Do you think the rest of my system will be ok, its running fine, there are still 2 black pins on the mb but it seems to be ok?
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
24 pin MB Connector Burnt!
- Thread starter RichardH754
- Start date
Soldato
Seems to be yes. From looking at it briefly it would seem that the board doesn't have enough grounding.
Tell me, are the pins burnt connected to black wires when the PSU is plugged in?
IIRC there are four earth wires on a 24 pin ATX but I can't remember exactly off the top of my head.
Yeah and it was a decent 850W Seasonic M12 psu, one of the best you can buy as well
This issue is not caused by a PSU. It's caused by heavy current draw from the stuff plugged into it and the lack of ground wires to return to ground.
Would the HX1050 have had enough power?
Its only the 2 yellow wires that melted and burned
Soldato
OK then that doesn't really point to a grounding issue.
The problem with power supplies is that they are not documented very well. With a Corsair you sort of have peace of mind because it only has one rail per voltage. IE it has a single 12v rail instead of multiple rails. This means that all of your yellow wires go back to a single part of the PSU.
I do find it odd though that the motherboard is the part that has done that and not the PCIE power cables. This indicates that you are drawing too much current through those two 12v wires and not through the GPUs.
Sadly boards and all of that are poorly documented also. But, I would imagine that the problem is caused by your CPU (overclocked would make it worse) and the current draw through the physical PCIE lanes themselves. This is why some boards have that molex for them so they don't overdraw current through the two yellow wires.
I would also imagine that the thing that EVGA make that has been posted in this thread does two things.
1. Adds an extra ground wire for the return and
2. Adds extra 12v power for the PCIE lanes (it plugs into one you see).
I think overall though your board is simply drawing too much power through those two 12v wires. If the wires have not burnt up or gone hard (usually caused by the sheath overheating then cooling and then repeat, in the end making it brittle like plastic) then it's definitely the board.
I can't remember off the top of my head what PCIE current spec is per PCIE slot. If I knew the spec in wattage then you can back convert it into ampage (wattage is deciphered by the amps)
But yes, this has definitely been caused by too much current being drawn and the board's connectors not being up to the job.
Yes. As an example my system in my old car. I had four amplifiers. One of them was modded to run @ 1/4 ohm. As you lower the ohmage you increase the resistance, meaning it's very similar to the light bulb theory. As the power tries to go through the cables it hits a bottleneck and starts to push and shove. This creates heat, which is what has caused your problem. Most of these issues are caused by inadequate grounds. Basically my system had four 4 AWG lives for the 4 amplifiers and the capacitor. However, my return feed was all done by one cable through a huge brass block (a distributor). If I'd used anything else it would have melted.
Obviously in a PC there are no fuses. If there were? your 12v fuses would have blown and that would not have happened. IE - a fuse is simply a piece of metal or wire that can only take so much power being shoved through it before it breaks, preventing any damage at the other end of the wire.
Your power supply was definitely up to the task of delivering the required current. However, the motherboard hanging off the end of it was not up to the task of receiving and distributing it safely. It obviously needs more current for the PCIE lanes for triple SLI and everything else. It tried to suck too much current in through those two pins, and the pins overheated and burnt.
The problem with power supplies is that they are not documented very well. With a Corsair you sort of have peace of mind because it only has one rail per voltage. IE it has a single 12v rail instead of multiple rails. This means that all of your yellow wires go back to a single part of the PSU.
I do find it odd though that the motherboard is the part that has done that and not the PCIE power cables. This indicates that you are drawing too much current through those two 12v wires and not through the GPUs.
Sadly boards and all of that are poorly documented also. But, I would imagine that the problem is caused by your CPU (overclocked would make it worse) and the current draw through the physical PCIE lanes themselves. This is why some boards have that molex for them so they don't overdraw current through the two yellow wires.
I would also imagine that the thing that EVGA make that has been posted in this thread does two things.
1. Adds an extra ground wire for the return and
2. Adds extra 12v power for the PCIE lanes (it plugs into one you see).
I think overall though your board is simply drawing too much power through those two 12v wires. If the wires have not burnt up or gone hard (usually caused by the sheath overheating then cooling and then repeat, in the end making it brittle like plastic) then it's definitely the board.
I can't remember off the top of my head what PCIE current spec is per PCIE slot. If I knew the spec in wattage then you can back convert it into ampage (wattage is deciphered by the amps)
But yes, this has definitely been caused by too much current being drawn and the board's connectors not being up to the job.
Yes. As an example my system in my old car. I had four amplifiers. One of them was modded to run @ 1/4 ohm. As you lower the ohmage you increase the resistance, meaning it's very similar to the light bulb theory. As the power tries to go through the cables it hits a bottleneck and starts to push and shove. This creates heat, which is what has caused your problem. Most of these issues are caused by inadequate grounds. Basically my system had four 4 AWG lives for the 4 amplifiers and the capacitor. However, my return feed was all done by one cable through a huge brass block (a distributor). If I'd used anything else it would have melted.
Obviously in a PC there are no fuses. If there were? your 12v fuses would have blown and that would not have happened. IE - a fuse is simply a piece of metal or wire that can only take so much power being shoved through it before it breaks, preventing any damage at the other end of the wire.
Your power supply was definitely up to the task of delivering the required current. However, the motherboard hanging off the end of it was not up to the task of receiving and distributing it safely. It obviously needs more current for the PCIE lanes for triple SLI and everything else. It tried to suck too much current in through those two pins, and the pins overheated and burnt.
Last edited:
ALXAndy is spot on. My old board had an extra molex for the pci-e lanes, my new, more expensive one didn't
I was running my i7 cpu overclocked to 4.65Ghz and my gtx470's were over volted and overclocked to the max.
I put it down to the mb drawing too much power through the yellow lines. However, what surprised me was that I would have expected a quality psu to have some kind of ampage draw limiter and shut down before too many amps were drawn down the wires?
My new mobo with my Corsair 1250W psu is running fine but it does have an additionally molex for the pci-e lanes.
I was running my i7 cpu overclocked to 4.65Ghz and my gtx470's were over volted and overclocked to the max.
I put it down to the mb drawing too much power through the yellow lines. However, what surprised me was that I would have expected a quality psu to have some kind of ampage draw limiter and shut down before too many amps were drawn down the wires?
My new mobo with my Corsair 1250W psu is running fine but it does have an additionally molex for the pci-e lanes.
The 12v rail when 3 cards are in reads 11.30v in the bios after all this happened, so something must have been damaged in the PSU.
It use to read 11.60 when it was new
It use to read 11.60 when it was new
Soldato
Most PSUs have protection on them. And, if power is overdrawn they will realise it and cut out before any clipping happens (IE voltage spikes sent down the wires blowing up components).
All good ones have a safety shutoff.
Voltage does not equal current or ampage mate.
Put it this way.
My car used to have four amplifiers, a stiffening cap to prevent damage to the battery and alternator, and a DVD player, two screens, stereo, active parametric EQ and an awful lot of speakers, including an 18" SPL subwoofer.
One day when I was out driving the capacitor failed (I'm not surprised in all honesty). I cranked up my stereo and killed the battery AND my alternator. Basically I sucked them dry lol.
I would imagine your voltage shows lower in your bios now because the blackened pins are now buggered and not as conductive as before.
However, the issue wasn't caused by the PSU. It's basically something like this.
Motherboard "Give me power. Give me lots and lots of lovely power !"
PSU - "With pleasure, take, that !"
Motherboard "Oh dear. Due to poor design it seems I asked for more power than my connector's could safely handle and I have no way of knowing how much is too much so I seem to be a bit burnt !"
Basically the PSU wasn't breaking a sweat. 1050w back equates into a bloody lot of ampage. So at no point did the PSU think anything was wrong.
Now if the motherboard was electronically programmed to sense the current being pulled? it would have shut you down. Fact is it wasn't, so it didn't
Last edited:
Can OCUK get hold of any of them EVGA Power Booster things??
Soldato
I would personally replace the board my friend. As I said above, it is now having trouble getting the current to the areas that need it.
Either that, or take the board to a professional with a soldering iron and get him to replace the damaged pins. It shouldn't be that hard to replace them, all you need is a dead donor board
Either that, or take the board to a professional with a soldering iron and get him to replace the damaged pins. It shouldn't be that hard to replace them, all you need is a dead donor board
Old problem, been around since Fermi came out, but it did first get reported with dual gpu Ati cards at the time.
Its just overdraw on the 12v wires on the 24pin.
If you were unlucky enough to have the early EVGA boards without the secondary molex power connectors, or as in my case the Asus P6T7WS. Thankfully manufacturers started putting extra power for the pcie and even for the cpu on some boards.
When I RMA'd mine (it took out both the motherboard and PSU) Asus replaced the motherboard with a rev2 P6T7 which had an extra molex, so that was a result. Unfortunately the PSU maker wouldn't take responsibility and I ended up with a token partial cash refund and now only buy certain brands of PSU
Interestingly the guy who designed the EVGA Classified, Shamino, acknowledged the problem (he noticed it on Quadfire 3870's) and came up with a hack to fix it, interesting read if you like to know the history of it here.
Good luck!
Its just overdraw on the 12v wires on the 24pin.
If you were unlucky enough to have the early EVGA boards without the secondary molex power connectors, or as in my case the Asus P6T7WS. Thankfully manufacturers started putting extra power for the pcie and even for the cpu on some boards.
When I RMA'd mine (it took out both the motherboard and PSU) Asus replaced the motherboard with a rev2 P6T7 which had an extra molex, so that was a result. Unfortunately the PSU maker wouldn't take responsibility and I ended up with a token partial cash refund and now only buy certain brands of PSU
Interestingly the guy who designed the EVGA Classified, Shamino, acknowledged the problem (he noticed it on Quadfire 3870's) and came up with a hack to fix it, interesting read if you like to know the history of it here.
Good luck!
Soldato
Infact come to think of it I would RMA the board if it is still in warranty.
Edit. Biffa, just had a look at that link. That could help. Basically you are doubling the "lanes" for the power, so all of the strain does not end up on the pins.
Edit. Biffa, just had a look at that link. That could help. Basically you are doubling the "lanes" for the power, so all of the strain does not end up on the pins.
Yeah but its a bit heath robinson, and its only worth RMA'ing the board if you can get one with extra molex power connectors for the pcie. I was pleasantly surprised when Asus had rev'ed mine with the extra molex, its been running fine for years now with quad GTX470s in fact its lasted longer than the cards (had to replace two of them during that time)
Hang on, is it the Rampage II Extreme? That has two extra power molexes on it already!
One where I circled, but one also at right angles on the bottom just above where it says Asus
Hang on, is it the Rampage II Extreme? That has two extra power molexes on it already!
One where I circled, but one also at right angles on the bottom just above where it says Asus
Last edited:
Soldato
Well right now I would say his board is a bit Donald Ducked.
Those pins are burnt and probably now not conducting as they should. It's sort of confirmed by the fact his bios is showing a lower 12v current than it did before.
So, it's either repair it (replacing the burnt pins with clean ones) and then modding it, or RMAing it and then modding it or using that thing EVGA make.
Edit. Something I forgot to add. OP - look at the connectors in the ATX plug and make sure they are not burnt or damaged also.
Those pins are burnt and probably now not conducting as they should. It's sort of confirmed by the fact his bios is showing a lower 12v current than it did before.
So, it's either repair it (replacing the burnt pins with clean ones) and then modding it, or RMAing it and then modding it or using that thing EVGA make.
Edit. Something I forgot to add. OP - look at the connectors in the ATX plug and make sure they are not burnt or damaged also.
Yes, the plugs are melted, but the PSU will fix that.
Its fine with 1 GTX 480 in 12.02v on 12v line
Its fine with 1 GTX 480 in 12.02v on 12v line
Soldato
It would be. But as it draws more current the damage will prevent it making it.
You don't want to starve your cards of power.
Can you not RMA it?
Its just out of warranty,ran out Feb 12!
Might have to try that mod
Might have to try that mod
Soldato
Get the pins replaced too man.
I can not find it now but lots of GFX cards have been exceeding PCIEX slot draw.
This means slots can be burnt and any feeds/connectors in the circuit back to the PSU.
Now add 2/3 cards that exceed slot limits you have a big problem. Any descent Board designed for big combined slot loads has to have some sort of AUX feed as Mentioned.
Depending on the quality of a PSU and it's connectors I have seen a few burnt ATX connectors...now this is not always load related so much as cheap metal that looses it's spring (over time and exposure to general heat) obviously it's down hill after that !
Just replacing one side of a bad connector is obviously asking for more trouble. The PSU connector can be quite easily repaired but the motherboard will need soldering skills.
Out of interest from memory in the GTX580 range of GFX cards I believe only the MSI lightning did not exceed PCIex design slot power limits
Also remember Never ever waggle connectors when fitting or removing, if you spread the female side connector they will never work reliably again.....I fully appreciate this can be next to impossible, but at minimum double check if you have abused them pulling them off
And I always check each wire (when possible) in multi connectors to make sure it's got a good connection and has not pushed out of the clip in the housing.
This means slots can be burnt and any feeds/connectors in the circuit back to the PSU.
Now add 2/3 cards that exceed slot limits you have a big problem. Any descent Board designed for big combined slot loads has to have some sort of AUX feed as Mentioned.
Depending on the quality of a PSU and it's connectors I have seen a few burnt ATX connectors...now this is not always load related so much as cheap metal that looses it's spring (over time and exposure to general heat) obviously it's down hill after that !
Just replacing one side of a bad connector is obviously asking for more trouble. The PSU connector can be quite easily repaired but the motherboard will need soldering skills.
Out of interest from memory in the GTX580 range of GFX cards I believe only the MSI lightning did not exceed PCIex design slot power limits
Also remember Never ever waggle connectors when fitting or removing, if you spread the female side connector they will never work reliably again.....I fully appreciate this can be next to impossible, but at minimum double check if you have abused them pulling them off
And I always check each wire (when possible) in multi connectors to make sure it's got a good connection and has not pushed out of the clip in the housing.