MembersPage/MarcellGal/EngineSwap/ElectronicProblemList (2006-01-27 06:57:44)

Electronic problems on the way - during MembersPage/MarcellGal/EngineSwap

I think that collecting these experiences can be useful. There are many things to watch out for, easy to overlook something.

Problems related to existing harness

It's now obvious that working with an existing harness is harder than make one from scratch. I decided to use it for the connectors that I didn't have separately.

It's a good idea to make a more detailed review and documentation of any existing harness than on MembersPage/MarcellGal/EngineSwap/DaewooHarness. Describe and categorize all connectors and their hidden connections, even the ones you know nothing about.

Problem Effect
bad identification of car's node15 and node30 delay with the starter motor. Starter motor and generator had to be connected after engine in place: harder
using AC refrigerant pressure sender connector on TPSruined MAP sensor due to reverse supply
bad grounding due to unknown connections in the daewoo harness delay (not yet fixed)
forgotten injector supply wire. After some minor rewiring and some change in external circumstances (temperature) engine did not start. It barfed a little, and sparkplugs were wet. This was confusing. Running detailed checklist revealed the problem in first minute, since very first step was measuring with DVM if ECM sees injectors: it did not, because of forgotten injector supply.Stupid mistake, isn't it?delay
bad connector between engine harness and injector harness. It came from the original Daewoo harness. It worked well for long. After about 9 months, I experienced occasional misfiring with "appr." 2 cylinders. WBO2 showed that O2 gets to the exhaust, but I first thought it's the sparkplug, or the ignition module, or the HV wires, or something related to ignition harness. It was very rare first, hard to debug (problem not occuring when you want). When it became somewhat more frequent, the IgnitionPage/TransformerWithAmplifier became first suspect since problem usually occured at high MAP and warm engine bay. When installing the spare ECM and executing the minimal checklist I discovered that injector (12.5 Ohm /2) = 6.3 Ohm resistance cannot be measured at the ECM connector. I had luck that the connector showed the hazard when I measured, despite the fact it usually worked OK. After finding this, the fix was obvious: new injector harness from connectors from WebShop drove appr. 5km with 2 cyl (out of 400km since when the problem first occured). It was annoying. Good side is that I made a spare ECM since I expected the problem to stay occasional which means little chance to investigate what of the 6..8 things is wrong. But it turned out GenBoard/VerThree was fine from the start

Problems related to v3.1 board

Remember I used an old leftover v3.1 board (v3.2 is higher quality and higher assembly level) that failed the tests.

I thought I can solder by heart and jump over most of the tests because I know the board very well. But I can make mistakes.

Problem Effect
stepper259 chip's reset resistor was not solderedminor delay
510 Ohm and 100k swapped at pump1 delay
nernst1 amplified not amplifying (killed by static electricity) delay
injector channel tests failed. In fact the channels operated OK from the start, I just forgot to consider that the lookup from h[0] takes place and issued the wrong mdh.. command delay

Unfortunately after getting up from my chair, I sparked (discharged) onto the board with my finger. It was a loud audible (and I felt it too) spark due to static electricity (the chair and my slippers helped build up voltage). It was so powerful that we first thought it was from an ignition coil, but there was no coil connected at all. It was no fun. We tested the functions, to see if anything is ruined. Fortunately the AVR and most functions survived, but the nernst1 OPA channel died. The pump- OPA channel in the same OPA survived. I connected the nernst1 signal directly to the AVR input (bypassing the nernst1 channel), and reconfigured the controller in config.txt for the lower wbo2_ri_target and higher wbo2_nernstdc_target. Surprisingly, the wbo2 heater control worked well even though the input voltage became 1/4 of original. I'll log the heater-control behaviour after this change in the car, but I'm confident that it'll be OK.

Always touch the ground connection before getting close to the board to avoid harming the board with static discharge.

In fact there were no problems with the board, just my negligence when completing (in a hurry). With v3.2 much less steps are needed with less chance of errors.


Battery

I had to change the battery, some fun in the middle of the night. Old one was only 5 years old (40000km), but this car takes very short, typically 4..8km trips (and sleeps out in the cold in winter). I expect the old (44Ah) should be useful (after recharging) for the lab, for CDI experiments.

I checked VBatt right after starting with the new battery, to make sure the alternator works OK. Seems to be OK:


See also