Interesting. As a person deeply involved in computational electromagnetics, I don't think the high current line caused the ECU to malfunction. Surely it's possible but it depends on your board content and angle to that transmission line thus it looks like a long shot to me. There are lots of parameters to be considered in your case but a 7-8A current radiates quite a lot of noise inversely proportional to square of the distance and you probably immersed your board to this radiation from various angles at close proximity. If it's not breaking down, it's not breaking down. Without knowing the characteristics of that line you mention (it can't be a 384kV transmission line..), I believe those rocks you mentioned might be the source of this malfunction. It's not that common actually but an occuring phenomena. DOI: 10.1023/A:1021199018553 (Journal of Mining Science) On Mon, Jun 11, 2012 at 9:14 PM, Electron wrote: > > Howdy, > > I'm finally testing my dirk bike ECU. It worked wonderfully till a point, > where the engine turned off and didn't want to start anymore. I thought i= t > was maybe a bad/cold soldering, so I just replaced it with the original a= nd > didn't bother anymore for that day. > > But when back in the lab, the ECU worked! I made a lot of tests and my EC= U > proved to work absolutely perfectly, even when hammered. ;) So what was i= t > that made it suddenly stop working? > > I immediately recalled where it happened (I even have GPS logs): I was in= a > dried river bed full of rocks, and a thick gray cable was making impossib= le > for me to pass. I stopped in front of it, the engine turned off but I tho= ught > it was just because of carburetion problems (too low idle), etc.. > > I noticed it was a very high tension/current cable that fell down from a > tower. Regardless of that, it felt natural for me to grab it with my hand= , > raise it, and thus push the bike beyond the obstacle. Then I tried to sta= rt > the engine but, as said, it didn't run, so (as I wasn't alone and everybo= dy > was in a hurry) I quickly swapped the ECU with the original, and the engi= ne > restarted. > > I am thinking to return there but it means to do 200 kms alone.. it's not > safe to do it off road, but thanks to the GPS log I found a way to do 99%= of > those 200km via asphalt, and I will return there to clear my doubts. > > I tried to disturb the ECU in my lab with 230V 2000W appliances, but > evidently that wasn't enough. I will have to return there, near that cabl= e, > and see if the fault is repeatable, also because I've been using it for > many hours today and it worked flawlessly. > > But as it's very distant from home, I want to make 1 trip and 2 jobs: I w= ant > to bring with me a 2nd prototype, with a supposed workaround/fix/improvem= ent > in place, and see if this new prototype is immune enough to that terribly= noisy > cable. > > Will shielding help? It probably won't, as it's just 50Hz magnetic field = I think, > but I wish to ask you. > > Will placing the Xtal nearer to the MPU help (currently 10mm), as well as= moving > the Xtal caps to the bottom of the PCB board, so they're as near as possi= ble to > the Xtal and MPU too? (currently the caps are on the same PCB side of the= Xtal > and MPU and are comprehensibly not really close (another 10mm) but, worse= , their > ground return doesn't go into the MPU Vss pin, but to the opposite side o= f the > PCB). Can this have made a difference? > > What else would you make to ensure maximum immunity to the noise coming f= rom > a very high tension/current cable that is there although it should not be= there? > > It's a chance I have to make my circuit more robust and gain experience b= efore > they fix that cable (because it can't be there, it's dangerous, and I wil= l notify > the local police about it then, i.e. hopefully tomorrow morning already). > > Oh yes this time I won't touch it, I will reach it near with my bike and = ECU but > not grab it anymore. If it was the cause of the malfunction, it indeed di= d it even > at 1 meter of distance, so I'll do it from a small distance. > > Cheers, > Mario > > -- > http://www.piclist.com PIC/SX FAQ & list archive > View/change your membership options at > http://mailman.mit.edu/mailman/listinfo/piclist --=20 http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .