MS3: Noisy TPS Signal
 

I think this is my first post here. I'm hoping to get some more opinions on an issue I'm having. Word of warning: I'm not an electrical engineer, so terms/phrases I might use are probably incorrect. Please don't shoot me. ;)

Anyway, I've had a thread about the issue I'm having on Miata.net but I seem to just be going in circles. I know there's more MS, etc support here so I figured I'd try getting any more insight that I can.

Miata.net thread: Urgent Megasquirt/Engine help - MX-5 Miata Forum

The car: '96 with '01 engine and MS3 from Reverant.

So, I've had an issue with my car that has progressively gotten worse since I swapped an '01 engine over the winter. At first it was only at autocross. On maybe (read: at most) one run out of the 8-12 runs per event, the car would seem to lose power for a couple seconds. Throttle input wouldn't do anything. Then it would come back. At that point I figured it was a fuel issue. I replaced the fuel pump with a Walbro 190, and a new Mazda fuel sock. I also replaced the Fab9 coils with the factory '01 coils and new spark plugs. The injectors that are in the car ('01) only have 11k miles on them, just like the engine.

Not all of those changes happened at the same time. Through each change, the problem still persisted and eventually got much worse. The car would have the issue more frequently and I couldn't really get through an event without having the issue multiple times. Except now, rather than just losing power, the car would start backfiring and if I didn't get the clutch it fast enough, would stall. Cycle the ecu, and it would come back fairly normal.

At this point, I noticed the TPS signal in the ECU was extremely noisy. It would fluctuate 4 or 5% at idle, and at full throttle it seems to bounce around slightly more than that. In the calibration screen you can see the bar bouncing all over the place. This wasn't the case when I first started the car after swapping the engine and installing the MS3. It also wasn't an issue when I got the car dyno'd. It's something that has developed over time.

So in diagnosing this, I feel like I've tried everything. I've swapped TPS's with a friend. Mine worked fine in his car, his has the same issue (noisy) in my car. A friend of his had an oscilloscope and tested the 5v reference for the TPS and found it to be no where near constant.

I feel like I've tested/cleaned the grounds a million times. I've redone all my wiring connections. All the connections from the TPS connector back to the ECU connector seem good (ground wire has 0 resistance).

What else can I do?

is the TPS the only noisy signal?

what is your lag factor in software for the TPS? Should be under general settings. MS3 firmware has significantly improved code to handle input signals compared to the rest.

As far as I can tell, the TPS is the only thing causing an issue. IAT is stable. IAC works well. I originally thought the issue might have been the cam sensor, but I've used three different ones and had the same issue. Logs never show an issue with that or the crank sensor.

I will take a look at the lag factor tonight.

my only other thought is that the 5v signal in the harness from the ECU isn't solid, or you're missing an important sensor ground.

Add a ground from the TPS black/blue to the cylinder hear/intake manifold and see if that helps.

It was at 85. I tried decrementing it by 10's down to 10% where it stabilized some. Even at 20% it still bounces around .8% or so. Is it reasonable to use something that low?

I will try adding a ground to the intake manifold.


EDIT: Smoothing didn't fix it as I still can't calibrate it correctly because the signal moves around so much.

Tried two things:

Grounded only to the intake manifold, and grounded to the ecu and manifold.

No change.

So apparently it isn't actually the TPS. It has run fine with it unplugged the past 4 or 5 autocross events, then today it decided it didn't want to play - at all.

First run, I got about 15 seconds in before it started acting up. Cycled the ECU and it came back fine, but I just coasted around to the finish. Second run, started about the same time. Cycled the ECU, it started up fine again. I attempted to finish the run at a reasonable speed, but it started puking again after about 10 seconds. Gave up on my car for morning runs.

At lunch, I went and filled the tank up. In the morning there was slightly less than a 1/4 of tank, so for kicks and giggles I filled it up to see if it was by some miracle fuel related. It's not. I barely got through the first element in my afternoon run (100ft is being optimistic) and it started sputtering and backfiring again. This is after the car hadn't been run hard for nearly 3 hours, and it was barely up to temp.

Pack it up later to go home; drives fine. Do a couple pulls and wiggle it back and forth as best I could - nothing. It took less than what I did on my way home to get it to happen during an autocross run, yet nothing I can think of was different.

This is sooooo frustrating. The only thing I can think of is whatever the hell the issue is it's temperature related. It was really warm today at ~80-85*. It hasn't been that warm in quite a while. But what would be so sensitive to something as minimal as like a 7* temperature swing? Where if it's 74* it works perfectly fine, but at 81* it won't run at all?

Please take a normal datalog showing the issue. IIRC the 2 logs you sent me were "composite" logs, only showing the crank/cam sensors and nothing else.

Updating this since I actually got a log of the issue happening. It now happens within 15-30 seconds of starting the car, every time. I emailed this to Reverant on Sunday, waiting to see what he has to say.

Did you ever find a solution? I am having a noisy TPS signal also.

what type of MS?

the filtering in the MS3 firmware should greatly reduce the noise youre seeing on the TPS. If youre getting noise on MS3 firmware something is wrong (i.e., the sensor is bad, grounding issues, etc.)

I have an MS3/x

I just tried running a fresh ground wire to the head off the TPS and it didn't seem to help. I turned the TPS lag from 80 to 70 and it seemed to help a little bit.

I will take some datalogs tonight when the roads are clear and can post them up. It isn't really causing any drivabilty problems, but I would like to have it sorted out.

Here are my logs and map.

What is the typical amount of noise found in the TPS signal?

When I hold the throttle cracked open in neutral at about 3k rpm I see my tpsdot oscillate and jump around between +/- 150%/s or so. Is this normal and I just set my tpsdot threshold slightly above that to 200%. Or do most people have a much cleaner tpsdot signal at steady throttle?

This is a 1.6 DIYPNP with BMW vTPS by the way.

I'm curious on this as well. My TPS signal is the only noisy one, but it seems to react just fine.

shoudn't really be more than 10 TPSdot at idle.

I'm seeing around that at idle but when I try to hold a steady throttle position in neutral at a low RPM (like 3000 or 4000) it seems to pick up a lot of noise.

Can anyone else run a similar test and let me know what you get for your tpsdot values please.

Also, at idle if your tps calibration ADC count is on the high side, does MS interpolate anything bellow that calibration point as 0% throttle position, or does it go negative? Because if your jumping around from say 50 to 55 adc count at closed throttle but your TPS calibration is 55 ADC count at closed throttle your TPS would always read 0% making tpsdot 0% despite any noise you might have.

However, if your tps calibration is at the low end of the noise variation at 50 ADC count then you would see more noise since the tps will be jumping around between 0 and maybe 3-5% throttle position or so making your TPSdot value higher.

This would explain why I don't see much noise at idle but do see more noise at a steady part throttle TPS.

Something I think I read on another forum for your case:

Double check that the TPS mounting screws tight and the TPS is solid.

I'm just about to go mess around with my car again and look at the values etc... and see what I can find.