Electrolytic corrosion damage

[BradC]

I have a fuel burning heater in my vehicle. At some unknown point, I suspect water has made its way into the electronics cavity and caused some neat-o electrolytic corrosion.
Damage wise, the only thing that seems to have occurred is the glow plug measuring circuit.

The glow plug is approximately 300 milli-ohms, so the device measures it by applying battery voltage through a 6.8 ohm resistor and measuring the voltage across the plug. This is a 1W resistor, and the measurement takes place using a 770us pulse at 9.8hz. This is switched with a PNP transistor. What appears to have happened is with the water in there it has got to the base drive circuit on the transistor which has turned it on enough to cook the 6.8 ohm resistor (the poor thing would have been trying to dissipate far in excess of its rating).

Everything else seems to pass the sanity check. I'm after some advise on the level of remediation I should be attempting. I've given it a good scrub with a nylon brush and PCB cleaner (the photos are as-found rather than this point in time), and I'm waiting on some dremel nylon brushes to arrive so I can get into the tight spots between the pins. The whole device barring the connector pins is conformal coated with a coating that seems to soften when exposed to the PCB cleaner.

There is some corrosion on/around a few of the pins to the CPU daughter board and of course the bottom of the power board below the high-tide mark.

I'm debating :
- pre-heating and removing the 7 pin MOSFET on the bottom of the board
- de-soldering the 39 pin daughter board connector

I'd rather not do either. I'm concerned about doing more damage desoldering that the current corrosion damage. Given the corrosion was electrolytic rather than chemical, I suppose I'm after advice on a path of least resistance (pun unintended) treatment as I figure if it's adequately treated it's not likely to march on as it would if it were alkaline for example. I've ordered a new 6.8 ohm resistor, and as far as I've been able to test that's all that is required. Every other sensor input and device output tests ok.

The resistor is a week away, so I'm not in a hurry and I'd appreciate some experienced input.

[m k]

Multilayer modules are a bad rap but it's many times better than it looks.
Those are also usually varnished or something like that.
Nasty bits are via to spot connections, you just can't see how they really are.

Connector pins are extra important.
You should see where the contact point was and it will be almost there later also.

[BradC]

Good call! Thankfully only 4 of the pins near the tide line are used. All the important stuff is up the top. I've checked the connector in the vehicle and there is certainly water ingress. Thankfully the connectors are easily dismantled so I can address that and worst case the sensor harness is easy to replace in its entirety and it's affordable.

I've gone over the vias and thankfully all the vias near the water line are all solder filled with the lightest of surface corrosion. A gentle scrape with a blade reveals clean and unbroken solder underneath.

It looks like the most damage was around the power semis with electrolysis to the case. The tab on the power MOSFET is quite pitted. Thankfully that seems to have borne the majority. Theres a couple of missing pads on an 8 pin SOIC footprint on the bottom left of the board, but again hasn't seem to have affected anything.

[BradC]

Progress. I ended up suspending the power board in the ultrasonic cleaner with DI water, a bit of IPA and a couple of drops of dish detergent. That did a reasonable job of all the loose stuff, but it shredded the acrylic conformal coating. I cleaned a lot of that up manually with the tweezers. A second DI rinse, then an IPA rinse and into the oven for a couple of hours to make really sure it's dry.

I was waiting for a dremel nylon pen brush to rock up. In the mean time I read up on the acrylic coatings, and they were all delivered on a primarily toluene solvent. When the brush arrived I used that and IPA to clean up the remainder of the burn/flux residue, any remaining corrosion grunge and the loose acrylic. I then gave the power board a good spray with some carb cleaner which is predominantly toluene and alcohol. That beautifully dissolved and redistributed the originally very heavy coat of acrylic. A quick rinse with alcohol and a dry in the oven and we're looking good. A couple of +12V vias burned out but thankfully in an unused part of the board. Put it through a set of diagnostics and all used I/Os pass tests.

Waiting for the repair parts to arrive now. Plan is to repair, remove the flux, put heatshrink over all the connector pins and give it a good extra layer of spray acrylic (which I've already checked for compatibility with some of the old acrylic flakes).

[BradC]

Not too pretty, but repaired, bench tested and coating applied. Now into the oven for a full dry/cure (acrylic really dries rather than cures) as it's 2 decently heavy coats.




Put Kapton tape over the motor terminals and heatshrink over all the pins to keep the lacquer off.
All i have to do once it's ready to go is scrape the lacquer off the earth pads. I was going to mask them, but they were already coated from the "laqcuer re-flow" with the carby cleaner.

[m k]

Looking good.

One more thing, fill up before it's too late.
(empirically verified)
(>20 below and diesel)