Overloaded PCIE 16x slot, saw smoke.

[gaximodo]

So my old PSU 1000watt Antec died last week and I replaced it with a Corsair CXM 750W. Motherboard is Asus Sabertooth Z77.

the Corsair CXM 750W was a modular PSU and I didn't push it all the way in on the PSU end for the PCIE 8pin connector.

When I powered the PC on I saw smoke came out from the first PCIE slot which I had a GTX1070 in there, must because the card was trying to draw too much power from the PCIE slot.

I then powered off, connected the card to secondary PCIE slot, booted fine. then I put the card back in the first slot, booted fine and tested fine too (few games, not stress tests).

I need advise - is it ok to just use the first slot like nothing has ever happened? What's the risk? There was smoke and I could smell burnt, no other visible damage tho..

 

[eidairaman1]

Its damaged, do not use, get a new motherboard

 

[biffzinker]

There was smoke and I could smell burnt, no other visible damage tho..

If it were me, I'd look for a replacement board.

 

[gaximodo]

You would think the motherboard packs some sort of protection over overdrawing power!

OK thanks guys, what if I leave the slot empty, put the graphics card in slot two? is there any risk?

 

[INSTG8R]

You would think the motherboard packs some sort of protection over overdrawing power!

OK thanks guys, what if I leave the slot empty, put the graphics card in slot two? is there any risk?

Well you can’t really know what the extent of the damage is. Slot 2 could go poof next it’s not worth risking damaging anything else

 

[eidairaman1]

You would think the motherboard packs some sort of protection over overdrawing power!

OK thanks guys, what if I leave the slot empty, put the graphics card in slot two? is there any risk?

Yes there is, once a board is damaged do not use it

 

[gaximodo]

Yes there is, once a board is damaged do not use it

Thanks, I hear you, if I need to replace the board I would rather have the CPU and RAM replaced too which is a big spending for something I don't use that much anymore. So long the harddisk is not in risk I'm ok with the rest of system gone. (I thought they were gone with the Antec PSU anyway...)

Will report back if something happens so other people in the same situation can learn.

 

[biffzinker]

If the system starts exhibiting stability issues at a later time you'll know which component to suspect.

 

[agent_x007]

I think your PSU's plug got burned (pins and plastic). Since that's where arching would be taking place (assuming it wasn't seated properly).
Since it was on PSU side, excess heating could have caused issues on it's PCB.
However, unless you see visible damage on or near a PCIe slot or card itself - it should be fine (arching isn't a good thing, but if it works, it works - may not live long though). Still, You must find what element got burnt, because wharever it was anything connected to it probably will fail in very near future.

 

[lexluthermiester]

Thanks, I hear you, if I need to replace the board I would rather have the CPU and RAM replaced too which is a big spending for something I don't use that much anymore.

If you are happy with the system as it is, you could likely just replace the board with the exact same model seamlessly. If you share your model spec we can help you find replacements.

I think your PSU's plug got burned (pins and plastic).

This, It would be a very good idea to replace your PSU at the same time as the PSU is likely part of the problem.

 

[gaximodo]

I think your PSU's plug got burned (pins and plastic). Since that's where arching would be taking place (assuming it wasn't seated properly).
Since it was on PSU side, excess heating could have caused issues on it's PCB.
However, unless you see visible damage on or near a PCIe slot or card itself - it should be fine (arching isn't a good thing, but if it works, it works - may not live long though). Still, You must find what element got burnt, because wharever it was anything connected to it probably will fail in very near future.

No PCIE connector was not connected to the PSU, I had visual of smoke coming out of PCIE slot not PSU.

 

[biffzinker]

This, It would be a very good idea to replace your PSU at the same time as the PSU is likely part of the problem.

Corsair power supply just replaced the Antec. Are you suggesting the OP replace the Corsair?

So my old PSU 1000watt Antec died last week and I replaced it with a Corsair CXM 750W

 

[gaximodo]

If you are happy with the system as it is, you could likely just replace the board with the exact same model seamlessly. If you share you model spec we can help you find replacements.

This, It would be a very good idea to replace your PSU at the same time as the PSU is likely part of the problem.

Its an old Z77 board and gen 3, 4 cores CPU. I don't feel like getting another old board. will see how much longer I can get out of it.

 

[ne6togadno]

System Builder

au.pcpartpicker.com

 

[lexluthermiester]

Corsair power supply just replaced the Antec. Are you suggesting the OP replace the Corsair?

The way that he stated the problem implies that the Corsair was the one connected when the problem occurred. If so then yeah, it sucks but replacing that PSU might be wise. However, he did say...

Its an old Z77 board and gen 3, 4 cores CPU. I don't feel like getting another old board. will see how much longer I can get out of it.

Fair enough. Have you tested the 1070 in Slot one properly connected to the PCIe power cable(s)? This wasn't clear from the conversation above. It is possible that a bug, hair or some other small bit of something may have momentarily shorted over two points of power contact that was then burned off. It is also possible that no permanent damage was done. Testing might reveal this. If your system is still booting with and without the 1070, then you might be ok. Just be sure to use a can of air or a vacuum to clean the slot out before you test. Once clean, check to see if any burn/scorch marks are present. If yes, don't try. If no, give it a try. If you're worried about your drive, disconnect it before testing.

 

[Vayra86]

Fair enough. Have you tested the 1070 in Slot one properly connected to the PCIe power cable(s)? This wasn't clear from the conversation above. It is possible that a bug, hair or some other small bit of something may have momentarily shorted over two points of power contact that was then burned off. It is also possible that no permanent damage was done. Testing might reveal this. If your system is still booting with and without the 1070, then you might be ok. Just be sure to use a can of air or a vacuum to clean the slot out before you test. Once clean, check to see if any burn/scorch marks are present. If yes, don't try. If no, give it a try. If you're worried about your drive, disconnect it before testing.

I would certainly try that but with some crappy old dummy card and not a working 1070.

 

[kapone32]

I had a similar issue. I had a power surge go through my system. It actually blew one of the caps on one of my GPUs (hope I can get it fixed). When I looked at the board though, the southbridge cover had evidence of a cap also exploding. The board would start and all the fans would run. I found out it also killed my CPU (I sent it infor Warranty). As the board is under warranty I was able to get a replacement board but I got the Amazon warranty so I exchanged the board for a different board. Unfortunately you are in a different situation but I would still get a new board regardless of what is working, as the board could just be starting to go bad. You could find CPU and MB combos for $300 or less that would be better than what you currently have like a 1600/2600 plus B450 combo. I only advise AMD because of the diversity of their product stack.

 

[Grog6]

Look closely at the area around the slot.
If you don't see anything, look closely at the Video card.

Unless you burnt a trace off the Mobo, I'd bet any damage is to the Video card.
Typically, the 3.3 and 12V for the slots come from planes in the PCB, so it's not real likely to be the mobo.

The PCIe regulator is going to be a linear LDO, if there is one.
It would reduce the 3.3V or 12V from the mobo by ~0.2-0.3V, at a maximum.
It would look like a SMT mosfet, ~10mm square for the power involved.
There's no power limiter, as such; there could be a mosfet on/off switch, but I can't imagine why.

Most do not have a regulator on the Mobo at all; the power supply for the card is on the card.

It's most likely any smoke came off the card, not the mobo.

Here's the pinout on the pcie slot.

PCI Express - Wikipedia

en.wikipedia.org

 

[eidairaman1]

I had a similar issue. I had a power surge go through my system. It actually blew one of the caps on one of my GPUs (hope I can get it fixed). When I looked at the board though, the southbridge cover had evidence of a cap also exploding. The board would start and all the fans would run. I found out it also killed my CPU (I sent it infor Warranty). As the board is under warranty I was able to get a replacement board but I got the Amazon warranty so I exchanged the board for a different board. Unfortunately you are in a different situation but I would still get a new board regardless of what is working, as the board could just be starting to go bad. You could find CPU and MB combos for $300 or less that would be better than what you currently have like a 1600/2600 plus B450 combo. I only advise AMD because of the diversity of their product stack.

What cap?

 

[Bill_Bright]

then I put the card back in the first slot, booted fine and tested fine too

I have mixed feelings about this. If you had come to us BEFORE trying the card in the first slot again, I would have advised against using it. But since you did it already, and it worked without further smoke, I'd say you are probably fine. My "guess" is you shorted some voltage harmlessly to ground and not through some other sensitive circuit or component. Ohm's Law tells us when resistance goes down (the short) and voltage remains the same, current must go up. And when current goes up, typically heat does too. And the smoke was that voltage frying some dust mixed with other contaminants that float around.

Since you have used it and experienced no further problems, I personally think all is okay. But you should still err on the safe side for now and don't leave the computer on unattended for a few days while you use it. And at the first sign/scent of trouble, be ready to quickly unplug the computer from the wall. Remember, simply shutting down through the Start menu or powering off with the computer's front panel power button does NOT completely shut down the power supply. Those actions merely put it in standby mode and +5Vsb standby voltage is still being distributed to multiple points on the motherboard. To stop an electrical fire, you have to completely remove the electricity.

If I have one concern, it is with that Corsair PSU. As I said before, Ohm's Law tells us that short would have caused current to go up. So I wonder why didn't that Corsair PSU sense the excessive current and immediately shut down on its own? It seems to me it should have.

And why did your Antec die just a week before? Did you have two faulty PSUs in 1 week? Why? A coincidence? I don't know but I recommend you immediately use a AC Outlet Tester to ensure your outlet is properly wired and grounded to Earth ground. I recommend one with a GFCI (ground fault circuit interrupt) indicator as it can be used to test bathroom and kitchen outlets (outlets near water) too. These testers can be found for your type and voltage outlet, foreign or domestic, (like this one for the UK) at most home improvement stores, or even the electrical department at Wal-Mart. Use it to test all the outlets in the home and if a fault is shown, have it fixed by a qualified electrician.

And just to ease my mind, since I'm not crazy about the CX series of Corsairs in the first place, and if not mistaken the CXM line stopped production years ago. So if me, I would get a new PSU. I like EVGA SuperNovas and Seasonic.

 

[kapone32]

What cap?

I am not yet sure I have to pull the card apart from the water block and check. I was going to take it to a repair shop up the street where the technician and I have struck up a great relationship because he told me he loves working on PC repairs but 99% of his customers are phone or tablet based.

 

[eidairaman1]

I am not yet sure I have to pull the card apart from the water block and check. I was going to take it to a repair shop up the street where the technician and I have struck up a great relationship because he told me he loves working on PC repairs but 99% of his customers are phone or tablet based.

Good then

 

[lexluthermiester]

I would certainly try that but with some crappy old dummy card and not a working 1070.

Considering that he's already had it in the slot and it was ok, I think there is a good chance that it was just a momentary problem. My "gut" feeling thinks the OP will be fine, but it's up to them to test.

 

[gaximodo]

I have mixed feelings about this. If you had come to us BEFORE trying the card in the first slot again, I would have advised against using it. But since you did it already, and it worked without further smoke, I'd say you are probably fine. My "guess" is you shorted some voltage harmlessly to ground and not through some other sensitive circuit or component. Ohm's Law tells us when resistance goes down (the short) and voltage remains the same, current must go up. And when current goes up, typically heat does too. And the smoke was that voltage frying some dust mixed with other contaminants that float around.

Since you have used it and experienced no further problems, I personally think all is okay. But you should still err on the safe side for now and don't leave the computer on unattended for a few days while you use it. And at the first sign/scent of trouble, be ready to quickly unplug the computer from the wall. Remember, simply shutting down through the Start menu or powering off with the computer's front panel power button does NOT completely shut down the power supply. Those actions merely put it in standby mode and +5Vsb standby voltage is still being distributed to multiple points on the motherboard. To stop an electrical fire, you have to completely remove the electricity.

If I have one concern, it is with that Corsair PSU. As I said before, Ohm's Law tells us that short would have caused current to go up. So I wonder why didn't that Corsair PSU sense the excessive current and immediately shut down on its own? It seems to me it should have.

And why did your Antec die just a week before? Did you have two faulty PSUs in 1 week? Why? A coincidence? I don't know but I recommend you immediately use a AC Outlet Tester to ensure your outlet is properly wired and grounded to Earth ground. I recommend one with a GFCI (ground fault circuit interrupt) indicator as it can be used to test bathroom and kitchen outlets (outlets near water) too. These testers can be found for your type and voltage outlet, foreign or domestic, (like this one for the UK) at most home improvement stores, or even the electrical department at Wal-Mart. Use it to test all the outlets in the home and if a fault is shown, have it fixed by a qualified electrician.

And just to ease my mind, since I'm not crazy about the CX series of Corsairs in the first place, and if not mistaken the CXM line stopped production years ago. So if me, I would get a new PSU. I like EVGA SuperNovas and Seasonic.

I really appreaciated your time going through my issue. Thank you Bill_Bright! And all others of course

I had the Antec PSU for 12 years, and it was hot here in Western Australia the day it blew up (42 °C ).

I was trying to simplify my situation, what really happened was I used one module cable (its longer) from Antec PSU on the Corsair, assumed it will be compatible but the board won't boot, so I unplugged all module cables on the PSU while the other ends are still connected to devices, so only 24pin and 8pin is connected to see if it will boot.

It booted, then the smoke happened so I shut the power immediately. I believe the card was drawing too much power from the PCIE slot on the board, with the PCIE 8 pin connected to the graphics card, not connected to PSU. I checked both card and board, can't see any visible damage. Rest you know.

I stopped using the Antec cable, everything boots fine, and now have the card in secondary PCIE slot, and a sound card in the first PCIE slot, so far no sign of abnormality.

I will test the outlets also, thank you again for the advice. it may have something to do with the old power board also died from the scene.

 

[Deleted member 185158]

I cannot give advice without seeing pictures of the area where the smoke had appeared. Nor can anyone else.

But given the age, and if it still works, rock it out until you can build a new rig.

My 2 cents, and it's worth it's weight in copper waterblocks.