Corroded PCB repair

[brabus]

Hello everyone,

I scored a nice BMW MK4 trunk navigation unit, sold as "not working". The seller says it stucks on boot logo, sometimes does not turn on at all, sometimes it even works. The cause seemed to be a corroded PCB, pretty usual in such an unit to be honest.

Following the motto "don't turn it on, take it apart", I immediately opened it; what I found is in the attached pictures.

It seems that the top side of the PCB is pretty fine, no sign of corrosion whatsoever. On the bottom side, the corrosion is evident but located on a specific side of the board, the input/power section.

I have two doubts, so I kindly ask your opinion:
- What should I use to clean up and deactivate the corrosion?
- A couple of ICs clearly need to be replaced; should I also replace all the passives on the bottom of the board? Some of them seem in really bad shape.

Thank you.

[HAL-42b]

Doesn't seem that bad to be honest. Squirt IPA and gently brush with a toothbrush then dry with compressed air.

That ST chip looks like it has one leg corroded off. I would try shorting that off with solder and booting it to see if the things improve. If not there will be a lot of work.

[Shock]

Check out this thread: https://www.eevblog.com/forum/repair/i-got-a-dirty-mac/

I used 100% IPA in a squirt bottle and a soft bristled toothbrush. The idea is to flood the area with IPA soaking as you go and just slowly drag the brush over the PCB using the weight of the bristles. Work on a small area at a time downwards towards the edge of the PCB. Once you have loosed the corrosion spray it with fresh IPA and it will wash away those contaminants off the board. Keep the toothbrush clean and soaked with IPA.

Use a container to capture the dirty IPA and do it in a well ventilated area with no ignition sources. I used magnification to inspect as I went. You have to be super patient or risk scraping off a component and the pads with it. It took me three rounds to really clean that PCB up and at the end there was no visible corrosion.

If there are brittle joints they need to be reworked also look if tracks have been eaten into and if solder mask is damaged, that could make it a bigger job.

[uncle_bob]

Hi

This is not all that uncommon for stuff that has been flood damaged. Occasionally you can see it from wreck salvage where the legal side of things kept it out in the rain for a while.

Here's the nasty thing to watch for:

Part leads are plated first and cut second. Yes, that sounds weird. I only really figured that out repairing a bunch of flood damaged gear. The problem that comes from that - you have bare metal on the end of the lead. It *should* have been covered in the soldering process on an SMT part. If it's a through hole part and was cut after soldering - not so much.

In normal use, this is not a problem. The parts never see enough crud to start to rust. With normal rain water, it's also usually not a big deal. Flood water is full of nasty stuff. What stays in the gear concentrates the muck. You rust out the middle of the leads. There is a shell of plating sitting there. The lead *looks* fine. Poke it and it disintegrates. I can't say that I ever figured out anything better than "poke the leads" to find the ones that had rusted out.

Lots of fun !!!

If the gizmo does not fully revive after IPA is used, the next step is really good distilled water. First you scrub to take anything off the surface. Next, you soak the board in it to pull out the salts. After a few soak / rinse / soak in fresh cycles you have done all you really can do. Next step is a bake in an oven at maybe 120C for a few hours. You need to run it with the door open so air moves through the oven. You want *dry* hot air in there. When it comes out of the oven, spray it down with your favorite silicon lube spray. If it got bad enough to soak stuff into the board, it will be prone to do so again (you can't get all the crud out). The silicone will help protect things. Keep in mind that once the silicone goes on, it's not coming off.

Bob

[Gyro]

I think I'd do it the other way around...

Scrub and soak the the area in distilled water to dissolve out the salts, followed by a dry, followed by IPA to bind up and pull out any remaining moisture in the PCB and underneath packages.

[Refrigerator]

Not that bad to be honest, should clean right away, also don't forget to coat it with some conformal coating to prevent future damage.

[uncle_bob]

I think I'd do it the other way around...

Scrub and soak the the area in distilled water to dissolve out the salts, followed by a dry, followed by IPA to bind up and pull out any remaining moisture in the PCB and underneath packages.

Hi

The theory was that the heat had driven everything out of the board, including whatever residuals the IPA had in it. We sprayed the boards while they were still warm and agin a day later.

Bob

[Gyro]

Sure Bob,

My point was that IPA is no good for removing aqueous salts (clearly present from the corrosion) and distilled water is. If not removed, the salts are just sitting there ready to be re-activated by any humidity/moisture. IPA is only good for removing moisture and organics, hence my suggestion of water followed by IPA.

With proper heat drying the IPA step is almost certainly redundant.

Chris

[TerraHertz]

Do you know what caused the corrosion? It looks a bit like the 'creeping death' from NiCd batteries - blackened solder surfaces, and creeping separation of soldermask paint from the copper fills.

Try melting any of the particularly blackened solder joints. It there's real solder under a thin surface discoloring and it melts, it's not NiCd creeping death. On the other hand, if the entire volume of solder in the joint has converted to a kind of brittle, fragile, non-meltable garbage, then it is. In which case give up.

I have no idea what the chemical process is by which old NiCds destroy everything around them, but they do.

[justanothercanuck]

doesn't look so bad...  i'd clean it, buzz it all out to see if the traces are still working, and if it still bootloops or whatever, possibly change the RAM chips.

[brabus]

Hello everyone,

Thanks a lot for all your hints, much appreciated!

I let the affected side of the PCB soak in distilled water for about 2 hours, then brushed gently away the crusts. Quite a lot of dirt came off with this first pass.
Then I dried it with compressed air.
Then came the IPA, I brushed all over the place with a bit more force, in order to remove the harder crusts.
I noticed that in some places the soldermask came off rather easily: I scratched it away with a pin, noticing that the affected area is luckily quite small (see photo). Not bad, I will flood it with solder and apply fresh coating.

I started to worry again after noticing that some components seem partially eaten away. It seems that the corrosion is more intense on some SMD terminals, coincidentially the ones connected to +12V. Take a look at the photo, the plating on the capacitors seem to have disappeared, and some other components exhibit the same.

I guess I'll replace those damaged caps (and resistors here and there), restore the missing connections and give it a try.

Thank you very much, I'll keep you updated!

[Gyro]

I started to worry again after noticing that some components seem partially eaten away. It seems that the corrosion is more intense on some SMD terminals, coincidentially the ones connected to +12V. Take a look at the photo, the plating on the capacitors seem to have disappeared, and some other components exhibit the same.

I guess I'll replace those damaged caps (and resistors here and there), restore the missing connections and give it a try.

Not a coincidence, the +12V supply is the one with enough voltage / current to get a decent amount of electrolysis going. Yes, it would be a wise precaution to replace any passives (capacitors at least) in the affected area as they may no longer be internally connected.

[station240]

Board is still a big grubby, main thing to watch out for is buildup around the tiny SMD capacitors/resistors. This can result in leakage currents, which upset the circuit.
I got a small scrubbing brush and paid a lot of attention to these areas when I had to fix something that fell into a pond, got it working again.

[brabus]

Yes, the board is still dirty, even if I scrubbed it thoroughly. I will proceed with the repair/rebuild, maybe some flux can help. At the end, I will let it soak in IPA and scrub it again before drying it and apply the protective coating.

Regarding protective coating, is there a specific product that you strongly recommend?

Thanks!

[helius]

VpCI-286 looks promising. No experience with it, sorry to say.

[Tomorokoshi]

In the cleaning step put some 409 in the water, then rinse that out with clean water.

Watch out for deposition of contaminants under the SMD devices.

I agree with coating the board after you get it to work. It will be somewhat more "brittle" now. Just make sure to mask the connectors.

[Qmulus]

I deal with this kind of stuff all the time.  That one doesn't look bad at all.  Personally, I would scrub the damaged side of the board with a water/detergent mix with a plastic bristled brush, like a tooth brush.  Sometimes this takes more than one wash.  You CAN be fairly aggressive.  Follow that up with a thorough rinse.   Dry with compressed air and heat of you can.  Setting it in the sun for a couple of hours is good.  If there are components that you know that you are going to replace (like that 8 pin SOIC) or components that can't or shouldn't be washed (relays, electrolitic caps, etc.) I would remove them first.  I also have an ultrasonic cleaner with my own concoction of chemicals for cleaning PCBs with very fine pitched components, but that wouldn't be needed here.  Don't worry about using lots of water, chances are the PCB assemblies were cleaned with hot water after assembly.  As for the caps or other passives that may have slightly corroded terminals, I wouldn't touch up the solder unless you plan on replacing them. 

From the looks of it, I doubt that you have any damaged traces.  In my experience, you only get significant corrosion (in the case of water damage - no battery acid) on pins with power applied, or sometimes ground connection.  Normally the damage results from electrolysis on the power pins of ICs, and people figure it out once the device stops working.  The worst if if they are allowed to sit wet for days or weeks with the power connected.  Often that results in the power pins on connectors corroding off, ending the electrolysis.

Personally, I would NOT apply conformal coat.  I honestly hate the stuff... You are better off just not letting it get wet to begin with.  In my experience, unless the PCB assembly is dipped in conformal coat, water will wick under components, even through wires or connectors, and the resulting corrosion will be extremely hard to repair as first you need to remove the coating, then fix the damage.  Whatever you do, do NOT put on the conformal coat until you are 100% sure that your repair is complete. 

You would be amazed by how bad a circuit board can look, yet be brought back to life.  I have had a few with corrosion and mold that looked like write-offs, but with a hour or so of work ended up looking and operating like new.

[brabus]

IT WORKS!!!

I just finished the restoration of the navigation unit. Guys, you wouldn't believe what a job it was.
After scrubbing a bit more, I realized that I needed to change many components. I quickly changed ALL the resistors and capacitor in the corroded area, and checked the diodes and transistors. I found a SOT-23 double diode with a partially corroded leg, which seemed intact but had no electrical continuity (!).

After restoring all the components, I fluxed the board and did a reflow with my hot air gun. Nice to see how all the components line up nicely.

I then put in the car, and the splashscreen appeared! 

At the beginning, there was no communication via the IBUS (a basic serial bus used in BMW systems), and it required a thorough reverse engineering to find out what did go wrong. Basically the IBUS transceiver was missing a pull-up on the line; strangely enough, the corresponding pin on the PCB was not connected to anything. I soldered a 10k resistor to 12V et voilà, communication was restored!

I still do not fully understand how such a modification has been necessary to restore communication. Deeper analysis is needed, together with nice photos showing the final result!

Thank you very much for all your precious hints, the PCB is now shiny as new.

[helius]

Great news!

I have no idea what the chemical process is by which old NiCds destroy everything around them, but they do.

Attack by potassium hydroxide, with bonus heavy metal catalysts. KOH is neutralized by CO₂ in the atmosphere, so corrosion is worst in sealed enclosures—yielding black copper oxide. In well-ventilated environments the main product is bluish-white potassium carbonate, and clean up is relatively minor.

[Warhawk]

IT WORKS!!!

I just finished the restoration of the navigation unit. Guys, you wouldn't believe what a job it was.
After scrubbing a bit more, I realized that I needed to change many components. I quickly changed ALL the resistors and capacitor in the corroded area, and checked the diodes and transistors. I found a SOT-23 double diode with a partially corroded leg, which seemed intact but had no electrical continuity (!).

After restoring all the components, I fluxed the board and did a reflow with my hot air gun. Nice to see how all the components line up nicely.

I then put in the car, and the splashscreen appeared! 

At the beginning, there was no communication via the IBUS (a basic serial bus used in BMW systems), and it required a thorough reverse engineering to find out what did go wrong. Basically the IBUS transceiver was missing a pull-up on the line; strangely enough, the corresponding pin on the PCB was not connected to anything. I soldered a 10k resistor to 12V et voilà, communication was restored!

I still do not fully understand how such a modification has been necessary to restore communication. Deeper analysis is needed, together with nice photos showing the final result!

Thank you very much for all your precious hints, the PCB is now shiny as new.

Congratulations ! Pictures please !

[Greg981]

I always advise caution to anyone who needs to use IPA on electronics.  If they don't know what they're doing, they might use the wrong stuff...

[Dave Hedgehog]

I deal with this kind of stuff all the time.  That one doesn't look bad at all.  Personally, I would scrub the damaged side of the board with a water/detergent mix with a plastic bristled brush, like a tooth brush.  Sometimes this takes more than one wash.  You CAN be fairly aggressive.  Follow that up with a thorough rinse.   Dry with compressed air and heat of you can.  Setting it in the sun for a couple of hours is good.  If there are components that you know that you are going to replace (like that 8 pin SOIC) or components that can't or shouldn't be washed (relays, electrolitic caps, etc.) I would remove them first.  I also have an ultrasonic cleaner with my own concoction of chemicals for cleaning PCBs with very fine pitched components, but that wouldn't be needed here.  Don't worry about using lots of water, chances are the PCB assemblies were cleaned with hot water after assembly.  As for the caps or other passives that may have slightly corroded terminals, I wouldn't touch up the solder unless you plan on replacing them. 

From the looks of it, I doubt that you have any damaged traces.  In my experience, you only get significant corrosion (in the case of water damage - no battery acid) on pins with power applied, or sometimes ground connection.  Normally the damage results from electrolysis on the power pins of ICs, and people figure it out once the device stops working.  The worst if if they are allowed to sit wet for days or weeks with the power connected.  Often that results in the power pins on connectors corroding off, ending the electrolysis.

Personally, I would NOT apply conformal coat.  I honestly hate the stuff... You are better off just not letting it get wet to begin with.  In my experience, unless the PCB assembly is dipped in conformal coat, water will wick under components, even through wires or connectors, and the resulting corrosion will be extremely hard to repair as first you need to remove the coating, then fix the damage.  Whatever you do, do NOT put on the conformal coat until you are 100% sure that your repair is complete. 

You would be amazed by how bad a circuit board can look, yet be brought back to life.  I have had a few with corrosion and mold that looked like write-offs, but with a hour or so of work ended up looking and operating like new.

As someone who also does this sort of thing for a living, this is good advice and pretty much exactly what we do.
We wash with soap/water, rinse with plain water then IPA, and then dry in an oven at 70c for a couple of hours.
I always give the PCB a very thorough initial clean with a reasonably stiff brush - Don't be gentle - if anything is so fragile that it falls off under brushing then you're better off knocking it off now. Take pictures of the board beforehand if you don't have a spare to compare.