Voltage weirdness
 

Still chasing problems with Reverant-built MS2 that's in my 2000 Miata for 4 years with the same tune. Here's a new one this weekend:

I'm measuring 13.5V at the battery, about 15.1V at the alternator and Tuner Studio shows 18V.
The car first tries to idle at 4K RPM, goes down to 1K for a bit and then stalls.
Car runs fine with stock ECU, though also measures about 15.1V at the alternator

I'm checking the MS2 to car harness right now but it seems OK so far. What else could this be? Bad grounds or connection on the alternator maybe? Where does TunerStudio manage to find 18V?

I pray to the car gods almost constantly about not having issues with my MS3. Good luck.

The voltage sensor in Megasquirt can be calibrated. In tunerstudio, in the menus under Tools / Calibrate battery voltage. It's a bit tricky to calibrate properly (the settings you give it are a bit awkward) but as long as it's in the ballpark for readings around 12-13V it doesn't need to be overly precise. The only thing I am aware that it's used for is compensating for injector dead time, but if MS thinks it's as much as 18V it is probably "overcompensating" with a much too short pulsewidth, giving you too little fuel especially at idle.

I calibrated mine before I even installed in the car, when I had it on a variable power supply. I compared multimeter and megasquirt readings at different voltages (around 9-15V) and got it to match to within 0.1-0.2V or so, which should be enough (at default values it was about 0.7V too low). I just lowered the "voltage at max ADC" a few points until things lined up and that seemed to suffice for me.

measure your voltage at the injectors and calibrate the MS to match that.

I can try to do voltage compensation. But I'm concerned with why the MS reading and behavior changed suddenly, when it was OK for 4 years before? What could have happened and what are we masking here by adjusting the voltage reading?

Mr. Brain, you got a PM, sir!

OK, moving step by step here. Rang out the harness, everything looks OK.

The car starts and idle OK with MS2. The problem now looks consistent and repeatable. When cold, it shows about 15V on TunerStudio. After a while of running the voltage goes to 16, then closer to 18V at TunerStudio gauge.

Even if I do a voltage calibration adjustment (what value should I put in?), I think it's just masking some serious issue either with car or the MS2. Ideas or thoughts? I already cleaned up the ground point at the front top of the engine block. Cleaning out the ground point under the headlights? Where else? Or is it something internal to MS2? I mean, no way the car actually produces 18V.

Whatever value makes the voltage reading come out correctly.

If you're reading, say, 20% too high at various voltages, lower the "voltage at full adc count" by 20%. As señor Brain says, best is to measure positive at the injectors to use as your reference, since that's the voltage that matters. If that makes it show correctly at some voltages and wrong at others, you have to adjust both values, and shit gets a bit more tricky. Maths and stuff. Then a couple days later when you find the bad ground point or barely connected power wire or whatever you messed with when this happened, redo the calibration to match reality again. You could also take the MS box out and run it off a separate power supply (ideally one that is variable around 12V) to fiddle with the calibration, then just do a small adjustment once in car.

Yeah, thinking of doing a bench test at 12V. It ran fine for 4 years, so I agree, wiring or ground issue somewhere in the car. Though it seems to get worse as car warms up and I still don;t get how it sees 18V when either alternator, nor battery are above 15V.

BTW, after a while, the MS disconnect and refuses connection. Probably voltage related as well?

Is this a Black DIYPNP-enhanced unit or a Silver box unit?

This is Reverant-bult (in 2012) silver-boxed MS2 Enhanced.

By the way. If I get PnP MS2 for '00 Miata, does the same tune just translate over? I have a bad feeling. Today, after idling for a few minutes, the voltage in TS shoots toward 18V and the car dies. The weirder thing... I hear clicking from under teh dash and it almost seems like it's coming from near the MS. I know, there's nothing that can be clicking there, it must be the fuel relay, but why would fuel relay keep clicking after the car stalls (key in ON position).

If it's been running fine on the MS2 for the past 4 years until recently and it runs fine on the stock ecu right now, then it has to be components within the MS2 that've gone bad, or the wiring between the MS2 and the stock harness that's gone bad.

You'll have to do some old fashioned sleuthing to figure out what's going on with it, or pay money and send it to someone that's able to test if you don't have the time/resources to figure out what's wrong. Anything else at this point is 100% speculation.

Open the ECU and make sure the MS2 daughterboard sits tightly in the sockets.

This really sounds like your alternator voltage regulator is failing more than anything. The MS has mechanical over-voltage protection (MOV1) that will shut it down if there's voltage spikes--IIRC it clamps at 18V.

If it was something like the components bad (specifically the 50K resistor on R3 and/or 10K resistor on R6) on the voltage input circuit, the voltage could show >18v but the MS would continue to run.

When the fuelp pump clicks like that, it usually indicates you have a conflict error on the tune, it's possible your tune is corrupted and simply needs a reflash and reload of the msq. You'll have to make sure that you use a MS Labs approved firmware file for your particular ECU.

Its very weird. Even if you're reving his car up when it decides to die some relay under the dash spazzes out and the engine dies violently. Like not even a sputter.

I'm glad you pulled the harness apart. There's a whole lot of electrical tape and crimp connectors in there. Surprised there wasnt anything else wrong besides the LC1 ground not being connected.

​Did done that, seems really tight in there. Will double-check again, but it seemed fine.

Leafy, that's how the harness came to me initially. I did get pinout from Reverant (maker of this MS2 and connector), figured out every signal and re-checked every single pin from the two DB37s to the rectangular connector with a Fluke. All there, though, as expected, my wideband was disconnected. I've disabled EGO correction and will re-connect and re-check LC-1 after other things are operational.

PS. LC-1 ground is connected, The signal itself was disconnected, I briefly connected it and saw AFR change with throttle input, as expected. So that part will be an easy fix.

Aha! I do believe this is what might be happening as I do see voltage creeping to 18V on TS before the car shuts down now.

Will check those, thank you!

Interesting. Because tight before the MS shuts down and dies, TS on my laptop disconnects from the MS and the message is something to the tune of: "Project mismatch" (wish I'd record it). I've done some burns after that problem started, like TPS calibration, but that probably just an incremental write, not the whole project re-write?

So, what exactly, do I need to do to re-flash?

First you need the correct firmware file for your particular unit if yours is one of Reverant's custom firmware units. FWIW, your MS controls the alternator (unless you retrofitted an NA alt).

project mismatch happens when the laptop has values that the controller doesnt have. you can see this happen if you're autotuning and dont send the changes to the MS before disconnecting. But it could be an indication that the tune file is corrupted and doing funky stuff. It's rare that firmware get corrupted anymore, but it can happen. We saw it way more frequently with MS1.

In his ECU, the alternator is controlled by the expansion board which has nothing to do with the MS2 firmware. For all intents and purposes, you can treat this expansion board like Jason's alternator control board. If you have the correct voltage at the alternator and battery, then there's something wrong with either the 2 resistors on the V3 board (R3, R6) or with the MS2 daughterboard itself.

Will be checking the resistors next week at the shop, possibly swapping this MS2 into another Mata to try. Strange, though, that when voltage spikes the Tuner Studio disconnects with that kind of message, no?