Corsair Grey Label CX750M Burning Smell [Fault Finding And Repair]

[iamdarkyoshi]

I got a CX750M Grey Label PSU (Model 75-002019) for free from a friend's friend. I was told it was making hot smells, so of course I had to test this myself.

 

And it sure as heck did, I connected it to some old hardware to put a load on it, and it STANK. It did actually stay running, however.

 

It was making an awful burning smell, but it didn't smell like burning semiconductor. I asked Corsair if I could RMA it without a proof of purchase, and to make a long story short... No.

 

So that meant I could do my favourite thing:

 

 

Void that pesky warranty sticker 

 

 

 

Once I was inside, I looked around on the board for things that look like they got hot.

 

I instantly noticed that this inductor looked a bit... burnt. Its supposed to be copper coloured like the shield behind it.

 

 

The capacitor next to it also looks like the radiant heat from the inductor discoloured the outer shell

 

 

Lets get those capacitors out of the way for a better look:

 

 

Something is clearly wrong here. I removed the shield as well, which had a skidmark on it

 

 

Lets remove that inductor as well, to see what the heck happened to it.

 

 

That certainly explains what was making the heat and smells...

 

But why did it fail? I took a closer look, and next to where the skidmark on the shield was, there appeared to be a short where two windings touched.

 

I unwrapped the inductor, and the insulation just flaked off the wire, and that suspect area was really burnt. I measured a short between the two windings, which confirmed my suspicions.

 

Here's what it looks like with the thick winding removed.

 

 

These windings sure got hot. I definitely cannot reuse this wire, the insulation is literally flaking off.

 

 

This inductor had twelve windings with the single conductor, which was 1mm thick. In the same direction was the four paralleled windings of 1.25mm wires, also twelve windings.

 

I've bought some replacement wire of the same gauge, so I can rewind this inductor. I have also cleaned and painted the ferrite ring with engine enamel to re-insulate it since the original paint had been burned off.

 

 

I should have a result this week documenting the repair.

 

What I'm hoping was that this was a manufacturing defect causing the windings to short, perhaps due to being wound too tightly, not a failure somewhere else on the power supply.

 

I'd be interested to see Corsair's response, I know we have quite a few reps on the forum now!

[Energycore]

This is so cool! Keep updating us, I want to see you fully rep that PSU

 

Also since you want Corsair's Input let's tag @Corsair and @Corsair Nick

 

Also @jonnyGURU might want to look at this

[Beowulff83]

Favorite picture is definitely pic #1. 

[aisle9]

This is what happens when you let the Logisys guys tour your facility without an escort.

[ItsMitch]

51 minutes ago, Energycore said:

Also since you want Corsair's Input let's tag @Corsair and

Corsair is banned  

[Stefan Payne]

1 hour ago, iamdarkyoshi said:

I instantly noticed that this inductor looked a bit... burnt. Its supposed to be copper coloured like the shield behind it.

You're not the first one to see that.

I've already seen a pic of that in Badcaps Forum. Someone there also had a Corsair CX with a burned secondary Coil as well.

Don't know how common this is. But if you see one coil or two, others aren't far away because Coils normally don't just burn...

 

The first time I've heard about burned coils was with an FSP unit where the material was faulty wich caused the Coil to overheat. But that was over 10 years ago...

[iamdarkyoshi]

6 minutes ago, Stefan Payne said:

You're not the first one to see that.

I've already seen a pic of that in Badcaps Forum. Someone there also had a Corsair CX with a burned secondary Coil as well.

Don't know how common this is. But if you see one coil or two, others aren't far away because Coils normally don't just burn...

 

The first time I've heard about burned coils was with an FSP unit where the material was faulty wich caused the Coil to overheat. But that was over 10 years ago...

I'm curious as to what could have caused it though. I checked the daughterboard behind it for shorted transistors and came up empty.

 

The really strange part is that the supply seemed to run fine, albeit hot and smelly.

[Hiitchy]

I liked the first pic tbh. I don't know much about power supplies, but is it possible that this could have been caused by overvoltage? 

[Stefan Payne]

37 minutes ago, iamdarkyoshi said:

I'm curious as to what could have caused it though.

The most likely is that the Coil was faulty and or the wrong type.

ie manufacturing error or used wrong materials despite specification of Coil.

[iamdarkyoshi]

4 minutes ago, Stefan Payne said:

The most likely is that the Coil was faulty and or the wrong type.

ie manufacturing error or used wrong materials despite specification of Coil.

I'm hoping it was the first option.

 

This PSU has had a noticable amount of dust in the fan, so its not like it was shipped out like this.

 

Given that there was what appeared to be a hotspot where the two windings shorted, I'm pretty sure that it was faulty, possibly wound too tightly.

 

Only one way to find out!

[W-L]

1 minute ago, iamdarkyoshi said:

I'm hoping it was the first option.

This PSU has had a noticable amount of dust in the fan, so its not like it was shipped out like this.

Given that there was what appeared to be a hotspot where the two windings shorted, I'm pretty sure that it was faulty.

Only one way to find out!

Usually for something like that to occur it would need to short out to blacken/burn like that, it's quite possible the insulation on the magnet wire was damaged. It's kinda hard to test for it unless you have a meter that could have tested the coil for inductance and that is if we knew the original value of the coil beforehand. 

 

Interesting to see if it still functions with a new inductor installed. 

[iamdarkyoshi]

Just now, W-L said:

Usually for something like that to occur it would need to short out to blacken/burn like that, it's quite possible the insulation on the magnet wire was damaged. It's kinda hard to test for it unless you have a meter that could have tested the coil for inductance and that is if we knew the original value of the coil beforehand. 

 

Interesting to see if it still functions with a new inductor installed. 

I have a meter that can do insuctance

 

I should have tested before and after repairing it...

 

Either way we know there are two windings, both 12 turns, and both in the same direction around the torroid.

 

As long as the ferrite's properties didn't change, we should have a coil with the same properties as the original one.

[Guest]

Your friend is a fool for giving you a PSU for free instead of using it's five year warranty to get a new one.  Must be nice to have money to throw away.

 

But I have seen this before and an ECN was implemented a long time ago to improve the choke manufacturing process some time ago.

 

What's the lot code of the PSU (the beginning part of the serial number before all the 0's)?  I can tell if it was before or after the ECN was implemented based on that.  

[iamdarkyoshi]

4 minutes ago, jonnyGURU said:

Your friend is a fool for giving you a PSU for free instead of using it's five year warranty to get a new one.  Must be nice to have money to throw away.

 

But I have seen this before and an ECN was implemented a long time ago to improve the choke manufacturing process some time ago.

 

What's the lot code of the PSU (the beginning part of the serial number before all the 0's)?  I can tell if it was before or after the ECN was implemented based on that.  

15367187 is the start of the serial before the 0's

 

I'm not complaining, I got a fun project to work on because my friend's friend didn't RMA it

 

Thanks for the info!

[Guest]

Yeah.  36th week of 2015.  That's a pretty old unit.  One of the first.

[iamdarkyoshi]

1 minute ago, jonnyGURU said:

Yeah.  36th week of 2015.  That's a pretty old unit.  One of the first.

That could explain it.

 

So this burning up choke was a known issue?

 

Was it just the wire being wound too tightly and compromising the insulation?

[Guest]

Yeah.  Just inconsistent assembly due to lack of SOP.  The magnetics manufacturer wasn't audited hadn't been audited back when this was made.  

 

China sucks. Even when you think you have a PSU factory up to snuff... one of the component manufacturers is going to come along and screw things up.

 

Just wait until they move things to Vietnam because China labor costs are getting too high.  I might have to retire for my own sanity sake.

[iamdarkyoshi]

14 minutes ago, jonnyGURU said:

Yeah.  Just inconsistent assembly due to lack of SOP.  The magnetics manufacturer wasn't audited hadn't been audited back when this was made.  

 

China sucks. Even when you think you have a PSU factory up to snuff... one of the component manufacturers is going to come along and screw things up.

 

Just wait until they move things to Vietnam because China labor costs are getting too high.  I might have to retire for my own sanity sake.

Thanks a lot for the info man! Sounds like this PSU is on its way to becoming a repaired unit!

 

What does amuse me to a certain degree is that it actually ran my old P4 testbench fine. Voltages were well within spec, and yet it smelled like a hairdryer thats been plugged up...

[vanished]

25 minutes ago, iamdarkyoshi said:

Thanks a lot for the info man! Sounds like this PSU is on its way to becoming a repaired unit!

 

What does amuse me to a certain degree is that it actually ran my old P4 testbench fine. Voltages were well within spec, and yet it smelled like a hairdryer thats been plugged up...

Well obviously that means this component was just added as a conspiracy for the purpose of planned obsolescence and was not actually required for good operation! /s

 

It should be interesting to see if just rewinding this and reinstalling it is all it needs.  Given that the core was probably heated to at least 100 C if not much higher, it's possible that the metal has changed somewhat, but then, this is basically an electromagnet, not a permanent one so losing any acquired field shouldn't be an issue.  I'm unsure if other changes could have occurred that would make a difference though.

[iamdarkyoshi]

18 minutes ago, Ryan_Vickers said:

Well obviously that means this component was just added as a conspiracy for the purpose of planned obsolescence and was not actually required for good operation! /s

 

It should be interesting to see if just rewinding this and reinstalling it is all it needs.  Given that the core was probably heated to at least 100 C if not much higher, it's possible that the metal has changed somewhat, but then, this is basically an electromagnet, not a permanent one so losing any acquired field shouldn't be an issue.  I'm unsure if other changes could have occurred that would make a difference though.

I'm sure it was well above 100C. It managed to char a fibreglass spacer under the choke, whats the char point of fibreglass?

[vanished]

Just now, iamdarkyoshi said:

I'm sure it was well above 100C. It managed to char a fibreglass spacer under the choke, whats the char point of fibreglass?

If you really want to investigate, I suppose you could take another similar core, wind it, test the inductance, then unwind it, bake it or blast it with a heat gun, and then rewind it and test it again to compare (or just leave the winding on if you think it can take it).  If this changes (or doesn't), it's likely that the other has done the same.

[iamdarkyoshi]

1 minute ago, Ryan_Vickers said:

If you really want to investigate, I suppose you could take another similar core, wind it, test the inductance, then unwind it, bake it or blast it with a heat gun, and then rewind it and test it again to compare (or just leave the winding on if you think it can take it).  If this changes (or doesn't), it's likely that the other has done the same.

Well of course this has to be tested, I'll wind one tomorrow and bake on my bench supply.

 

I'll also see if I can measure the temperature of the wire, but it ain't gonna be easy.

[vanished]

3 minutes ago, iamdarkyoshi said:

Well of course this has to be tested, I'll wind one tomorrow and bake on my bench supply.

 

I'll also see if I can measure the temperature of the wire, but it ain't gonna be easy.

I suppose one thing I forgot to mention is while you're testing if bringing the temperature up will affect it, remember that you're not testing whether it is different at high temperature.  In other words, make sure it cools off fully before testing it again

[Stefan Payne]

Isn't the laquer of the wires OK for up to 200°C or something like that??

[iamdarkyoshi]

8 minutes ago, Stefan Payne said:

Isn't the laquer of the wires OK for up to 200°C or something like that??

Depends on the wire. Whatever this was rated for, it has been exceeded.