Rewinding burnt choke coil

[PeteS in CA]

Re: Rewinding burnt choke coil

Hmmm. Good point!

Thanks for posting that link - I learned a lot of new things from it today.

So according to that article, increasing the number of turns decreases the core loss?... at least in the last example they showed with the APFC coil.
- Very interesting!

I'll probably read it a few more times just to make sure I didn't miss something. We did some simple boost and buck regulator calculations back in one of our classes in university, but I've forgotten a large part of that. I don't think we got anywhere near as deep as core material or anything like that. Mostly it was just calculating the inductance for the coil needed based on current and voltage requirements and whatnot.

That article's been on Micrometals' website for as long as I can remember. No author is given, but I suspect the/an author may have been Dale Nichol, an applications engineer there for many years. Before Micrometals Dale was the magnetics designer-guru at Boschert back in the late 70s and early 80s.

[momaka]

Re: Rewinding burnt choke coil

I believe we are discussing an output inductor, which would have a powdered iron core, not a ferrite core. That powdered iron cores can lose their magnetic properties over time by being run too hot is not news. Here's Micrometals' webpage on the topic.

Hmmm. Good point!

Thanks for posting that link - I learned a lot of new things from it today.

So according to that article, increasing the number of turns decreases the core loss?... at least in the last example they showed with the APFC coil.
- Very interesting!

I'll probably read it a few more times just to make sure I didn't miss something. We did some simple boost and buck regulator calculations back in one of our classes in university, but I've forgotten a large part of that. I don't think we got anywhere near as deep as core material or anything like that. Mostly it was just calculating the inductance for the coil needed based on current and voltage requirements and whatnot.

[PeteS in CA]

Re: Rewinding burnt choke coil

I believe we are discussing an output inductor, which would have a powdered iron core, not a ferrite core. That powdered iron cores can lose their magnetic properties over time by being run too hot is not news. Here's Micrometals' webpage on the topic.

[Dannyx]

Re: Rewinding burnt choke coil

For curiosity's sake, here's the choke in question off the board. I hadn't unwound it at that point. True, it's hard to get a sense of scale since there's no ruler beside it, but what I know for sure is the wire is 0.3mm in diameter and you can kinda get an idea by comparing it to other components on the board...can't remember the other dimensions of the rings right now, though I DID measure them. May post those too at some point.

[momaka]

Re: Rewinding burnt choke coil

The core material has probably lost its properties.

I've heard the same, but I am not entirely convinced myself. Ferrite cores are often painted and then baked in an oven to cure the paint. As for the coating that has peeled on Dannyx's core, that's probably plain varnish insulation, painted on after the coil was wound with wire (to help prevent vibrations of the wire and audible oscillations).

I've successfully rewound two small buck-regulator toroid cores from eVGA GeForce 7600 GT video cards that got burned due to bad Sacon FZ caps (which caused the RAM to short-circuit and pull a massive surge through the coil in the buck regulator until it overheated).

After rewinding the coil (and changing the RAM chips!), I got the buck regulator to at least output the right voltage without the coil or anything else overheating. Unfortunately, the GPU core must have been damaged too, as I get artifacts with the new RAM chips. But hey, at least I have video output now!

Can I take some 0.3mm enamel copper wire and just rewind the thing ?

Yes.

Just unwind the old wire and measure its length. Then rewind with new wire. As long as the core hasn't gotten an internal crack from all the heat, it should work fine again with the new wire. However, if it is cracked, you will likely get an overheating core again. So it's worth giving this a shot, if you want to do it just for an experiment.

Otherwise, if you really need to fix the PSU professionally, then you find the same core size and type online, and wind it with your own wire (as finding an already wound core with the same number of turns and thickness of wire is unlikely).

Also, no need to coat the wire in varnish - at least not for testing. But if you find the PSU is making nasty whining noises, you might have to do that. Wood varnish isn't the most ideal substitute, but it should do the trick.

The DC output caps failed and that caused too much stress on the choke.

Yes. But they likely failed only because the fan seized.

The only time I've seen a burned coil without the fan getting seized and the caps cooked (or showing bad ESR) was in a really cheap PSU, where the output toroid core was just too small, along with the wire wound on it was too thin.

[PeteS in CA]

Re: Rewinding burnt choke coil

I'm not sure if this helps: http://imajeenyus.com/electronics/20...ge/index.shtml
Looks like a similar coil
and I found this http://transformers.en.hisupplier.co...oke-coils.html

I'm going to guess that the two stacked cores in that inductor are -52 material, based on this Micrometals chart. This chart compares core losses for the various materials.

The inductor in that second link is a common mode inductor, a very different function, and it has a ferrite core.

[Dannyx]

Re: Rewinding burnt choke coil

I'm not sure if this helps: http://imajeenyus.com/electronics/20...ge/index.shtml
Looks like a similar coil
and I found this http://transformers.en.hisupplier.co...oke-coils.html

Yeah the first one looks very much like what I've got and so does the coil. Apparently it's greenish-yellow...doesn't matter too much at this point since it's obliterated and beyond repair, but perhaps I could still find a replacement from China somewhere...

[R_J]

Re: Rewinding burnt choke coil

I'm not sure if this helps: http://imajeenyus.com/electronics/20...ge/index.shtml
Looks like a similar coil
and I found this http://transformers.en.hisupplier.co...oke-coils.html

[Per Hansson]

Re: Rewinding burnt choke coil

Going without a coil sounds like an idea for using this thing in low-precision applications like I dunno, powering a fan or light something

I was not proposing to go without a coil.
What I meant is you could use it as a learning example if you get a scope:
That is to fine tune in a coil that makes the unit behave it's best.

Though looking at the new pictures you posted I'd for sure call it a goner!

[stj]

Re: Rewinding burnt choke coil

about the colour, there is a vid on utube of one being repaired.
search "cheap psu" or something like that.

[redwire]

Re: Rewinding burnt choke coil

Hahaha, that PSU is so baked even the solder melted on the underside, wow!
The PCB topside needs to be fixed first, if it's too bad it's not worth it IMO.
And you can throw the coil away: the core is toast and needs to be replaced.
But just jerry rigging it would work, it requires s scope for testing though.
I would recommend you get a analog scope on eBay instead of some crap digital thing as you mentioned.
It will allow you to learn allot more, and get a deeper understanding.

P.S: This PSU failed because of the use of general purpose caps, the switching converter UC38458 it uses supports operation at 500Khz!

+1
this thing is burned badly, it's dead it's dead, lol.
I would expect the main transformer to also be damaged, from heat. Look at it closely because your time to fix it may be wasted.

[Dannyx]

Re: Rewinding burnt choke coil

I've yet to provide exact measurements of the coil and I don't have it on hand right now (friggin' busy day), so leaving color aside, another thing I need to mention, though it should be visible in the pictures as well is that it's actually got TWO rings stacked back to back and wound together....don't know how that affects parameters and calculations but I thought I'd point it out.

[budm]

Re: Rewinding burnt choke coil

I use Micrometal -26 in a lots of our products.

[PeteS in CA]

Re: Rewinding burnt choke coil

I BELIEVE the coil was yellow, though it had already started to peel off when I first got it, so that's just a hunch....doesn't matter anyway.

Mostly yellow with one face being white would be Micrometals -26 material. -52 is the same permeability but lower core loss above about 50KHz. The two materials can be used interchangeably at 20-30KHz though -52 is a bit higher in cost. I think -52 was introduced in the late 80s or early 90s; -26 has been around since at least the late 70s.

[Dannyx]

Re: Rewinding burnt choke coil

What brand (and Series) are the other capacitors that were not replaced?

If you're referring to the caps on the secondary, ALL of those have been replaced in the same go, so I can't remember what the originals were. You're only seeing two left in the pictures because I didn't bother removing them, though they failed at the same point. If you mean other capacitors on the whole board, including the primary, those are just chinese caps not worth writing home about.

[Stefan Payne]

Re: Rewinding burnt choke coil

What brand (and Series) are the other capacitors that were not replaced?

[Dannyx]

Re: Rewinding burnt choke coil

Just for fun, here's some more photos of the board, the rectifier and the switching device...I don't know how good this thing was even when new, Dave of EEV would probably call it trash straight away, obviously but it was apparently just fine for powering CCTV cameras

[Dannyx]

Re: Rewinding burnt choke coil

The caps adjacent to that inductor are rated for 16V. It's an output inductor. The input common mode inductor is that E-core inductor near the input fuse.

Output inductors are typically powdered iron, not ferrite. Powdered iron cores lose their magnetic properties permanently if they are too hot for too long. Micrometals, the major manufacturer of powdered iron cores, has information about this on their website. That the paint peeled off the core is a good indicator that the core is probably no good. Also. unless you saw what color paints were where on the core, you could only guess what material type it was (I'd guess Micrometals -26 or -52, but that is nothing but a decent guess).

All in all, that power supply is probably good for spare parts and trash can filler.

I BELIEVE the coil was yellow, though it had already started to peel off when I first got it, so that's just a hunch....doesn't matter anyway.

[PeteS in CA]

Re: Rewinding burnt choke coil

The caps adjacent to that inductor are rated for 16V. It's an output inductor. The input common mode inductor is that E-core inductor near the input fuse.

Output inductors are typically powdered iron, not ferrite. Powdered iron cores lose their magnetic properties permanently if they are too hot for too long. Micrometals, the major manufacturer of powdered iron cores, has information about this on their website. That the paint peeled off the core is a good indicator that the core is probably no good. Also. unless you saw what color paints were where on the core, you could only guess what material type it was (I'd guess Micrometals -26 or -52, but that is nothing but a decent guess).

All in all, that power supply is probably good for spare parts and trash can filler.

[Dannyx]

Re: Rewinding burnt choke coil

Hahaha, that PSU is so baked even the solder melted on the underside, wow!
The PCB topside needs to be fixed first, if it's too bad it's not worth it IMO.
And you can throw the coil away: the core is toast and needs to be replaced.
But just jerry rigging it would work, it requires s scope for testing though.
I would recommend you get a analog scope on eBay instead of some crap digital thing as you mentioned.
It will allow you to learn allot more, and get a deeper understanding.

P.S: This PSU failed because of the use of general purpose caps, the switching converter UC38458 it uses supports operation at 500Khz!

The area underneath the coil which sustained the most damage is just a whisker away from poking all the way through, so there's no repairing it since there IS no trace left. This thing will never go past the "Frankenstein" state as it stands, but it's not like I'm investing time and money to repair it, which is why I'm trying to patch it rather than buy stuff to fix it

Going without a coil sounds like an idea for using this thing in low-precision applications like I dunno, powering a fan or light something