Disassembling/recylcing ferrite transformers

[Buriedcode]

I searched the forum and found a few posts, about this, and whilst I can just buy cores sets, I've having trouble sourcing the appropriate formers (EFD15 SMT).  This is a hobby project, so out of curiosity I thought I could try and salvage the cores and former from some surplus ferrites I had in stock, that I believe were for PoE.   I've done this before with some hefty ATX PSU ferrites with success.

Again, this isn't for the sake of saving money, its become more of a interest/challenge.

Unfortunately, rather than just varnish, whoever manufactured these employed black epoxy-like substance to fix the cores, a blob at one end, and most likely inside the former for the centre legs.

Tried so far:
- Soaked in White spirit ("mineral turpentine" ?) for 48 hours:  seems it attacked the enamel on the copper wire, but nothing else.
- Soaked in acetone for 48 hours: seemed to soften the black crap slightly, so I could scrape some of it off, but thats about it.
- Hot air started at 120C, and tried several times in increments up to 170C.  Each time spent a good 5-8 minutes heating the entire thing:  nothing budged.
- Oven at 180C then 200C for 15-20 minutes.  The insulation taped curled, and former softened to the point of breaking off.

Again, this is a lot of effort just for some SMT coil formers and cores, but once I start something I find it hard to let go 

Anyone have any other things I could try?  Right now I'm thinking, as I have 8 of these, half I'll break the ferrite to get the former, and the other half destroy the former to get the cores.

[Benta]

The only thing that really "bites" on epoxy is acetic acid.
But you'll need an outdoor place to use it.

[TimNJ]

Usually I use some hot air and then give it a bit of a twist to break the epoxy in the center leg...Unfortunately about a 50/50 chance of cracking the core.

Plus, with ferrite, you need to be careful about rapid heating/cooling and mechanical stress, in general. You may accidental introduce micro-fractures that can seriously affect permeability, and magnetic performance in general. We had a quality issue with MnZn ferrite common-mode chokes failing, caused by the FR-4 separator used to maintain isolation between the two windings. If tolerance of FR-4 or toroid was off a little, the FR-4 put too much stress on the ferrite and caused the inductance to drop by over 50%! Solution was to use a more compliant material, soft-ish plastic.

For hobby use, these things don't matter as much, but in general, I usually don't bother re-using ferrites because it's hard to know what kind of effect my prying, heating, and so forth may have had.

[julian1]

I have read about boiling in a water-filled saucepan for 20mins - 30mins. It would be easy to try - and help resolve if it is epoxy. 

[Jay_Diddy_B]

1+ boiling in water.

I suspend the transformer on a piece of wire supported be a stick over the pan.

It may depend on the type of varnish used originally.

Jay_Diddy_B

[Wolfram]

I've had good success buying cores and bobbins from Semic in the Czech Republic. Unlike most other electronics distributors, they have a good selection of cores and bobbins at a decent price. I didn't find any EFD 15 bobbins, but they have EFF 15 ones listed which seem to be very similar or identical https://www.semic.info/lj-eff1505h-10p-1s-ff-2/

[T3sl4co1l]

Also, methylene chloride.  It will swell epoxy...and strip most everything else, too.

Heh, I wouldn't recommend glacial acetic acid (GAA), it will dissolve the core.  It's ceramic, sure, but it's not an inert silicate ceramic like porcelain, it's literally ferrite -- that is, ferrate(III), a compound of metals and ferric oxide.  Possibly it doesn't dissolve quickly (possibly, the salts don't even dissolve in GAA themselves, so it kind of passivates -- though that seems quite unlikely to me), but it will be etched regardless, I think.

Cores and bobbins are also available from the usual suspects (Digi-key, Mouser) and probably other more regionally specific distributors.  Failing that, magnetic suppliers are usually somewhere, broadly local, or continental at least, let's say.  I know of several in the US, there ought to be more than a few in EU.

And failing that, well, there's always Ali Express... don't get your hopes up of course, expect everything you order to be a scam, pay accordingly.  When received, set up some tests to see if it's as expected.

Tim

[Buriedcode]

Thanks lads,  I'll try the boiling water, as pointed out, its simple enough to try.

The only acetic acid I have to hand is.. distilled vinegar, which I believe 4% so probably far too weak,  But I suppose if the boiling thing does nothing I could try it. 
I managed to get one core out of one (with the other core breaking....) that still has the goop on it, so I could use that as a test.

The cores themselves, I found then on farnell - gapped and ungapped pretty cheap. But the SMT formers aren't there.   I might indeed try aliexpress - just for the formers - because their specs aren't as critical as the cores (just dimensions really).  I have used it before, but mostly for generic parts.

[chris_leyson]

1+ boiling in water. Won't reach the curie temperature and is OK with phenolic bobbins. Works almost every time for varnish impregnated transformers but you might have to apply a little leverage for epoxy glued cores. Farnell and/or RS components stock EPCOS TDK and/or Ferroxcube EFD15 smt bobbins, they come with 8 pins. EDIT Haven't bought any smt EFD15 bobbins in a while and hadn't noticed that RS have dropped them. EDIT Farnell 1422727

[coppercone2]

keep in mind if you remake it you need special glue for ferrite, particularly if you pop off the bobbin top or bottom. I thought it was paranoid bs but I made long inductors with superglue and I had fails that were uncharacteristic of super glue, those same joints made in the same way with any other material would not fail nearly so much. Same reason why it does not work well with glass. You need an acrylic glue. Actually not sure if epoxy works well for ferrite.

Superglue holds it really good, for some short amount of time, then it fails for no good reason, at room temp, with the slightest flexing. I suspect that it would benefit from acrylic epoxy rather then standard epoxy also.

When you look at the price of those acrylic glues it becomes less appealing to make your own. Also not sure how well the bond between the wire and the core holds up with age and thermal cycling, so it might stay together and look good, but you might have decreased thermal performance because the core can't sink heat because of some fracture line (that you can't detect because its solid on top). Almost feel like you need to give ferrites some sort of coating before you pot them with normal TC glue, before you wind and impregnate, for ultimate stability and high load performance and durability.

Not sure how well transformer varnish works on ferrite cores. I see the more expensive ferrite cores come with a enamel or something (hopefully done right at the factory).

i.e. glass glue sold in the hardware store is a different type of super glue that smells the same.

[Buriedcode]

Well, I boiled one of them for 40 minutes.  There were some sort of fatty deposit that came off and sat on top of the water - I'm guessing a varnish - but the cores and former didn't budge.

As I had fracture the cores of one, I had a former, so I thought I'd cut away the former of another to get the cores, then experiment with heat.  Turns out 220C was the magic number, I had tried 200C hot air for a good 5 minutes, nothing, but about 30 seconds into the 220C they came apart, with some light mechanical persuasion.   That temperature is way beyond what the former could have coped with.

As farnell do the cores, I think I'll just break the ferrites off a few and use their formers.

Re: superglue.  I've found it to be so brittle that anything that remotely flexes/expands doesn't last long.  I generally use two part epoxy for most things, but as this thread has highlighted, it can be  a nightmare to "undo".  I might look into some silicone adhesive.

[coppercone2]

There are three types of common super glue, methyl, ethyl, and butyl. You might be surprised to know the de-facto way of fixing a flexible o-ring is with.. super glue. An O-ring is made of elastic rubber. But they sell a special type of super glue for it.

I think you will be surprised what proper glue choice and primer allows for. I mounted a microwave oven transformer to rubber with rubberized super glue (locite prism black) to rubber and then rubber to a steel plate (vibration), it stood up to a 3 pound maul (that is the hammer that looks like a octagon mounted to a handle) after I messed the project up. I can glue polyethylene to things so strong the poly rips before the glue joint does.

I don't think epoxy is great for glass either. I think you need an acrylic epoxy to do it right. Not sure what the correct engineering method of bonding glass to metal is (i mean 2 part standard epoxy might be decent but I don't think its the greatest bonder).

https://www.acehardware.com/departments/paint-and-supplies/tape-glues-and-adhesives/glues-and-epoxy/10581
That is a butyl super glue meant for glass work. The normal ones are ethyl type.

[Buriedcode]

Just to finish this thread off.  It seems for the transformers I have - some with black epoxy, some older with varnish, the only things that work are heat and acetone.  Acetone doesn't do anything to the epoxy, at least it didn't for the epoxy used by the small transformers, but it did seem to dissolve the varnish on the older ones - specifically the ones from old ATX power supplies.  These were left to soak over a couple of days.

I found a bunch more for tests - not for saving money as the larger EFD20 are pretty damn cheap anyway, and it just confirmed the 220C figure.  200 wasn't enough, even after a good 5-7 minutes of heating with hot air, nothing budged. After a minute of 220, I could gently pry them apart, without destroying the former, or fracturing the cores.  The former plastic does get soft at this temperature, so best to be gentle.

Note this "220C" is what my cheapy 898D is set to.  I'll have to check the actual temperature with a thermocouple.  And as its hot air, this probably isn't the temperature the ferrite gets to.

[mag_therm]

I see : 2 new cores for US$0.41
They probably go to landfill eventually anyway.
Can such a supplier link be posted here?

Edit: the same vendor has 2 bobbins for $0.56

[coppercone2]

try nitromethane for epoxy

https://www.zoro.com/loctite-remover-175-oz-bottle-clears-235018/i/G2420013/feature-product?utm_source=google&utm_medium=surfaces&utm_campaign=shopping%20feed&utm_content=free%20google%20shopping%20clicks&gclid=Cj0KCQjwtsv7BRCmARIsANu-CQdsVJA7KtxlKk8C0LUTWIqNRJOTCzw4RhFXE11JJttNk4X1y8JeTlAaAqviEALw_wcB

Also, some people recommend MEK. You leave it submerged for a while. I think its a bit safer then acetone (smells really bad, less volatile, slightly harder to ignite).

[cdev]

Where can we get ferrites like the kinds we get from Amidon, Fair-Rite, for cheaper, maybe from China or Russia/Eastern Europe? For cheap?

ferrites especially but also iron powder cores.

For use in antenna baluns, tuners, filters, to suppress RFI, for example?

[coppercone2]

maybe no one makes them because the market is small and people are reluctant to put filters on stuff in the first place

I think there will be plenty for switching converters, but for filters I am not so sure

[TimNJ]

http://www.cwsbytemark.com/

https://www.mhw-intl.com/


Not sure if considered cheap but hey, at least you can get them?

[Oleander]

Instead of water, place the trafo in a (used) engine oil and heat it slowly well beyond 100C.
Worked 100% for me several times.