Very Rapid Resets. Its killing me (almost literally)

[fryea]

I feel like I come around this neck of the woods just to post my megasquirt woes...

So heres my problem this time. This is on a v3 board triggered off of an 004 distributor.

I get very very rapid resets on megasquirt that will end up in it dieing. Doesnt matter the gear, rpm, or speed. Say for example I'll be cruising down the freeway at 60, and ill get a reset-wait 2 seconds-reset-wait 1 second-reset-resetreset-resetresetreset-resetresetreset-resetresetreset-resetresetreset-resetresetreset. I can tell its a full reset because i can hear the main and/or fuel pump relay clicking on and off.

Usually 3 resets a second, and then eventually it will just die. THe wideband goes very lean. The IRQ and Spark LED's on megasquirt will go out as well.

Similiar situation at idle, however since the engine is turning over much slower, it dies without nearly as many resets.

The first one usually happens after 5 minutes after driving, and then after that its anywhere from 1-10 minutes between shut downs.

It can be cured by letting the car sit for about a minute(completely off), and starting it again. That allows me to drive an indeterminate amount of time.

My first guess is the hall sender, since I lose the IRQ trigger LED whilst driving, giving it a 0 on the rpm, telling megasquirt that the car is off.

However, sometimes megasquirt will do the rapid resets with the key in the on position (but engine not running) while parked. You can hear the relays go nuts.

I would really ask for your input, as it is quite scary to randomly lose power when your driving uphill and a semi truck is barreling toward you in the rear view mirror.

(cross posted at msextra)

[crazy0000]

WOW, this sounds exactly like my problem, started out with random reset here and there, then started clicking like it resets 5 times in a row, then your letting it idle in a parking lot of your school, and it shuts off and won't go back on for over a week now and MS won't put out an RPM signal!

I'm using a VR sensor though.

[Karl Buchka]

Do you have a datalog from when it does this?

[fryea]

Do you have a datalog from when it does this?

I can cook one up for ya.

Im also going to take a blow drier and compressed air can to the board, to see if it reacts to heat/cold.

[permieshane]

What is the ignition setup? I had similar problems with my first setup & ended up cutting out MS in frustration & installing a carb (lpg setup), but after I sold MS I had the urge to setup the gas injection (after turbo installation). So went down MS path again (actually cheated & spent the big $$$ on ems-pro).

To cut a long story short I found a major mistake when doing the fresh install. The main power earth for the LS2 coilpacks was ran back to the MS earth point instead of a point very close the the coils. Anyhow the setup works perfect now, not sure how much of that is due to the better power supply setup in ems-pro though.

Hope this is of some help, but interested in ignition setup anyhow.

Shane

[fryea]

To cut a long story short I found a major mistake when doing the fresh install. The main power earth for the LS2 coilpacks was ran back to the MS earth point instead of a point very close the the coils. Anyhow the setup works perfect now, not sure how much of that is due to the better power supply setup in ems-pro though.

Right now im just using a single coil with the distributor, triggered off of the onboard vb921. So the only ground for my coil is the onboard chip.


In another strange twist of events, i datalogged my entire trip to work today, and did some running up and down the gears to try and make it reset. Nothing. I even took a blow drier to it to see if it would react to heat, nothing there, and I also drove with the heater at full blast. Nothing. No resets. Im at a loss for words

It has literally reset at least 3 times per trip for the past 2 weeks, and my commute to work is 14 miles 1 way, giving plenty of time for it to manifest itself.

Since I was datalogging, the only difference is that I reburned my MSQ with megatune, and ran with the serial-usb adapter as well. Perhaps the rs232 has something to do with it? Gah, i feel like the boy who cried wolf. Phantom megasquirt ghosts. It might have something to do with halloween nearby...

[Karl Buchka]

Right now im just using a single coil with the distributor, triggered off of the onboard vb921. So the only ground for my coil is the onboard chip.


In another strange twist of events, i datalogged my entire trip to work today, and did some running up and down the gears to try and make it reset. Nothing. I even took a blow drier to it to see if it would react to heat, nothing there, and I also drove with the heater at full blast. Nothing. No resets. Im at a loss for words

It has literally reset at least 3 times per trip for the past 2 weeks, and my commute to work is 14 miles 1 way, giving plenty of time for it to manifest itself.

Since I was datalogging, the only difference is that I reburned my MSQ with megatune, and ran with the serial-usb adapter as well. Perhaps the rs232 has something to do with it? Gah, i feel like the boy who cried wolf. Phantom megasquirt ghosts. It might have something to do with halloween nearby...

See if you can elicit a reaction with the computer either plugged in or un plugged. I used to have a weird wiring problem where it would reset every time i connected the laptop.

[fryea]

See if you can elicit a reaction with the computer either plugged in or un plugged. I used to have a weird wiring problem where it would reset every time i connected the laptop.

Alrighty. I will drive back from work half way with the laptop plugged in, and drive the other half with the serial-usb converter disconnected from MS

[fryea]

THe verdict on the serial is a no go. It will reset even with the serial plugged in. I must have just had a lucky run that one day. On the way back from work today I got really bad resets, serial plugged in and datalogging. Unfortunately the datalogs don't show anything suspicious.

The most perplexing thing to me is the resets with the car on, but engine off. Example today, i pulled over on I-5 to let MS rest for a bit after some resets (stopping the car, turning off MS will allow me to travel for an indetermined amount of time). I tried keeping the car on, engine off, immediately after some resets, and I got massive *clickclickclickclick* over and over.

How often do standard automotive relays go bad? Could that be the problem? After having the relays running for a while, will they get hot and turn off and on? Is there some kind of protection built in?

I will give anyone 10$ if they can pinpoint the problem.

[crazy0000]

I will win that 10$, is your board still covered in solder flux? or is it nice and clean?

[fryea]

It is clean. It was assembled by protocar (http://www.protocar.net/catalog/index.php) in the fall of 2005.

I will go over it again just because I'm at my wits end. My one suspicion that I am looking at right now, a far as the board is concerned, is when I re-assembled the fidle circuit, my holes were somewhat mangled. I might just remove all those components again, since I am using the bosch-kjet auxiliary air valve (heated coil).

[Karl Buchka]

Are you using any of the pins on the DB37 connector for non-standard functions?

[fryea]

Are you using any of the pins on the DB37 connector for non-standard functions?

knock sense and E-fan

[Karl Buchka]

knock sense and E-fan

What pins are those functions going through? Like, are you using any of the IAC pins?

[fryea]

JS10 (knock) > IAC1A

JS3 > transistor > IAC2A

Also, I ruled out the system relay being NOT the problem. It powers the wideband as well, and that stays on while the resets happen.

[Karl Buchka]

Yeah, disconnect both those from the IAC pins. I had originally connected my boost control and fan out through a couple of unused IAC pins, and even with all those features turned off I had completely unpredictable resets. I moved the boost and fan control outputs to the two SPR (spare) pins and the problems immediately went away. It seemed to be causing some kind of weird feedback loop.

[fryea]

Huh, that is very wierd. Thanks for your input karl, I will try that.

HOWEVER... might have got it fixed via another method

So i took crazy's advise and inspected my board. No soldering flux per-se, but something else. I thought my board was relatively clean, as less than 3 months ago I scrubbed the whole thing down with 90% iso-alcy and a tooth brush. However... I did not assemble it. One suspicious area I saw was Q12, one of the transistors on the heat sink. On the reverse side of the board, it looked like some of the solder had slightly melted and were just barely a flys hair away from each other. I realise thermal expansion is quite small, but it might have been the last straw on the horses back. I got some solder removing braid, and redid the 2 suspect holes. Was able to drive to school (2 miles of stop-lights, avoiding drunk fratters) and then to work (14 miles I-5) without a reset. I was lucky that one time, so I will have to report back once I get home.

But I will probably end up moving those things off the IAC spots anyways. Ill report back when I get home.

[crazy0000]

:) 10$ please. I'm totally kidding.

[Karl Buchka]

:) 10$ please.

It clearly wasn't the solder flux. I'll take the $10. Me too.

[fryea]

I think the real culprit is sub-par solder quality :grrr:

dont trust anything! even if it is "professionally built"

The jury is still out though. Im going to drive home right now to vote out in north bend (35 mile one way). If I can make it there and back without 1 reset, I will consider it case closed.

[crazy0000]

Annnnnd?

[fryea]

fixed!

The fix to Q12 definately cleared up all the resets. I used solder-removing braid, and resoldered the points on the underside of q12. Two solder points had almost bled into each other. I have no idea how this would have happened. The odd thing is that I did not touch these since i have had the board. It was "professionally assembled". Also, this problem hasn't reared its head before in the 20,000+ miles i've put on this MS3.0 board.

[crazy0000]

We'll soldering flux does absorb water, and it becomes ionic when it does that... so slowly it could be getting water vapour and getting progressively worse? anyway glad its fixed!

[Boris740]

We are lucky that MS lead-tin solder. Imagine this horror:
http://www.aciusa.org/leadfree/leadfree_verdi-11-5-04.htm

[fryea]

GAHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH

It's back

Not as bad as before. I had 2 weeks of relative peace.

I was driving some LADIES to a bar the other night and, like the pimp I am, have my volvo die en route. Luckily we were only 2 blocks from our place but... not impressive to say the least. The best way to describe the problem without dorking out on those girls was best summed up as "volvsie doesnt want to drive anymore", and out they went.

Then it happened today while I was sitting in traffic. CLICKCLICKCLICK-DIE. "**** ME!" was the best way of describing this encounter.

So back to the drawing board. After the bar incident, i resoldered ALOT of points to make sure it is C-L-E-A-N, but still clickity click click death.

[crazy0000]

Hey funny story, mine was a champ for about 2 days, then the resets came back, pretty much everything came back.

[permieshane]

Hope you guys can sort it...Really takes me back to the days when I felt like putting a match to her....So frustrating in fact that I could justify spending the big bucks on ems-pro.

Saying that it would be good for someone not to give up like me to sort this out (mind you it was our family car & the only car too, so no the best idea to play with:)

But I am so happy that she purrs like a kitten now:) It did take a little while to have the confidence with her.....When you have spent so long trying to sort such an annoying problem you expect the car to keep behaving that way.

[fryea]

Well, took the board out to inspect again. It looks like that one spot I mentioned earlier (q12) and 2 other to220 based transistors have got spots on the back of the board that are very close together. Its like they have sunken into a pool though, and I cannot desolder it at the moment. This is my current speculation for thing shorting each other out. Whenever they get hot (eg, after running for about 5-20 mins) is when it happens.

Im thinking of either moving to resistor based injectors, which would bypass the heat being generated from those suspect transistors... or sticking some additional heatsinks to the top of the suspect transistors and mounting a small fan on top of megasquirt to keep them cool. My PWM% is currently set at 40. Would a higher or lower number affect the heat being put out by the PWM transistors? I just have ****ty greentops in there right now, so if I damage them I am not too worried.

[Karl Buchka]

Did you try moving your extra outputs off the IAC pins?

[fryea]

I moved the fan control directly to the db-37 (pin 29)

Knock control is still plugged into IAC1A, but I have unplugged the fuse to knock sense, so there is nothing coming into it. I need to walk out to the garage (ITS COLD!) and get a longer piece of wire to move knocksense off of the IAC pins. Will move that directly to the DB37 pin 25

[Karl Buchka]

I moved the fan control directly to the db-37 (pin 29)

Knock control is still plugged into IAC1A, but I have unplugged the fuse to knock sense, so there is nothing coming into it. I need to walk out to the garage (ITS COLD!) and get a longer piece of wire to move knocksense off of the IAC pins. Will move that directly to the DB37 pin 25

Yeah, I had resetting problems even with nothing connected externally.

[fryea]

Yeah, I had resetting problems even with nothing connected externally.

Alright, so finally got around to ripping MS out (again) and took out the IAC jumpers, and did a bit more board cleanup. Going 24 hours strong with no resets... my time test is 3 weeks to see if it really REALLY worked or not. Wish me luck.

[crazy0000]

Good luck, I got so fed up last time it reset, I didn't drive for a week. Tonight I had a bit of energy and time(aside from school work) and decided to beef up the grounds by adding a engine ground straight to where my MS is grounded and where all my sensor are grounded. Took the car out for about a half hour drive, doing anything I can to make it reset, but couldn't get it to... So far I have done several little things that make the resets go away for an hour or so then they would come back strong, but the ground seems to have done something, but in the back of my head I know they are just waiting to catch me off guard and smash me into the back of another car... So good luck, I also need it.

P.S. My board is now spotless, there is not a drop of solder flux, residue anything, you could eat off it, or use it for a mirror, thats what taking MS out 4 times to clean it does, not to mention soaking it in 99% isopropanol then soaking it in distilled water, instead of just dripping it over it.

[fryea]

P.S. My board is now spotless, there is not a drop of solder flux, residue anything, you could eat off it, or use it for a mirror, thats what taking MS out 4 times to clean it does, not to mention soaking it in 99% isopropanol then soaking it in distilled water, instead of just dripping it over it.

I use the isopropanol with a toothbrush and scrub the board a few times

Also, its hard to MAKE it reset, because it only wants to do it when I don't want it to do it

[crazy0000]

Well I did scrub it also, but I found if I just did that and didn't fully rinse it, it would leave residue which I'm pretty sure was a thin film of flux left on the board, especially areas like the db37 connector because it was so hard to get at.

and yeah, I know what you mean, it will only start resetting like crazy when I have to go somewhere and can't be late.. If I just go for a casual cruise, its the perfect car.. lol

[permieshane]

Have you guys got access to another working MS that you could plug in for a while & see what happens to the resets?

[crazy0000]

I was thinking of trying this, but I do not have someone near me that I know who has a V3 board

I would just like to point out, that after I threw an engine ground to where I ground all my ms stuffand sensors, that my MS hasn't reset yet. And normally would after 20 min of driving.

[fryea]

Have you guys got access to another working MS that you could plug in for a while & see what happens to the resets?

I have a MS-II that a friend just put together that I might try out if my recent round of corrections doesnt work. Driving 40 miles out to my home for thanksgiving tonight, we shall see... It only WANTS to do it when I am driving friends places, where megasquirt KNOWS it will embarrass me.



Hey crazy, where on the engine did you add the extra ground to?

[crazy0000]

I added A wire with a loopy connector right on top of where the engine ground is already on the valve cover, so now there are two of those things on one of the valve cover acorn nuts, and it goes straight to the MS grounds in my car which happen to be right where the old LH 2.4 unit used to be screwed into

[crazy0000]

I just wanted to post that, I finally trusted my car enough to take it to school, with it running for two days zero reset, this morning it reset within a minute of drivng, and it kept resetting, I said **** it and took the bus again. The only difference between those day and this day is that I drove at night with zero resets and today I drove in the morning... I kind of hate my car

[fryea]

I just wanted to post that, I finally trusted my car enough to take it to school, with it running for two days zero reset, this morning it reset within a minute of drivng, and it kept resetting, I said **** it and took the bus again. The only difference between those day and this day is that I drove at night with zero resets and today I drove in the morning... I kind of hate my car

IT'S HELL!

On that note, I drove to North Bend last night (40 mile aformentioned trip) and only had 1 minor reset on the way up a big hill. Still more than I'd like, but no stopping on the side of the road for 2 minutes to let "whatever it is" cool down.

The real test will be driving in downtown traffic sometime... echk

[permieshane]

I just wanted to post that, I finally trusted my car enough to take it to school, with it running for two days zero reset, this morning it reset within a minute of drivng, and it kept resetting, I said **** it and took the bus again. The only difference between those day and this day is that I drove at night with zero resets and today I drove in the morning... I kind of hate my car

I know that feeling, it is amazing after spending so many hours rebuilding my car that I could loathe it so much. I had a break from her (purchased a nice 960 wagon) for some time & then started to work on her again. Once I got it all sorted it was amazing how much different I feel towards her now - she is such a pleasure to drive.

So there is hope.

[crazy0000]

I dont want to clutter this thread, but I bought a 1990 740 turbo for the winter, and hopefully I'm going to spend some time redoing the wiring for the 240 over the winter, hopefully that will solve the problem, because everytime I drive my car and its perfect(very rarely), I love it so much, it makes it so hard to even think of parking it, even if its just for a few months.

But its just going to have to be that way.

[fryea]

Alright, so another twist in the plot of this neverending nightmare:

I have gotten rid of 90% of the resets. However, now after, say, 30+ minutes of driving, instead of getting a reset, megasquirt will pop its fuse instead. For the lifetime of it I was running a 3 amp fuse with no fuss. It popped once and I switched it to a 10 amp fuse. Popped that as well. I dont think I will go higher than a 10amp fuse to megasquirt, but my resetting problem has definately transformed into something... uglier.

[Boris740]

MS draws miliamps. If you are popping 3A fuse there is something seriously wrong. Instead of the circuit board, have a look at the soldering job on the 37 pin D connector. I do electronics for living and that connector is a cheap pita. Keep in mind that one row of pins is all ground. You could have a something shorting out there. Check if incoming 12V is not shorting out intermittently against of one the opposite row pins. You also have Vref going out (pin 26 on MS1) if that is shorting out anywhere in the harness you will get interesting results.