[SOLVED] How to remove the dreaded "yellow glue of death"?

[cigmas]

There is a yellow glue used presumably used for supporting large components, which becomes a problem over time by turning brown and becoming conductive and corrosive. This is a common problem on CyberPower UPSes, and is documented in a video:



What are the best ways to remove it? I have tried as much as possible to remove it mechanically from the most common failure point which just started turning brown, as seen in the images below, but I am afraid of going further, for fear of damaging components, breaking solder joints or pads, or slipping and scratching the PCB or traces.

I tried soaking the glue in 99% isopropyl alcohol, and also put a piece of the glue in nail polish remover (mostly acetone with a bunch of oils and fragrances added), but neither seemed to do anything.

[BravoV]

Get a drink, go and read my discussion thread below ...

 -> How to remove/clean off these annoying glue easily ?

Click image to enlarge.

[Veketti]

One other thing you might consider testing is WD40. I can't tell whether it works for this or not, but for example if you have to remove double sided tape, just soak WD40 and it comes off easily without mechanical rubbing.

[Whales]

> I tried soaking the glue in 99% isopropyl alcohol, and also put a piece of the glue in nail polish remover (mostly acetone with a bunch of oils and fragrances added), but neither seemed to do anything.

I was thinking of suggesting acids and bases as well as solvents, but alas they'll eat the metal faster than the glue.  Soaking in water might be worth trying (it might get between the adhesive and other plastics, slowly separating them) but it might infiltrate the relays or possibly the electrolytics.

Perhaps it's a polyurethane?  In that case you might be SOL, they tend to be very chemically stable and insoluble.

Do you have a soldering iron tip shaped like a flat bladed screwdriver?  Perhaps setting that to something low might help when chiselling it off.

[jpanhalt]

I wouldn't consider just one method for every type of mechanical stabilizer:
1) If it is hot snot (i.e., a hot melt), which is very common, isopropyl alcohol (IPA) is the answer.  Those fixants are generally whitish/clear but there is no reason they cannot be colored (e.g., black).  They also do not have a detectable odor.
They are flexible and once IPA is added, they peel right off.
2) If it is a rubber, like Pliobond, which is tan an has a distinctive odor and is light tan as the TS shows, freezing might be the answer.  One should be able to clean up with acetone or MEK.  MEK is less volatile and is what I would use.  Chlorinated solvents (e.g., methylene chloride, chloroform, carbon tetrachloride, PERC) would also help for clean up.  Try a chlorinated brake cleaner.
3) If it is a rubber like silicone, the only reasonable approach is mechanical, but freezing might help.  Final clean up with a mixture of IPA and light mineral spirits will help.  That's what I do with silicone caulk.
4) If it is hard, then I suspect an epoxy, polyester, or acrylic.  Carefully applied heat would be my approach, but I have not run into those materials on PCB's. (EDIT2: Except as tiny dots to affix components during assembly.)

Edit: There are simple tests to identify polymers.  They may include "flame tests."  Lots of detail is on the web.

[coppercone2]

man that thread you linked is hilarious, it reminds me of the movie 'the blob' or something. Like there is a menace and no one can figure out how to deal with it and you see the government try all sorts of whacky shit like freeze guns.

I wonder if this is why the old APC battery I had broke, I checked so much stuff but the glue was unassuming. Wow, if only I had a time machine ....

that UPS failure made me like shy away from a entire field of engineering because I felt like it was a lost cause, that its too complicated to get right. I wonder if its as stupid as the glue. Its like mad evil.

and I suspect jpanhalt forget #5 : currently unknown substance in use as trade secret 

btw thanks, maybe I can give having battery backup another shot. I began to disregard it as a bogus technology lol

Maybe UV exposure and freezing?!

[Terry Bites]

It looks like a urethane. If so use MEK, the safer less toxic and less agressive alternative to acetone.

[vk6zgo]

I had to remove some really strange gunk, which seemed to be a compound of adhesive & disintegrated rubber carpet underlay from a wooden floor.

The Hardware shop sold me some really nasty stuff, (maybe toluene,--can't remember) that you could only apply for 10 minutes at a time, then you had to leave the house, leaving the doors & windows open to air it out.
I struggled on with it, but it was useless----melting the "gunk", but just making it bond harder to the floor.

I eventually tried using "Mineral Turpentine" which didn't melt the stuff, but penetrated between it & the wood, allowing it to be removed with a spatula.
The Turps really stunk, but it isn't near as toxic as the original stuff I tried.

Much more recently, I had occasion to remove a bunch of defunct nicads from a battery pack belonging to a handheld ham radio.
They were held in place with the yellow crud.

Having had a bad experience trying to remove this stuff before, I decided to try Turps, so using a 500g Coffee can, I dropped the assembly inside, poured in enough of the stinky stuff to cover it, & left things for several weeks.

On removal, the nicads could be relatively easily extracted from the clutches of the glue, & enough of the latter removed from the case to allow it to be re-loaded with new cells, although I haven't got round to that, instead just "bodging" two 9v alkalines in parallel, to allow me to use the radio.

I don't know what adverse effects Mineral Turpentine would have on more fragile electronics, nor do I know what it is called in other countries.
All in all, the method decided upon in the link posted by Bravo might be the best bet, as freezing seems a bit gentler.

[Jwillis]

The description appears to me  like  Neoprene (Chloropene) . Super cheap crap becomes acidic after a while when it  goes brown and cracks when long term heat is applied. Hexane disolves it and pretty much everything else. Nasty  stuff. I heard that freeze spray works to make it brittle enough to break away with a little work.   

[cigmas]

Thanks for all the great suggestions!!

I tried lighter fluid on a piece of the stuff, and it slightly softened it permanently, without making it sticky. So I used a cotton swab to apply many rounds of the fluid on the spots of glue, and used ceramic-tipped tweezers, small flat blade screwdrivers, an old credit card, a dental pick, a thumb tack... anything I could think of, to get a grip on the glue and pry it off.

The lighter fluid gets it off insulated wires quite easily, and makes the glue stretchy and loose, but it doesn't seem to penetrate the glue too deeply, so it takes many iterations to get larger clumps off. I also found that lighter fluid has an effect when the glue is yellow, but seems to do nothing where it has turned hard and brown.

I got the USB board (not shown in pics) almost spotless clean, and made some progress on the other spots, but it is very slow going. I'll keep at it, but some areas are really difficult for access and even when softened and stretchy, the glue is difficult to grip well and remove from the surfaces, as opposed to just pulling a small piece off the top.

Freezing it sounds interesting, but I'm kinda scared to try the butane trick..

As for acetone/MEK, is that worth pursuing if nail polish didn't do anything? Similarly, IPA seemed to do little to nothing.

I do have chlorinated brake clean, WD40, and other oils/lubricants. Thoughts on how safe those are on electronics?

[cigmas]

It took many hours over several days, but it's finally done:





Lighter fluid / naphtha helps soften it, but ultimately it took dentistry to get the stuff off.

I found the ceramic tipped tweezers helpful as a hard scraping tool especially for getting between the glue and PCB without scratching like a metal tool would. Getting under the glue and prying it off a surface was great when it worked.

Also very useful was the dental pick for getting under capacitors and into tight spaces and pulling the glue away gently but firmly. For the browned and hardened glue, the sharp point was useful for wearing it down. For the really hard stuff, the only success I had was by pressing down with the point of the pick horizontal to the surface. The pressure would cause a small piece to fracture and fly off.

Afterwards generous amounts of IPA were used to clean up and avoid leaving residues.

Overall, a royal pain, and I'd rather never have to do this again. Hopefully there's no damage to components or traces and the UPS can live a long life.

Here are the tools used:

[5U4GB]

It looks like a urethane. If so use MEK, the safer less toxic and less agressive alternative to acetone.

Yeah, polyurethane or some type of epoxy would have been my guess as well, it doesn't have the mozarella look of hot-melt glue.  For the OP, if you can get Gunwash, technically a lacquer thinner but in practice a cocktail of carcinogens (MEK, acetone, xylene, IPA, toluene, and others, the MSDS for it has its own MSDS), you could try that on it, just not indoors.

[BravoV]

In my case, acetone (mine was > 95%) was useless, not even soften that glue. 

[coppercone2]

you can try epoxy remover

[5U4GB]

you can try epoxy remover

Epoxy remover is typically acetone which I think the OP has already tried.

[coppercone2]

no, some of them have more advanced formulations. THF, Nitromethane, etc.

I was amazed by super glue remover from the hardware store, I think it was THF based, wow... it actually removed it. The other stuff seemed like a struggle. I won't use super glue unless I have the little thing of goof off brand super glue remover handy.

If you get it on your body by accident, you can actually dab it off with a wipe. I mean you should be wearing gloves but accidents happen. I Had some seriously bad experiences with super glue in the past, particularly the fluid one.

[5U4GB]

I Had some seriously bad experiences with super glue in the past,

Picking your nose after getting it on your fingers is a definite no-no.

I've actually found the very liquid super glue to be quite useful for stabilising cut edges in plasterboard/sheetrock, if you need to cut a small hole or slot in it the liquid will soak into the gypsum plaster and form a plastic-rock-like edge that doesn't crumble over time.

[Master IPC Trainer]

Kudos to CIGMAS for posting this.
I earn a good living training electronics manufacturing personnel ranging from quality inspectors, management to process workers in both the IPC-A-610 and J0STD001 quality standards used almost universally by the world's top electronics manufacturers. https://www.ipc.org/ipc-standards.

Being IT related products, UPS's should be built to IPC J-STD001 and inspected to IPC-A-610G class 2 standards where the use of adhesive is mandatory for mechanically supporting (bonding to the PCB substrate) heavy thru-hole parts to eliminate the risk of lead breakage or solder fracture occurring. Not only must a suitable adhesive (meeting IPC-SM-817) be used, but it must be in the correct places and NOT in other places. IPC has specific standards covering this.
 
Failures of the adhesive are visible in the photos in the vicinity of the fuse and R47.

Firstly, adhesives meeting IPC-SM-817 should not become conductive and fail like this. The yellow adhesive is of unknown type and therefore may not meet this standard. The heat dissipated by R47 is the likely catalyst however the real culprit lies elsewhere as conductive adhesive shouldn't be in places where conductivity  matters.

Secondly, while IPC standards have a requirement to support the tall electrolytic capacitors nearby, small beads of adhesive should be placed around the circumference of the cap adhering to the PCB substrate alone and not be placed on areas to be or which have been soldered as appears to be the case.

It is a manufacturing DEFECT under all IPC classes for any bonding adhesive to be present in areas required to be soldered. There is also no purpose in covering the numerous small adjacent parts with this material unless the intention is to obscure the components. Does CyberPower's supplier claim their products meet IPC standards or is CP an ISO9001 quality certified company ? Their website doesn't say.....  I think it should.

[coppercone2]

you know what it probobly is, there is a faulty or crappy glue dispenser. The good ones have a kind of back lash where after they dispense it retracts because the container is flexible, for caulk guns. I can imagine all sorts of funky stuff going on because someone does not want to tune up some valves or whatever and the factory does not wanna fix it. At least that is part of the problem.

Semi-stuck/dry caulk gun + low pay employee. Like where the tube dried up and should have been disposed of two months ago but someone stabbed it with a screw driver and squeezed it out so there is no waste. Factories are atrocious with dealing with adhesives if its not their main stay.

Because if there is a 1/3 of the cylinder left uncured and the manager gave it squeeze and it was still flexible you know its gonna get stabbed on the side and someone is gonna try to bomb the board with whatever comes out and shovel it into place with a flat head.  And its gonna rip the opening while its being squeezed out side ways and get glue on someones sleeves and then they will make a bigger mess.