Blown inductor - repair feasible? (HD3870)

They make noise, too.

Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee

Post Reply
Cistron
Posts: 618
Joined: Fri Mar 14, 2008 5:18 am
Location: London, UK

Blown inductor - repair feasible? (HD3870)

Post by Cistron » Fri Aug 01, 2014 1:22 pm

Hi silent folks!

Yesterday my graphics card gave up the ghost (in the middle of a Dota session!). There was a somewhat nasty smell, a few seconds before it gave.

At least I think it did, as the PC refused to boot. And according to Abit-beep-speech *beeeeeeeep-beep-beep* means graphics card is naughty.

Finally found the time today to take off the cooler and found an inductor rattling underneath it. At least that what I think it is (says Magic R60 on the side). Though it doesn't seem to have been damaged - see (grainy) photograph.

Well, the PCB doesn't look to bad either, though a bit of the top-coat has come off, revealing the shiny coppery insides. I've highlighted where the inductor used to sit (according to other images of the webz).

Feasible to re-solder it? I thought it might have been a dry joint that finally heated up too much. And the nasty odour was a result of that.

But since I have zero experience with this, I thought I'd shoot a question into the blue.

The PC is 6 1/2 years old, but still does it's duty. Buying a new card would be kind of stupid for such an old machine. Now my loan Quadro FX550 is driving me mental (doesn't do 1440p and the tiny fan buzzzzzzzzzes).

HELP! :)

Cheers,
Mike
You do not have the required permissions to view the files attached to this post.

sipitai
Posts: 46
Joined: Mon Jun 20, 2005 7:08 am
Location: Perth, Australia

Re: Blown inductor - repair feasible? (HD3870)

Post by sipitai » Fri Aug 01, 2014 8:11 pm

Hi Mike,

Yeah I'd say go for it.

A few issues off the top of my head:

1). If you've never done soldering before, it might be worth doing a little bit of practice beforehand. Or getting someone with soldering experience to do it for you. You want to make sure you form a solid join and don't accidentally connect any nearby tracers. That said the job is a simple one, so there's no harm in just giving it a go if neither of those options are feasible. Just look up some how-to videos on Youtube and make sure you wear eye protection.

2). I have no idea how heat sensitive the component, and those around it, are. It's possible, although I'd consider it fairly unlikely, that you could do some damage if you let things get too hot. The trick here is to make sure you're not applying heat to the join for any longer than you need to - a few seconds max ideally. That said you might be fine either way.

3). Given the component fell out during operation, other parts of the card may have been damaged. It's possible that even if you repair the component properly, the card still might not work. Hopefully this wont be the case, but its a possibility.

Anyway that's my 2 cents. You might want to wait and see what other users here have to say.

Good luck with it all :)

Cistron
Posts: 618
Joined: Fri Mar 14, 2008 5:18 am
Location: London, UK

Re: Blown inductor - repair feasible? (HD3870)

Post by Cistron » Sat Aug 02, 2014 12:55 am

sipitai wrote:Hi Mike,

Yeah I'd say go for it.

A few issues off the top of my head:

1). If you've never done soldering before, it might be worth doing a little bit of practice beforehand. Or getting someone with soldering experience to do it for you. You want to make sure you form a solid join and don't accidentally connect any nearby tracers. That said the job is a simple one, so there's no harm in just giving it a go if neither of those options are feasible. Just look up some how-to videos on Youtube and make sure you wear eye protection.

2). I have no idea how heat sensitive the component, and those around it, are. It's possible, although I'd consider it fairly unlikely, that you could do some damage if you let things get too hot. The trick here is to make sure you're not applying heat to the join for any longer than you need to - a few seconds max ideally. That said you might be fine either way.

3). Given the component fell out during operation, other parts of the card may have been damaged. It's possible that even if you repair the component properly, the card still might not work. Hopefully this wont be the case, but its a possibility.

Anyway that's my 2 cents. You might want to wait and see what other users here have to say.

Good luck with it all :)
Thanks for your comprehensive advice sipitai! :)

Cistron
Posts: 618
Joined: Fri Mar 14, 2008 5:18 am
Location: London, UK

Re: Blown inductor - repair feasible? (HD3870)

Post by Cistron » Mon Aug 04, 2014 10:23 am

Well, I soldered the damn thing back on and it revived the card! Yeah!

A few things of note. Desoldering holes is a real pain. And soldering honking huge inductors takes a lot of heat.

So think for my first soldering job, I was thrown into the deep end.

Vicotnik
*Lifetime Patron*
Posts: 1831
Joined: Thu Feb 13, 2003 6:53 am
Location: Sweden

Re: Blown inductor - repair feasible? (HD3870)

Post by Vicotnik » Mon Aug 04, 2014 10:43 am

Sweet. :) I thought is was a lost cause. Well done!

xan_user
*Lifetime Patron*
Posts: 2269
Joined: Sun May 21, 2006 9:09 am
Location: Northern California.

Re: Blown inductor - repair feasible? (HD3870)

Post by xan_user » Mon Aug 04, 2014 3:38 pm

:mrgreen:
awesome!
:mrgreen:

CA_Steve
Moderator
Posts: 7650
Joined: Thu Oct 06, 2005 4:36 am
Location: St. Louis, MO

Re: Blown inductor - repair feasible? (HD3870)

Post by CA_Steve » Mon Aug 04, 2014 4:19 pm

Congrats on the Lazarus job.
Cistron wrote:Desoldering holes is a real pain.
Back in my big soldering days, we had what was effectively a small turkey baster/ huge eye dropper kind of suction. Get the solder melty then suck it up. The end material was teflon or something similar to resist melting.

Cistron
Posts: 618
Joined: Fri Mar 14, 2008 5:18 am
Location: London, UK

Re: Blown inductor - repair feasible? (HD3870)

Post by Cistron » Tue Aug 05, 2014 12:10 am

Thanks guys, I was doing a little victory dance yesterday. Well, it was more of a Gary Newman chair dance, but you get the picture :mrgreen:
CA_Steve wrote:Back in my big soldering days, we had what was effectively a small turkey baster/ huge eye dropper kind of suction. Get the solder melty then suck it up. The end material was teflon or something similar to resist melting.
That sounds cool. I wish that sucker had been available. Working with copper wick was ... unintuitive - never remove the iron before the wick (learning by doing).

I recently read an editorial about companies specialising on component repair with some truly amazing desoldering contraptions. The way hot air heats up the parts depending on what exactly it is, etc.

There probably is quite a potential for micro-electronics repair, it just isn't done.

sipitai
Posts: 46
Joined: Mon Jun 20, 2005 7:08 am
Location: Perth, Australia

Re: Blown inductor - repair feasible? (HD3870)

Post by sipitai » Tue Aug 05, 2014 3:29 am

I'm a bit late to the party, but good to hear it's working again :)
Cistron wrote:There probably is quite a potential for micro-electronics repair, it just isn't done.
Yeah its a bit of shame, the economic viability just isn't there. The cost of labour is generally more than the cost of the replacement, so a lot of broken computer parts just end up in the bin.

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Re: Blown inductor - repair feasible? (HD3870)

Post by ~El~Jefe~ » Wed Aug 06, 2014 9:23 pm

that was an impressive fix :)

Very 70's / 80's of you. Hobbyist approach gets my respect :)

Cistron
Posts: 618
Joined: Fri Mar 14, 2008 5:18 am
Location: London, UK

Re: Blown inductor - repair feasible? (HD3870)

Post by Cistron » Thu Aug 07, 2014 4:20 am

~El~Jefe~ wrote:that was an impressive fix :)

Very 70's / 80's of you. Hobbyist approach gets my respect :)
Thanks, man!

As a precaution, I've added a variable resistor to the fan in front of the graphics card. I'll just turn it up for gaming (~900 vs 600 rpm), which keeps the maximum around 70 instead of 85 degrees.

~El~Jefe~
Friend of SPCR
Posts: 2887
Joined: Mon Feb 28, 2005 4:21 pm
Location: New York City zzzz
Contact:

Re: Blown inductor - repair feasible? (HD3870)

Post by ~El~Jefe~ » Thu Aug 07, 2014 9:08 pm

makes me wonder how many things i throw out that I could have tried something with.

I am always afraid of shorting and burning a socket on a motherboard. Nowadays though, the motherboard is like 50 bux and the cards are 200.

Cistron
Posts: 618
Joined: Fri Mar 14, 2008 5:18 am
Location: London, UK

Re: Blown inductor - repair feasible? (HD3870)

Post by Cistron » Sun Sep 21, 2014 4:23 pm

Riiight, so it seems the choke heated up high enough to free itself of the solder again. Bummer.

It's possibly that the part itself is running a hot. I'm not quite sure what can go wrong with a piece of wire, ferrite core and insulation, but hey... let's give it a shot. Now is not a good time for me to buy a new graphics card.

Does anyone know what the specifications for a replacement inductor would be? I googled around for quite a while (R60, hole through, inductor, etc.), but I'm not much the wise. I think 600nH (whatever that means) and the ceramic encasing is about 10.5mm square, 9mm high.

Zolishoru
Posts: 85
Joined: Sun Feb 06, 2011 11:19 pm
Location: North of the 49th parallel

Re: Blown inductor - repair feasible? (HD3870)

Post by Zolishoru » Sun Sep 21, 2014 9:22 pm

Solder it back, then cover it with RTV or conformal coating(I think the vibration breaks the choke free, so it needs some dampening).

Cistron
Posts: 618
Joined: Fri Mar 14, 2008 5:18 am
Location: London, UK

Re: Blown inductor - repair feasible? (HD3870)

Post by Cistron » Mon Sep 22, 2014 12:17 am

Thanks for your reply, Zolishoru.
Zolishoru wrote:Solder it back, then cover it with RTV or conformal coating(I think the vibration breaks the choke free, so it needs some dampening).
It did actually fall apart. I think it got so hot that the dampening material slowly deteriorated, which was likely the foul odours that preceded the first death. Now I have a spool and the case. Before the second death it buzzed badly and casing is now brown, not grey any more.

Does anyone know how precisely do I have to match the inductance of the inductor? As mentioned above the R60 implies 600 microHenry. I guess the wire gauge must be somewhere around 1mm, so 17-19 and 4 turns (or is it 3 1/2?).

Cistron
Posts: 618
Joined: Fri Mar 14, 2008 5:18 am
Location: London, UK

Re: Blown inductor - repair feasible? (HD3870)

Post by Cistron » Mon Sep 22, 2014 10:04 am

Progress!

After some time on google (sadly, hours, not minutes), I'm a little wiser.

I believe it is a shielded power inductor, listed as APL-PQ on this website: http://www.magictec.com.tw/magic%20web( ... 0Inductors

So it probably is 600uH and somewhere in the 10-38A rated range.

Now where to order one. All I seem to find are bloody surface-mounted ones.

Zolishoru
Posts: 85
Joined: Sun Feb 06, 2011 11:19 pm
Location: North of the 49th parallel

Re: Blown inductor - repair feasible? (HD3870)

Post by Zolishoru » Mon Sep 22, 2014 9:20 pm

You can mount back the original inductor without the casing, BUT:
1. Do damn good soldering, and
2. Once in place, cover the inductor completely with acid-free RTV.

Cistron
Posts: 618
Joined: Fri Mar 14, 2008 5:18 am
Location: London, UK

Re: Blown inductor - repair feasible? (HD3870)

Post by Cistron » Tue Sep 23, 2014 3:33 pm

Zolishoru wrote:You can mount back the original inductor without the casing, BUT:
1. Do damn good soldering, and
2. Once in place, cover the inductor completely with acid-free RTV.
The problem is, I don't have any RTV silicones. Maybe some of my friends can help.

In the meantime, I've gotten in touch with Powercolor and they've come back to me quite promptly. Sent them some photographs and hopefully they'll be able to sort me out.

Zolishoru
Posts: 85
Joined: Sun Feb 06, 2011 11:19 pm
Location: North of the 49th parallel

Re: Blown inductor - repair feasible? (HD3870)

Post by Zolishoru » Tue Sep 23, 2014 8:51 pm

Cistron wrote:[The problem is, I don't have any RTV silicones. Maybe some of my friends can help.

In the meantime, I've gotten in touch with Powercolor and they've come back to me quite promptly. Sent them some photographs and hopefully they'll be able to sort me out.
I hope that Powercolor will help, it would be ideal; until then here's some info:
I've had good experience with the DOW CORNING 748 non corrosive sealant; expect to spend at least 10 pounds to get some(enough to cover the coil).

Cistron
Posts: 618
Joined: Fri Mar 14, 2008 5:18 am
Location: London, UK

Re: Blown inductor - repair feasible? (HD3870)

Post by Cistron » Tue Sep 30, 2014 11:10 pm

So apparently PowerColor don't have access to their own data, because the archives are at some external warehouse. Hmmpf.

I guess it's time for a replacement card. I just wish the GTX960 were out already for the price drop on the used market.

WARDOZER9
Posts: 121
Joined: Thu Jan 28, 2010 7:41 am
Location: Rockford, IL ( USA )

Re: Blown inductor - repair feasible? (HD3870)

Post by WARDOZER9 » Sun Oct 19, 2014 9:29 am

If everything still seems stable, you need to find out why this happened so it doesn't come back in spades. It is possible and it has happened that this is due to faulty solder joints causing arcing but the other possibility is a sever surge due to faulty components up the line. You want to make sure that the PSU wasn't the cause of a possible surge because if that is the case, your whole system is at risk.

Cistron
Posts: 618
Joined: Fri Mar 14, 2008 5:18 am
Location: London, UK

Re: Blown inductor - repair feasible? (HD3870)

Post by Cistron » Thu Oct 23, 2014 2:11 pm

WARDOZER9 wrote:If everything still seems stable, you need to find out why this happened so it doesn't come back in spades. It is possible and it has happened that this is due to faulty solder joints causing arcing but the other possibility is a sever surge due to faulty components up the line. You want to make sure that the PSU wasn't the cause of a possible surge because if that is the case, your whole system is at risk.
I would assume that more than just inductor would be affected by a surge, however, the rest of the card looks quite fine.

I currently piece the event together that a bad solder joint caused heat-up, which over turn deteriorated the dampening and eventually the vibrations caused the inductor to break free. I might give the soldering another try and will insulate the component with high-temp hot glue (or such).

Currently, I'm using a 8800GTX (alas! with fan), which our IT people had lying around. Had to bake that as well before it would work. The card is pretty ludicrous as it doesn't even reduce clocks on idle - won't even support automatic profile switching with MSI Afterburner. But then, you don't look a gift horse ...

Post Reply