Deathtrap

[R-P]

Being in lockdown and all, this is the perfect moment to try to get to the bottom of this.

So I did this: put leads on every connection between the CEM-PCB and the outer housing it slides into. This way I can get to the processor and flash memory in my CEM.

First pic is the bare CEM on the hood of my white Volvo that is parked under a tree with lots of birds...
Second one while attached to the rest of the car.

I had a great setup where I heated the processor, scanned the housing of the component I was heating with a infrared thermometer and video'd it all with my phone. But nothing happened for 10 minutes while the processor went up to 60ºC. So seeing that vid isn't worth your time

Then I did the same with the flash memory and the car failed at 35ºC!!! Unfortunately you cannot read the thermometer in that clip, so I ddin't bother to upload that either. Will try again tomorrow and will try to heat individual pins on the flash to see if it is a bad solderjoint, but so far they all seem OK (but I can only view one side due to the tin-can still around it, see third pic of the flash-chip and part of the tin housing, the white goo is heatpaste from when I had a heatsink attached).


It is often said it is not the processor, but the flash memory that causes CEM failure. In my case, this is indeed true.
I watched for overheating from the processor and flash from within, but there was no noticable rise in temperature with key on OFF or II position. Obviously I was not driving, so it may not have been doing a lot of work, but it is my belief that these issues come mainly from the environmental temperature. All the heatsinks and fans in the world will never keep MY flash memory below 35ºC when it is parked in the baking sun.

[R-P]

Here's the redone video.
Watch for the puddle light without sound or the clattering of the relays if you have the sound on.

https://youtu.be/szoaJSq3guI

[R-P]

Here's me pinpointing it to a specific pin.

Still have to clean it up (can't find my isopropanol), as I cannot find a bad soldering joint*, so still not sure what causes it. Technician who looked at my vid/pic said he sees failures on the "101" resistor array in Lucas ECU's. But I tried shortly heating it in the video and it didn't cause the failure.

Below a magnified pic of the pin I suspect the most.

It may be pin 12, which accordign to the specsheet is:

Chip Enable: Activates the device when LOW. When CE# is
HIGH, the device is disabled and goes into standby power
mode.

* I tried examening the pins with a USB-microscope, but the whitebalance on it sucks. The pic attached is with the 48MP setting on my cheapo Xiaomi and this is a cutout of the entire pic.

[R-P]

Here's me pinpointing it to a specific pin.

Still have to clean it up (can't find my isopropanol), as I cannot find a bad soldering joint*, so still not sure what causes it. Technician who looked at my vid/pic said he sees failures on the "101" resistor array in Lucas ECU's. But I tried shortly heating it in the video and it didn't cause the failure.

Below a magnified pic of the pin I suspect the most, but it is NOT the one connected to the "101" resistor-array.

I think the pin responsible is the 12th pin, the "CE":

Chip Enable: Activates the device when LOW. When CE# is
HIGH, the device is disabled and goes into standby power
mode.

* I tried examening the pins with a USB-microscope, but the whitebalance on it sucks. The pic attached is with the 48MP setting on my cheapo Xiaomi and this is a cutout of the entire pic.

[Panu]

Did you managed to find the right leg to do soldering? Does it any mater which way you put the fan? I have seen both ways. You will get it blowing towards CPU by changing the polarity.

[Simmy]

Originally Posted by R-P

Here's me pinpointing it to a specific pin.

Still have to clean it up (can't find my isopropanol), as I cannot find a bad soldering joint*, so still not sure what causes it. Technician who looked at my vid/pic said he sees failures on the "101" resistor array in Lucas ECU's. But I tried shortly heating it in the video and it didn't cause the failure.

Below a magnified pic of the pin I suspect the most, but it is NOT the one connected to the "101" resistor-array.

I think the pin responsible is the 12th pin, the "CE":



* I tried examening the pins with a USB-microscope, but the whitebalance on it sucks. The pic attached is with the 48MP setting on my cheapo Xiaomi and this is a cutout of the entire pic.

lead free solder is the curse of many electrical components on lots of circuit boards........

[Panu]

Originally Posted by Panu

Did you managed to find the right leg to do soldering? Does it any mater which way you put the fan? I have seen both ways. You will get it blowing towards CPU by changing the polarity.

The fan works only by red to + and black to -. The other way at least my fan did not turn. But it is done and working. I put the 40x40 fan and heat sinks for both CPU and flash. There is ground in the CEM and 15 (+ via ignition switch) you can find back of the cabin fuse box. So the leads which came with fan were long enough.

Now I can test it during next long trip. Thumbs up!

[R-P]

Just to update: error returned and I bit the bullet and had a Dutch expert at this stuff do a repair: he swapped the eeprom (the memory with the programming in it) and all has been flawless for 2 years I think.

He also editted in the one-button-down function for my passenger window which wasn't there when he read out the memory from the old chip before programming it into the new one.

Here's the link in Dutch, there's a button to switch to French/English/German.

He's a cranky dude when I rolled up in my Prius with the CEM in my hand, he wanted to measure the thing IN my Volvo and decide where to go from there, so I got some snarky remarks. After he went to work on the CEM and I drove around for 2 hours or so, he had mellowed a little and we had a nice talk about electronics where he carefully avoided explaining his business secrets (fair enough). Don't recall the price, but regret not doing it years sooner. But I needed these years to get some insight in where the problem might lie and gain some confidence that this would indeed solve something instead of just being yet another problem that eats money without ever being solved... (still haven't been able to fix my turbolag in 14 years of ownership and recently found out my AC issues were minor due to misdiagnosis of a specialist repairshop, so drove around without AC for >>5 years for nothing...)