Open circuit splices

A customer recently complained of their car occasionally dying while driving, then starting only intermittently, and finally not starting at all.  We kept the car for quite awhile to make sure that the repairs we made would fix all of the problems.

I found a burnt precharge resistor, and replacing it got the car running again.  But the customer knew not to cycle the key with the heater on.  And not only that, a burnt precharge resistor couldn’t have been responsible for dying while driving.  I did find that the ribbon cable near the precharge resistor had been burnt, and I could see bare conductors with their insulation melted away.  (That repair was not done by me; don’t you know by now that I never make mistakes while working on cars?) I don’t have the schematics to indicate what signal each wire of the ribbon cable carries, so I couldn’t really guess the result of any of them shorting together, but I do imagine it could cause enough of a problem to shut down the battery and thereby shut down the car.  So I spliced in new wires to the ribbon cable inside the traction battery where it had been damaged.

The repair could have explained all the symptoms, but could there be a new root cause to the precharge resistor overheating?  I had really hoped that the heater/key cycle snafu was the only cause.

I kept the car and continued to drive it as my own personal vehicle to make sure all the problems had been solved with the repair.  And of course they hadn’t.

There was no fixed periodicity for the car to die while driving.  It would just happen without any identifiable pattern.  The contactors would click open, the power limit light would flash, and it would coast down.  Eventually I got the hang of managing it in traffic: Shift to neutral, cycle the key (it would reliably turn back on), and shift it back in to drive.

That was the norm as I was troubleshooting until on one occasion the car died just as I was maneuvering to parallel park outside my apartment building, but it died in an entirely different way:  Everything went dark.  No lights on the instrument panel or center console.  The condenser fan was blowing even with the key off.  I had just made a very sharp u-turn to the left.

It was a small parallel parking spot even for a Think, but I managed to park it under human power.  I disconnected the condenser fan so it wouldn’t run the battery down and walked back to work while thinking the situation over.

I drove the shop’s car back with some diagnostic tools. One of the first and easiest places to start is the fuse box.  I didn’t find any blown fuses, but what I did find with a test light was certain fuses had no power on either side of the fuse element.  I noted their numbers, went to the circuit diagrams and found that they all were powered by a common splice out of the power distribution box.

All the wiring for power distribution is on the far right under dash toward the firewall.  As I found the wires I was looking for I heard some quiet noises from the car– maybe it was the hum of integrated circuits or relay or two clicking. I extracted myself from under the dash, tested my fuses again and this time found power. I had changed nothing; I had only poked around at some wires.

The car started up like normal and I drove it back to the shop.  It was dying more reliably now, but in the manageable way that allowed me to drive it on side streets without causing too much of a traffic nuisance.

The next day I cut open the heat shrink around the intermittently open splice.  It looks quite neat and professional, all the shiny copper strands with their insulation stripped back just the right length and squeezed into a very solid-looking squarish cross-sectioned joint, two larger wires coming from the power distribution junction connector and about five smaller wires continuing out the other side of the splice to their respective terminals in the fuse panel.  I don’t know how those splices are made; I assume there is a powerful machine that squeezes them tight while applying heat.  I have since noticed in Honda circuit diagrams “Thermal junctions,” and I wonder if they are the same.  The splices certainly don’t appear to have any sort of solder applied.

I had run into a failure of this type once before in a Think.  That time, though, it was on a ground circuit and required much more interference removal to make the repair.  The only symptom of that failure was that the horn did not work, and I had previously diagnosed it as having had an open circuit somewhere in wire between the clock spring and the relay panel, so I had run a new wire in parallel for the entire length.  The car came back for a different reason– maybe the driver was so easy-going as to never have tried the horn, or maybe it worked for him on the occasions that he did try to use it– but I really clenched up when I found out that my repair had failed.

When I made the correct diagnosis (knock on wood), and unearthed (pardon the pun) the faulty ground splice, it looked just like the faulty power splice:  shiny cooper neatly, tightly crimped together– impossible to imagine that there was any way electrons could not glide right through it.  Such an insidious fault.  At least they had the decency to quit their intermittent nature for long enough that I could nab proof of their fault.

In each case I added solder to the joint.  The one on the power distribution side was easy enough to remove the pins from the junction connector and slide heat shrink back to where it was before, but there was no reasonable way to get heat shrinkable tubing around the ground-side fault, so I taped it up thoroughly.

So what’s to be learned from these experiences:  Is there another MLEC failure mode?  Unfortunately, yes, if just the wrong splices open circuit for a brief second it can have the same effect as turning your key off then back on (or worse).  If your heater is on during that time you may have just cooked your precharge resistor.

But wait, I haven’t even finished the story/repair yet: The customer’s car was still dying after the intermittently open power slice was repaired.

There were no useful trouble codes, so instead I started watching raw data on the CAN bus using PCAN-View.  At first everything would look normal– not that I have enough familiarity with the message IDs and data to know that it looked good, but there weren’t any errors with reading data off the CAN bus.  But as time went on, after a minute or two errors would start to show up every once in awhile, and they would get more and more frequent until finally the car would shut down, unable to get its critical data back an forth between control modules.  Something was talking over all the other control units, clogging up the bus.

I put out an email to the other Think technicians asking for their advice, and Rick Steiner of Tom Wood in Indianapolis told me that it was probably the CDCM.  I disconnected the CDCM on his advice and drove the car.  The car stopped dying while driving, and I didn’t see any more errors on the CAN bus.  That’s not to say that messages weren’t going back and forth containing information about errors– when the CDCM is disconnected the high voltage warning light flashes and you obviously have no heat or A/C– but the data on the CAN bus was clean.

Long story short, replacing the CDCM fixed it.  And I believe that it was the intermittent open circuit that caused the issue with the CDCM.  The CDCM and VCU can also be ruined by disconnecting the high voltage heater connector from the PCU before disconnecting the low voltage heater connector.  That’s discussed in a bit more detail in the Think Technician ESSENTIALS post.

Automotive repair stories involving me, though, usually end up being long stories.  So when I went to replace the CDCM I managed to ruin it and the VCU by unnecessarily and improperly disconnecting the high voltage heater connector– I have probably done this at least 4 other times, scrapping previously perfect good, new control units. I know there are sayings about how mistakes are only really mistakes if you make them more than once, but this particular issue is just so counter-intuitive to me it just continues to confound me.

And to top that we didn’t have any more new CDCMs in stock.  So I put in a used one and a new VCU, reflashed them to the latest firmware and all was good. Except that after I made a harrowing trip half-way from Portland to Seattle to return the car to the customer, the customer found that the heat did not function. It’s not that I didn’t notice that the heat didn’t work after the repair.  (By the way, Think owners probably already know this far better than me: if you are planning a long trip that is going to draw your battery way down to near zero, you MUST NOT drive over about 55 mph.  The energy consumption of keeping pace with interstate freeway traffic is so, so, so much greater than going just 5 or 10 mph slower.)  Having done lots of 12,000 miles services on Thinks during the summer months I have notice that the heat will not turn on if the ambient temperature is not low enough.  I suppose there was some amount of wishful thinking when I decided that this was the case with this new CDCM– after all the A/C worked, why wouldn’t the heat work?  (From now on I force the heater to turn on by dipping the ambient temperature sensor in ice water.  I’ll show you exactly where it is in another post.)

So the story is not actually over yet.  I still need to get the heat working, and hopefully that will simply require replacing the CDCM (without killing it and a VCU again *sigh*).  And that will be the last repair it will ever need, and everyone will live happily ever after.

Update: So I replaced the CDCM, which would only be a relatively small amount of labor if it didn’t entail a 3-hour drive to Seattle.  And there was no change.  So I got as much diagnostic information as I could and headed home to plan a new strategy.  To be continued…

 

The End

But if anyone knows anything about the procedure that is used for making these wire harnesses, specifically their splices, I would like to learn more about it.  Would it be reasonable to go in and add solder to every single splice in every single car on the off chance that one might open for an instant in the future?  (I would probably be more likely to cause a lot of problems doing a repair of that magnitude.) (I’m compiling a list of pre-emptive repairs that will make Thinks reliable and improved against all pattern failures.)  Was there a problem in the wire harness manufacturing process or was it handled roughly when it was installed in the car?  Could someone in the passenger seat have kicked it accidentally and that caused the open circuit?  There was no one in the passenger seat when it happened to me– was it the minute centripetal force of my u-turn that caused it?

-JM