Megasquirt "Resetting" Causing Tune Issues
 

I promise I'm not just on here to make a bunch of threads about my inability to work with megasquirt. I think I have an actual problem which has led to all my frustrations with megasquirt.

Ok so when I'm driving my car sometimes when I get into boost my AFR's richen up significantly, and I lose power. Afterwards the car runs extremely rich, even when cruising, to the point that the car will start bucking, and feels like its backfiring. However if I power cycle the ecu the issue goes away. I did this a couple times today. I drove for approximately 15 minutes then got in it in 2nd gear, when I slowed down AFR was low and car ran poorly. As I was coasting I flipped the key off and on, the car came right back on and AFR was back to normal. I was able to repeat this several times.

Doing some searching I read people had issues with grounding causing the megasquirt to "reset". However my car doesn't actually shut off it just starts to run poorly. I did notice my battery cable on the positive terminal was loose so I tightened it and went for a drive and the issue persisted. I also idled the car and went and shook wires and the ecu. This didn't have any effect.

So my question is could this be a grounding issue or have I possibly damaged the megasquirt? I did have incidents in the past where my battery came loose and contacted metal while the car was running. Unfortunately I can't check old datalogs as my tuning computer recently crashed.

Any insight or ideas would be greatly appreciated. I'm going to check electrical stuff today and try unplugging the EBC to see if that helps. Also going to put in new spark plugs.

and the log of this is where?

I went for a drive this morning after changing the spark plugs. I thought the issue was gone but it happened again. So I pulled over to hook up my laptop. I did not power cycle during this time. The issue went away and I wasn't able to get a log of it yet. I will keep trying though.

Edit: I do have one log but it's long so I'll have to figure out how to shorten it.

Put it in a ZIP file or use MegaLogViewer's trimming option.

Wow I wish I had seen this before going through loading it to excel and trimming that way. :slap:

Anyway thread resurrection! I finally went out and drove around till I got the issue to occur. It's unusual in that the issue will persist for a few minutes then go away on its on if I don't power cycle the ecu. However, power cycling immediately fixes the issue, or so it seems.

It happened today when I was driving around but this time I didn't power cycle the ecu. I drove a few more miles and after siting at a red light for about 30 seconds the issue seemed to go away.

In the log at 12273.365 seconds there is a dip in the afr's when I power cycled the ecu. It can easily be seen that the afr's went from lower to higher after power cycling.

When the car is stuck in the super rich mode low speeds (like neighborhoods) it becomes almost unbearable to drive because it starts to buck. If I give it gas it stops but if I maintain a steady throttle it will start again.

The only actual reset in the log was when you power cycled the ECU. The dip afterward is caused by the warmup enrichment being on, and possibly the wideband resetting.

I'm having trouble spotting a clear-cut issue in the log. Can you try logging and press the space bar when the issue starts? This will help spot what points to compare.

QFT.

I will do that. It's not "resetting" I guess but thats the closest I could find to this issue when I was googling.

Ok I went for a ride tonight with the laptop hooked up and got a couple instances. So when it finally happened, the first time I hit the spacebar, I immediately came to a red light and the issue seemed to go away. Then a few seconds down the road it came back, so I marked it again. Then I pulled over and unplugged my ebc, just to eliminate that as an issue. As I pulled away the issue seemed to be gone. Again, about a minute or so down the road the issue came back.

I'm not sure if this is a potential wiring issue but I think next I'll unplug the VVTuner to eliminate that as well.

Ok so here's something new. I unplugged the vvtuner to see if that was causing an issue. With it unplugged the car will not start. The vvtuner is wired in so that it shares the sensors crank/cam sensor signal and I used to be able to start it in this configuration if it was unplugged. I also unplug it when I have to put the stock ecu back in and that has never been an issue.

Edit: I just remembered I had to change a resistor pull-up on the megasquirt to get it work with the vvtuner. So I'm assuming thats probably why it won't work when it's unplugged.

Ok so I fixed a crappy ground last night and reflased the newest firmware on the megasquirt but the issue still persists. I'm wondering if something with the megasquirt unit is damaged itself or if there is some setting I'm missing in tuner studio.

You have some weird RPM oscillations that I have not seen before. Is that the problem you're referring to?
To me that would indicate an ignition problem. What's the condition of your coils?
Or of course it could be some weird VVT stuff. Can you disable VVT and see if that helps?

I too noticed that I got weird oscillations in the rpm, the main problem is that the tune suddenly goes much richer then it would when operating normally.

The ignition system is stock with one coil powering two plugs. One of the coils was replaced about 30k miles ago because I accidentally fried it when I installed the megasquirt. The spark plugs are brand new with probably 200-300 miles on them.

I can unplugged the oil control valve to for the VVT to see if that changes anything but other then that I'd have to change a resistor pull-up on the ecu to run without the VVtuner.

This may be unrelated but I also burn absurd amounts of oil, and I have high leak down results. I know this could potentially cause drivability issues but I wouldn't expect the issue to be remedied by power cycling the ecu. However, after initially experiencing the issue and power cycling the ecu, the issue seems to come back more quickly then the first time.

Could you please post your MSQ?

Certainly

Finally isolated the issue to the VVTuner. Just have to figure out why its acting up and I have some theories.

So I unplugged the OCV drove for about 30 minutes and had no issues. The car felt great. Pulled over and plugged the OCV back in and made it about 50 feet before the car started acting up.

So here is my theory. When I was initially struggling with getting the VVTuner hooked up DIY told me if I was confident that all wiring was correct to have the sensors shared to remove the R2 and R3 resistor from the VVTuner circuit board. This didn't work and I reinstalled them. Later I found out that the OPTO needed to be switched from a 12v pull up to 5v pull up on the megasquirt board. So I did that and got everything to work, or so I thought.

Over the weekend I pulled the VVTuner and checked the board and everything looked ok except that resistors R2 and R3 had a terrible solder job (I didn't soldier it, as I wasn't confident soldiering to a circuit board). So my friend and I pulled the two resistors again, which didn't help anything. So my friend, who was supposed to be better then me at soldering, tried to get them back in. After awhile they were finally in but the solder looks pretty bad. Cloudy and not shiny like its supposed to be.

So my theory is that these bad solder joints are causing the vvtuner to act up, subsequently messing with the tune. Does this sound reasonable? I'm going to get some new resistors to today to replace the ones that have been re-soldered.

If the solder is on there and making contact, which I assume it is, then that's not your problem. It's actually quite hard to screw up soldering in some resistors....
Could the oscillations be caused by improper tuning of the VVT tuner? I don't have VVT, so I can't answer this. I can't even find any VVT related settings in your MSQ.....

That's because vvt tuning is handled by another program, because its the separate controller. I only messed with it a little last year when I finally got it working. I then re-uploaded the map DIY provides that is supposed to be like stock and left it. Perhaps it was somehow corrupted? I'll have to hook up and see.

Ok so I plugged into the VVtuner, reloaded the base map. Started fine. Selected NB passthrough and it started to make cam changes to hit target, fine. I go for a drive, after about 10 minutes the cam advance is just freaking out, doing whatever it pleases and the OCV duty cycle is bouncing all over the place. So I unplug the OCV again but the duty cycle and cam advance are still all over the place, its just not driving like crap. So I drive for about 10 more minutes with the OCV unplugged before the cam advance finally drops back, and stays at 4 degrees.

So now I'm wondering if the vvtuner is damaged or if its somehow getting noise in the signal? I don't know why it would run fine then have the cam advance freak out. I did take a datalog from the VVTuner software but somehow screwed it up so I can't upload it.

The logs you posted are for RPM input troubleshooting. Do you have a regular data log, from the Data Logging menu?