Since then we've made 100% sure everything was back together correctly. Think it could just be getting noise from a butt splice behind the seat?


Matt: Crap, Sorry! I thought I posted the whole file, forgot;

http://jugrnot.com/datalog200809012044.xls

There is a single reset in the log, and it's a real one. There isn't anything to suggest its cause though.

The RPM readings look pretty steady to me - a little jitter, but not much of any spiking.