Potting Compound Removal Success Stories?

[knobtwiddler]

Hi,

From the many threads I've seen in Google, the general advice seems to be 'it'll destroy any functionality and will probably take off component IDs. You'll need to use such a heavy duty process to remove the potting, it'll likely take most valuable info with it, including PCB laminations'. However, I haven't seen too many threads here about potting removal, so I thought I'd ask out of curiosity in case the Brains Trust here have anything to add...?

A post at one group suggested that a firm had made a CNC jig that milled off microns of layer at a time, photographing the excavated section and then reconstructing the photos to make a 3D model. Which makes me think that this potting removal business is a pretty desperate course of action... When I tried it before, I managed to do what I wanted, i.e. identify which part failed, but this was in an SMPS module and I made a beeline for the MOSFET, noting that D-S had failed short, i.e. the classic SMPS failure. It was hardly a forensic exercise.

The module I have in mind is from the mid 70s. Is definitely some form of epoxy (it is hard-ish, but easily scored with a blade). You used to be able to buy replica ones, but the firm making them went out of business, and the schematic's never found its way out. I could replace module with modern IC, but that would devalue the item (guess I could pot the IC once repaired...sneaky...).

I dipped a cotton bud in acetone and ran it across the compound, which left a brown residue on bud, along with a dull line across compound. Methinks leaving in acetone overnight will only soften up a few outer microns of coating.

I'm not sure if the compound goes all the way into the potting box... I could just try sawing off the lid and seeing if components are visible. This could be a start.

I doubt that I will get many optimistic replies to this thread... Please prove me wrong!

TIA!

edit - here is a fun (not) thread: https://www.diyaudio.com/community/t...hannel.358649/

[stevehertz]

It depends on what type of potting material is used. If it's a 'rubbery' type I've seen people have success using a scalpel to gradually hack away and remove small chunks at a time.

[Electronpusher0]

Take a look at this youtube video. It is an episode from CuriousMarc's series on restoring an apollo guidance computer.
In this episode they dig into a potted module and repair it.

https://www.youtube.com/watch?v=I7VLIjsKxkU

Peter

[McMurdo]

hard epoxy can be softened with a heat gun but it's a delicate exercise not to damage the pcb and it's easy to get burned. I suppose the fumes are probably toxic too.

There is an automotive part my brother uses and they're potted in hard black epoxy. They send them to somewhere in eastern europe who 'de-encapsulate' them and repair them to component level. I haven't seen the state they come back in but they must work and I dont know how they do it.

[Bazz4CQJ]

Probably one of the well-know "victories" in dealing with potting compound is the successful effort to re-engineer the Wellbrook magloop amplifier.

I think the approach used is mentioned on the Wellgood website, and just used the simple but slow method of solvents. MEK may be better than acetone. It might be that a solvent mixture (80% MEK, 20% xylene) might be better still.

B

[Nickthedentist]

Originally Posted by McMurdo

They send them to somewhere in eastern europe who 'de-encapsulate' them and repair them to component level.

I wonder whether they have access to solvents which have been long banned here.

[Bazz4CQJ]

Originally Posted by Nickthedentist

I wonder whether they have access to solvents which have been long banned here.

That could be the case, or simply that they are taking risks in the way they use them that could more readily get you in deep trouble here.

B

[David G4EBT]

Originally Posted by Bazz4CQJ

Probably one of the well-know "victories" in dealing with potting compound is the successful effort to re-engineer the Wellbrook magloop amplifier.

I think the approach used is mentioned on the Wellgood website, and just used the simple but slow method of solvents. MEK may be better than acetone. It might be that a solvent mixture (80% MEK, 20% xylene) might be better still.

B

In George Smart's case (M1GEO, designer of the 'Wellgood' amplified loop antenna), he'd inadvertently transmitted into his Wellbrook loop and thought he'd have a go at 'reverse engineering' it. He didn't need to remove the resin potting compound as he had access to X-Ray facilities, which showed the the Wellbrook was nothing more than a couple of ZTX327 transistors on a little piece of Veroboard and not a lot else. That enabled him to create his 'Wellgood' loop.

Then another amateur (G7UVW), sent him pictures of his own Wellbrook which had failed in 2012 due to water ingress where the potting had failed and had corroded the Veroboard. Looking at the pictures, the potting compound didn't appear to have adhered well.

Gary Tempest's Wellbrook failed too due to water ingress, which inspired him to create his own designs.

George Smart's current Wellgood site in here:

https://www.george-smart.co.uk/projects/wellgood-loop/

The link to his historic site which covers his investigations, shows X-ray pics and pics of the Wellbrook with failed potting compound is here:

https://www.george-smart.co.uk/proje...-loop-history/

Clever chap.

[Bazz4CQJ]

I've not had problems with my Wellgood and it's been out there for a few winters. I didn't pot it, but it's inside a plastic box and anywhere water could get it, it's sealed with Blue Hylomar paste. Blue Hylomar has had the CQJ seal of approval for many decades .

B

[Ambientnoise]

When I was at a company which potted acoustic transducers in various epoxies it was the old “nitromors” (dichloromethane ?) which was routinely used to salvage or investigate parts. Even that took days and only some electronic parts survived.

[Bazz4CQJ]

This is what George Smart had to say about revealing the Wellbrook;

"Dave G7xxx, who works in an X-Ray lab, offered to x-ray the failed part to assess the feasibility of a repair. From the x-ray images, it was apparent the device was made on veroboard, with several capacitors, ferrite transformers, and some active transistors.

I decided to try and remove the resin compound from the device. Various methods were tried, but the most successful was a combination of using a craft knife and pick, as well as warm (around 35C) MEK (methyl ethyl ketone) to dissolve the resin. Over time, we managed to get through, until we could see all of the connections".

https://www.george-smart.co.uk/proje...-loop-history/

If that link does not work, go to the Wellgood website and find the Wellgood Loop History page.

The crucial detail, not discernible from the X-rays, were the windings on two transformers.

B

[knobtwiddler]

The 'new' Nitromors doesn't even remove paint (I'm not sure what it does - I have a full can and it hasn't worked on any paint I've tried it on). I don't fancy its chances on potting epoxy!

Thank you for the Smart link. I wonder what an x-ray lab charges for a handful of shots?

At this point I am leaning towards designing a new circuit. I'm not sure the ratio of time invested VS rewards gained will be worth it.

[kalee20]

In the 'day job' we have used Electrolube OP9003 Resin Remover. It works well enough to get things apart for fault analysis. Down side is it's probably not available to the man in the street as it uses some of the chemicals the 'proper' Nitromors used - dichloromethane (AKA methylene chloride) and methanol.

Originally Posted by knobtwiddler

I dipped a cotton bud in acetone and ran it across the compound, which left a brown residue on bud, along with a dull line across compound. Methinks leaving in acetone overnight will only soften up a few outer microns of coating.

MEK does work, but you need to be thinking in terms of leaving it for months, not days. Acetone might be a bit quicker.