Overclock.net › Forums › Components › Power Supplies › Melted 8 pin 12v
New Posts  All Forums:Forum Nav:

# Melted 8 pin 12v - Page 3

Quote:
I actually bought a quality unit from performance-pcs. They ended up sending me a quality 8 pin for another format (if I recall it right a pci-e one). After some discussion people here said it would be OK to use the cable if I could fit it there, so I modded it to fit an EPS header. Probably this is the case. I had to take my rig apart right before the melting started, I must have put it back together lazily.

Wait, hold on a minute. You said you're using a PCIe 8 pin molex connector for your mobo? You said you modded it, but I'm assuming you realize that the PCIe 8 pin has a ground where a 12v would be for a cpu 8 pin molex. (see pinout diagram I posted earlier) That could be very bad. Very bad indeed.
Quote:
Originally Posted by Arimis5226;15381788
I'm not sure I understand. For each independent current limit, there would be a different rail in the PSU, no? That PSU has 4 different current limits (+12V1-4). If the PSU isn't limiting the current like it should on one of the rails, then overcurrent is possible, no? I'm not sure what you mean by your post. Would you please clarify instead of just telling me I'm wrong, by telling me how I'm wrong?

Sorry, was in a hurry.

Current is pulled, not pushed. The power supply is designed to provide up to a given amount of current at a specific voltage. I'm not going to go into how it does that, because that would take a full college course.

The amount of current pulled is inversely proportional to the amount of resistance on that circuit. V=IR. So if you have 12V and 0.5ohms then you'd have a current of 24A. But if you had a resistance of .2ohms then you'd have a current of 60A. With computer circuits it isn't quite simple resistance because blahdeeblah complicated stuff, but generally the more complicated a task a processor is handling the more circuits are closed, so the lower the resistance, so the more current pulled.

If you have a short circuit (when resistance drops to close to 0ohms) or if you just have too current hungry a chip then you can exceed the current rating of the wires that deliver power to the chip, and they can burn up due to excess heat.
I would also like to point out that both of these people with the same problem also have PSUs with multi 12v rails (one has 4, the OP has 6). Basically, if I'm not mistaken, there is one major +12v rail, and then there are minor rails branching off from that one (depending on how many they PSU has). Problem with these PSUs is that a lot of people don't understand how to balance your current loads on each rail properly. That put to the side, they are awesome PSUs for an enthusiast....but I digress.

If these PSUs had a single 12v rail like many of the others out there, then I would say that it was a bad connection, as the remainder of your 12v components havent failed or have been damaged. Allow me to reiterate. The fact that the damage is done across all the +12v pins (and not just one) suggests to me that the problem is at the PSU itself, not dust in a connection. If you had dust in a connection, it would more likely cause damage at one connection, perhaps spread to an adjacent pin, but I would highly doubt it would hit them all. Seeing as these are both PSUs with multiple 12V rails, I'd imagine that a bad rail would damage all of the 12v connection pins at a single connection associated with that one bad rail, while still maintaining good current to the other +12v rails.

With the fact that EPS cables are being used, it could be a bad connection (poor modification or construction), which would be a good thing because you don't have to buy yourself a new PSU. However, just because you replace the connection and it "works fine" doesn't mean that it is fine. Either test the PSU, or keep an eye on the replaced connections. Things like these start fires and burn down homes. I can't stress this enough. Personally, I'd replace the PSU regardless for peace of mind, and believe me I wouldn't be happy about flipping the bill. At least I'd have another spare PSU laying around for troubleshooting down the road.
Quote:
Originally Posted by Phaedrus2129;15382492
Sorry, was in a hurry.

Current is pulled, not pushed. The power supply is designed to provide up to a given amount of current at a specific voltage. I'm not going to go into how it does that, because that would take a full college course.

The amount of current pulled is inversely proportional to the amount of resistance on that circuit. V=IR. So if you have 12V and 0.5ohms then you'd have a current of 24A. But if you had a resistance of .2ohms then you'd have a current of 60A. With computer circuits it isn't quite simple resistance because blahdeeblah complicated stuff, but generally the more complicated a task a processor is handling the more circuits are closed, so the lower the resistance, so the more current pulled.

If you have a short circuit (when resistance drops to close to 0ohms) or if you just have too current hungry a chip then you can exceed the current rating of the wires that deliver power to the chip, and they can burn up due to excess heat.

Yes, I get that. Current is pulled not pushed. Semantics, but I admit my post was up for attack on that note. I'm an electronics tech, so I know about ohms law. 12 years now in the profession, years of classes and training. I should have chosen my words more cafefully. This doesn't have much to do with taking out the word "rails" in my post, I don't think. Both of these power supplies have lower current limits on each rail as opposed to a single rail power supply, so I have a hard time believing that the wires reached thier current limit and burnt (not impossible I suppose), but the connections are a different picture. I'm not sure how a bad connection could create lower resistance. I can see how a bad connection could easily create a higher resistance, but that would lower current draw, not increase it. Foreign matter stuck in the connection could cause a decrease in the resistance due to the matter being more conductive than the wire/connection. I'm still not a big fan of the idea of foreign matter on one connection causing all 4 of the pins to fry. Aren't the connections for the cables coming from a power supply parallel (constant voltage, variable current)? If that was the case, then aren't we looking at over-current across the entire circuit (all 4 pins damaged)? I'm no electrical engineer (yet), but I know enough to follow what your saying and expand.
A PSU's +12V rail isn't "split". Or, well... Both single rail and multi-rail PSUs have a single +12V source. Both of them split that source into several contact points on the PCB where the wires that leave the PSU are soldered on. The only difference is that multi-rail PSUs monitor those separate contact points with over current protection circuits, which are there to prevent too much current being pulled from one bundle of wires, in case of a short circuit or overloading of the PSU. Single rail power supplies either have one OCP circuit monitoring all cables, or else (more commonly) have no +12V OCP at all.

(The only common "true" multi-rail design, where there are multiple +12V sources, is the Channel Well PUC design, used in the Corsair HX1000, some Thermaltake 1kW+ units, and some others. It's basically two 500W SMPS crammed into one housing with a shared EMI filter. There are also some server units that have multiple transformers, but that's not seen in ATX PSUs; there are also one or two units out there with multiple +12V inductors in parallel which could be considered multi-rail. Those are very rare.)

The fact that these are multi-rail PSUs probably prevented further damage. If it had been a single rail power supply, it likely would have been far worse. Like so:
http://www.overclock.net/power-supplies/944707-why-single-rail-not-better-than.html
Quote:
Originally Posted by Phaedrus2129;15383000
A PSU's +12V rail isn't "split". Or, well... Both single rail and multi-rail PSUs have a single +12V source. Both of them split that source into several contact points on the PCB where the wires that leave the PSU are soldered on. The only difference is that multi-rail PSUs monitor those separate contact points with over current protection circuits, which are there to prevent too much current being pulled from one bundle of wires, in case of a short circuit or overloading of the PSU. Single rail power supplies either have one OCP circuit monitoring all cables, or else (more commonly) have no +12V OCP at all.

(The only common "true" multi-rail design, where there are multiple +12V sources, is the Channel Well PUC design, used in the Corsair HX1000, some Thermaltake 1kW+ units, and some others. It's basically two 500W SMPS crammed into one housing with a shared EMI filter. There are also some server units that have multiple transformers, but that's not seen in ATX PSUs; there are also one or two units out there with multiple +12V inductors in parallel which could be considered multi-rail. Those are very rare.)

The fact that these are multi-rail PSUs probably prevented further damage. If it had been a single rail power supply, it likely would have been far worse. Like so:
http://www.overclock.net/power-supplies/944707-why-single-rail-not-better-than.html

Yeah, okay. I think we're on the same page now. Sorry for the confusion. I'm thinking that in this case, either there was something stuck in the connection (or short in the connection), or the overcurrent circuitry in the PSU failed (or possible bad ground for that rail), which caused the meltdown. With all 4 pins being damaged, I'm thinking that the rail associated with that connection is likely damaged as well. The instances I've seen at work, I opened up the power supplies and there was damage inside as well if I recall (in this case, one was 12v, one was a 5v). Either way, the PSU needed to be replaced. If it was the PSU, then duh. If it was the connection, that also damaged the PSU. I think the OP lucked out as there werent any components plugged in at that connection. I had to replace the daughterboard (backplane) as well as the PSU on the one with the 12v problem. Thanks for taking the time to square me off. I hope this discussion helps the OP, as I'm certainly not a fan for hijacking threads.
Too bad I can't see which rails are the problem. The EPS header and the extension are melted together and my CPU cooler blocks the way. I could take everything apart but that would take many hours I don't currently have to spare.

I found out about this problem when creating a new overclock profile and smelled something, like burnt plastic. Now I smell this every time my computer is at load. I went back to stock and only on high load I can smell something burning.

I bought a new extension to try it before buying a new PSU. I could test it without the modded extension right now, but then I would risk melting the PSU to the mobo. I don't mind about the mobo because I already have to change it for another reason.
 Prodigy (14 items)
 Prodigy (14 items)
Quote:
Too bad I can't see which rails are the problem. The EPS header and the extension are melted together and my CPU cooler blocks the way. I could take everything apart but that would take many hours I don't currently have to spare.

I found out about this problem when creating a new overclock profile and smelled something, like burnt plastic. Now I smell this every time my computer is at load. I went back to stock and only on high load I can smell something burning.

I bought a new extension to try it before buying a new PSU. I could test it without the modded extension right now, but then I would risk melting the PSU to the mobo. I don't mind about the mobo because I already have to change it for another reason.

Remove your PSU and get it out of the case. Take a smell and see if the burnt plastic smell is coming from the PSU. Maybe try reconnecting it while it's outside the case and powering everything on. I wouldnt think that it would matter, as voltage should be applied to the rails even if nothing is drawing current, as long as the PSU is turned on. I would try to replace the EPS as well, however if the burning smell persists I wouldn't move forward without isolating the bad component. If you have a PSU tester, use it. If you have a multimeter, check all of the connectors to make sure they are all good.

Assuming your PSU isn't under warranty (and if it is, see if you can get them to replace it, due to potential fire hazard), remove the cover from the PSU (looks like 4 phillips head screws on the bottom where the intake fan is, and notice one should have a sticker about removing it voiding warranty). Take a look at the components, see if you notice any physical damage (such as burning, melting or even bulging capacitors. Take a close look at the rail connections to see if you notice any scarring or charring.

I hope both you and the OP luck out and are able to simply replace the connection and keep on trucking without replacing anything else, I'm just trying to stress the importance of avoiding potential fire hazard. Not likely mobo drawing too much current, as your PSU is designed to restrict current flow on each 12v rail. Either the connection or the PSU itself would be the culprit. Thanks again Phaedrus2129 for taking the time to explain the minor/major rail setup in the PSUs.
Here is what happened with my 8-pin cpu connector. No damage to mobo or cpu but the connector kinda melted. I am sure Corsair would reject this RMA due to connector damage
 WC Rig (13 items)
CPUMotherboardGraphicsRAM
i7-930@4.05GHz ASUS P6X58D-Preminum XFX GeForce 275 896mb 6GB Corsair TR3X6G1600C8D
Hard DriveOptical DriveOSMonitor
WD 150gb Raptor LightScan Win 7 64-bit Acer 22"
KeyboardPowerCaseMouse
G15 Corsair HX620W Antec 1200 several various
none
 WC Rig (13 items)
CPUMotherboardGraphicsRAM
i7-930@4.05GHz ASUS P6X58D-Preminum XFX GeForce 275 896mb 6GB Corsair TR3X6G1600C8D
Hard DriveOptical DriveOSMonitor
WD 150gb Raptor LightScan Win 7 64-bit Acer 22"
KeyboardPowerCaseMouse
G15 Corsair HX620W Antec 1200 several various
none
Quote:
Originally Posted by PCCstudent;15394301
Here is what happened with my 8-pin cpu connector. No damage to mobo or cpu but the connector kinda melted. I am sure Corsair would reject this RMA due to connector damage

Pretty much what happened to my connectors-top 4 pins melted plastic of the pins them selfs but not the houseing there in, and to let others know-i did not have an extention on when this happened. It was plugged straight into the mobo, and yes i do try my best to keep rails balanced propperly.
Now that i have the extention on the same psu, the only end that get realy hot is were the psu 8 pin plugs into the extention and only when extreemly stress in prime 95-ibt-linx-occt and others-even gameing it only gets warm.

I would RMA it-sparkle replaced my 1000watt with this 1250 watt, no questions, and now that it happened again they are replaceing the 1250watt, and am working with me to try and solve this prob.
 Dragon In Training (17 items)
CPUMotherboardGraphicsRAM
8350/4.8 Asus Crosshair v formula z Sapphire Vapor x 7950 Crucial BalisticTactical tracer-1866/9-9-9-24 O...
Hard DriveHard DriveOptical DriveCooling
Crucial m4 128gb Western Digital Caviar Black 2 TB SATA III 7200... Asus Burner Corsair H100
OSMonitorKeyboardPower
windows 7 ultimate 64 bit Asus ve258q Logitech G19 seasonic series x-1250
InWin Dragon Rider Logitech G 500 thermaltake lighted logitech 5.1 506
Other
SB 3D recon sound card
 Dragon In Training (17 items)
CPUMotherboardGraphicsRAM
8350/4.8 Asus Crosshair v formula z Sapphire Vapor x 7950 Crucial BalisticTactical tracer-1866/9-9-9-24 O...
Hard DriveHard DriveOptical DriveCooling
Crucial m4 128gb Western Digital Caviar Black 2 TB SATA III 7200... Asus Burner Corsair H100
OSMonitorKeyboardPower
windows 7 ultimate 64 bit Asus ve258q Logitech G19 seasonic series x-1250