So I think I fried something

Liger88

2[H]4U
Joined
Feb 14, 2012
Messages
2,657
SOLVED: Turned out my top GPU fried from Folding@Home 24/7 for the past 1 1/2 years and the MOBO was smart enough to realize I ain't giving power to that PCI-E slot you idiot.

My computer randomly shut off around 5 minutes after a gaming session. It sounded like it was idling itself down doing a smooth Windows shutdown, but upon it instantly trying to restart itself I smelled a burning smell. Took it out to examine, dust it out, and the CPU heatsink smelled really bad.

Everything else smelled fine (well couldn't tell about memory because its proximity). No smells around the two GPU's (660/780 gaming and folding) so that sounded like hope. I immediately considered the CPU or MOBO as being fried. With nothing to lose I plugged it all back in and tried to start it. Spark near the CPU and a strong burning smell.

My speakers, monitor, external (with all my important data) seem just fine. Only the CPU area did I notice the spark and smell, but after I turned it on to see the spark it the MOBO completely ignores the power button.

Somethings dead. My guess is CPU or MOBO, perhaps both (worst case cross fingers). What are your thoughts?

Should I buy the exact same MOBO/CPU and then plug in the new MOBO to test the CPU and if its good return the new one? Do I just suck it up and get a cheaper MOBO/CPU just in case its both and would cost over $500 because its the Enthusiast class?

System:
CPU: i7 3930K (spark/smelled bad)
MOBO: Asus x79 Sabertooth (spark near CPU/smelled bad)
PSU: Corsair 1200 (smelled fine)
GPU: GTX 660 (smelled good)
GPU: GTX 780 (smelled good)
RAM: 4 x 4GB DDR3 (can't tell)
SSD: Corsair 480GB
 
Last edited:
I should also note since January our house has been having power issues and recently got resolved a month ago. We had an old power box that was known to start fires, and the circuit breaker was completely fried when we took it out and power often went out almost daily. Machine was always doing this random turn offs in the past month but turned back on with no problems. I'm wonder if that played a part.

It was all on a surge protector though.
 
If the motherboard sparked it is most likely what is dead.
The cause might have been the motherboard itself, a power issue or another faulty component putting voltage where it shouldnt be.
I would guess its just the motherboard, but it is a guess.

Do a test of the PSU voltages with a multimeter (unplug it from everything first).
You can power it up by connecting the green wire on the motherboards power connector to any black connection.
 
Thanks for the suggestion. Used a multimeter to test the voltages and the only plug that was working was the one that connected to the motherboard. The voltage on that was lower than the spec on Corsairs sheet. I ripped apart my dads machine to test the power supply on the motherboard with nothing connected and the heatsink fan for the CPU spun right up and so did the lights like nothing happened. Smells really bad still which has me worried about future reliability of the MOBO even if it works.

It wasn't good enough to test the GPU or CPU, but it was kind of odd it never made a beep without any power to the CPU. Didn't see the BIOS screen either, it was just a quick test to see if it was fried. At least I now know the PSU (Corsair AX1200) completely fried after 1 1/2 years. What a piece of shit for $300.

So I guess I'll get a cheap 850 overnighted from Newegg and go from there. I hope nothing got fried but that smell is strong. Even sticking to the back of my GPU's feeding paranoia.
 
Contact Corsair.
Their customer support is good.
Explain the result.
 
Thanks, I think I'm going to try that but I decided to overnight a Corsair 850 from newegg out of impatience. Maybe I can work out a situation with them when things are operational. I just have no faith that things are OK despite plugging in a new PSU and seeing at least the motherboard lights come on like normal because I didn't have a GPU or CPU plugged in to visually see.

Plus seeing smoke and a spark makes me think its unreliable even if it does work with just a PSU replacement when it comes in. I guess I'll see soon enough what else is broken if anything. Logic is telling me it can't be just the PSU from where I saw the spark/smoke.
 
At least I now know the PSU (Corsair AX1200) completely fried after 1 1/2 years. What a piece of shit for $300.
bad power kills power supplies don't blame corsair blame yourself for not looking into the power issue sooner.
Glad you caught it before it burned the house down.
 
I should also note since January our house has been having power issues and recently got resolved a month ago. We had an old power box that was known to start fires, and the circuit breaker was completely fried when we took it out and power often went out almost daily. Machine was always doing this random turn offs in the past month but turned back on with no problems. I'm wonder if that played a part.

It was all on a surge protector though.

At least I now know the PSU (Corsair AX1200) completely fried after 1 1/2 years. What a piece of shit for $300.

Blaming the wrong hardware :D
 
Everything else smelled fine (well couldn't tell about memory because its proximity). No smells around the two GPU's (660/780 gaming and folding) so that sounded like hope. I immediately considered the CPU or MOBO as being fried. With nothing to lose I plugged it all back in and tried to start it. Spark near the CPU and a strong burning smell.
Most likely the power mosfets for the cpu, bad power kills the power supply and it killed the mosfets when it puked.
It happens.
Id say 50/50 chance the cpu is toast also.
Check the mosfets for burn marks before putting the cpu back in and trying to powering it up.
I have had a mb with one of 4 blown and it would post and boot into windows but die under load when the 3 remaining mosfets overheated.
It could have easily took out the new power supply cpu and other components.
Now I check and if I am one bit worried about them I rma or toss the board.
 
bad power kills power supplies don't blame corsair blame yourself for not looking into the power issue sooner.
Glad you caught it before it burned the house down.

I had no control over the housing power issue, but took the steps necessary to prevent damage as best I could under the extreme circumstances. It may or may not have played a roll, with almost anything power related the one thing I've learned is you never know. I simply cast blame out of momentary anger on that particular model remember one of the last couple reviews I saw before I bought it way back then related to it literally burning out. I figured I'd never be that guy.

Could happen with any PSU really and could have been unrelated to it entirely. The housing power issue really made me react differently than I normally would have.

Blaming the wrong hardware :D


Well I did neglect to say it didn't turn right back on. I had to wait a couple seconds or more before it would turn back on, so in retrospect it kinda hinted a PSU issue. At this point I just have to wait for the new one though lol. I still stuck with Corsair because my old 750 lasted me 5 years under more harsh conditions with never an issue.

Most likely the power mosfets for the cpu, bad power kills the power supply and it killed the mosfets when it puked.
It happens.
Id say 50/50 chance the cpu is toast also.
Check the mosfets for burn marks before putting the cpu back in and trying to powering it up.
I have had a mb with one of 4 blown and it would post and boot into windows but die under load when the 3 remaining mosfets overheated.
It could have easily took out the new power supply cpu and other components.
Now I check and if I am one bit worried about them I rma or toss the board.


That's what has me worried. If I plug a NEW PSU that I ordered into everything and load up all the power cables do I run a risk of burning out anything else? If it's one thing I suck completely at its anything related to Power/PSU/MOBO's.

I haven't yet taken out the CPU just to avoid the painstaking process to visually inspect it, but I'll probably do that tomorrow since you gave me some new info to go off of. I didn't notice anything that looked burned around the CPU or the rest of the MOBO with the CPU still installed, but I don't really know what the MOSFET is until I just heard it lol.

I'm just so damned confused thinking the testing PSU plugged just into the motherboard plug should have confirmed the motherboard was dead, not lighting it up.

Thanks for the ideas and help. Right now all my normal "what if's" aren't making sense.
 
That's what has me worried. If I plug a NEW PSU that I ordered into everything and load up all the power cables do I run a risk of burning out anything else? If it's one thing I suck completely at its anything related to Power/PSU/MOBO's.

I'm just so damned confused thinking the testing PSU plugged just into the motherboard plug should have confirmed the motherboard was dead, not lighting it up.
Yes it can blow the new power supply and anything connected.
Do not hook the drives or cards up until you do more troubleshooting.
All the lights do is say there is voltage to the board.
If it did not post with the testing supply there are issues.
 
Yes it can blow the new power supply and anything connected.
Do not hook the drives or cards up until you do more troubleshooting.
All the lights do is say there is voltage to the board.
If it did not post with the testing supply there are issues.


Dang it. I don't know where to go from here now. I just took off the heatsink and removed the CPU and memory. Still notice NO signs of damage, burns, or leaks. The bottom pads of the CPU looked fine, the pins on the motherboard looked fine, the metallic square boxes near i looked fine, and all the little block looking things on the CPU looked fine in the middle.

I hope I didn't inadvertently blow someone elses PSU by plugging it into the motherboard. The PSU does have a faint smell to it, but not the same kind as the burning smell from where I saw the spark/slight smoke coming from that stunk up all the parts inside, and that whole area again looks like nothing happened.

I have no way to confirm it got to POST or not because I didn't have the CPU power plugged in,nor a GPU when I used a third-party PSU to see if the MOBO at least received some kind of power. The only way I can confirm it is by using the new PSU to test, but if you say that can blow then I'm back to the catch-22. Do I possibly throw away good parts in a freak accident and just buy a cheaper board/cpu combo or bother knowing for sure with the new PSU.

I feel like I'm at odds end with no right answer. Everything is saying just replace the board and cpu and suck it up, but all my troubleshooting is giving me contradictory responses. I expected my dads PSU to confirm the motherboad was toasted and to see damage and yet I see nothing anywhere on any part.

I trust you guys. Just tell me want to do. :\
 
Still notice NO signs of damage, burns, or leaks. \
Most all electronic failures have no visual indication of damage. Those indicators (bulging cap, smell) are the rare exception.

Had you measured voltages on six key wires and reported those numbers here, then much could have been posted as to what failed, why it failed, and what else might have been damaged. That means measuring without disconnecting a single wire or part.

BTW, if an AC electric anomaly damaged a supply, then also listed are other less robust and also damaged items including dimmer switches, smoke detectors, clocks, and bathroom GFCIs.

If a PSU was properly designed, then it did not do any damage to any other part. And no other part damaged that PSU. But again, to say more means three digit volt numbers so that fewer who know this stuff could then post something useful.
 
Most all electronic failures have no visual indication of damage. Those indicators (bulging cap, smell) are the rare exception.

Had you measured voltages on six key wires and reported those numbers here, then much could have been posted as to what failed, why it failed, and what else might have been damaged. That means measuring without disconnecting a single wire or part.

BTW, if an AC electric anomaly damaged a supply, then also listed are other less robust and also damaged items including dimmer switches, smoke detectors, clocks, and bathroom GFCIs.

If a PSU was properly designed, then it did not do any damage to any other part. And no other part damaged that PSU. But again, to say more means three digit volt numbers so that fewer who know this stuff could then post something useful.


My dad tested the PSU using a multimeter, but again I don't know if it was done right. It's something neither of us have ever done. They're all black cables with no green ones visible as the previous poster implied making it more difficult.

We tested the PSU with cables attached and without, separately one at a time, and the only one showing any signs of life was the port on the PSU > MOBO. All other 6 and 8 pin connectors didn't provide any current. The only number we saw was like 5.17 then something like 4.7x on a restest. Again someone told me you might need to keep the motherboard plugged in to test the others. I don't know if that's true, but we did each port one by one.

The new PSU comes in tomorrow and I guess I have nothing to lose testing it with a return policy. I don't think it'll blow it out because it didn't do that when I used someone elses to run a quick test.

I wish I had numbers to give and not be a n00b when it comes to this. Nothing turned on not even the PSU fan, which is not passive IIRC. Sorry I can't be of more help. Worse come to worse I can at least do further testing tomorrow and keep people posted. I already am writing off the motherboard/cpu as a loss regardless and the only real worry is blowing a new PSU and playing the wait game to do the full investment.

FYI I should say the power issues have been resolved house wise for the past month with the new breaker box.
 
I wish I had numbers to give and not be a n00b when it comes to this. Nothing turned on not even the PSU fan, which is not passive IIRC
All that is determined by a power controller. A defective power supply can also act good if not connected to everythng else. Since the power controller is not connected to see the defect. And since defective power supplies can look good when not connected to a load or if connected to the near zero load found in power supply testers. Just a few concepts to learn for the future.

Is the new supply good? Again, a defective supply can still boot and run a computer. But again, a meter may identify a defective supply that fails later whent he warranty expires. Just another reason for doing what was recommended. Get your father involved so he too can learn this stuff.

BTW, was your breaker box from Federal Pacific? Those have a long history of causing house fires. Mentioned because all lurkers can (should) learn from your experience.
 
All that is determined by a power controller. A defective power supply can also act good if not connected to everythng else. Since the power controller is not connected to see the defect. And since defective power supplies can look good when not connected to a load or if connected to the near zero load found in power supply testers. Just a few concepts to learn for the future.

Exactly, that's what my buddy who worked at Fry's was trying to tell me. Testing a PSU can be very tricky if you don't know what you're doing even with a multimeter because of their operation, but I figured getting nothing back was not a good sign. I'll keep that in mind and once this is resolved study up more to actually learn this dreadful topic.

Is the new supply good? Again, a defective supply can still boot and run a computer. But again, a meter may identify a defective supply that fails later whent he warranty expires. Just another reason for doing what was recommended. Get your father involved so he too can learn this stuff.

the *NEW* Corsair HX850W PSU coming in tomorrow should definitely be good. My dads work PSU was definitely good when I used it to plug only the cable from the PSU to the motherboard and noticed the lights turning on for a minute with no issue. The only thing I accidentally left plugged into the motherboard was the CPU fan and that was spinning like a champ. As I said before though it was a quick test with NO GPU/CPU power plugged in so I couldn't visually see the results. Motherboad could be erroring for all I know. The old Corsair AX1200 would do nothing when hitting the power or during testing except for one plug.


BTW, was your breaker box from Federal Pacific? Those have a long history of causing house fires. Mentioned because all lurkers can (should) learn from your experience.


Haha you nailed it. It was Federal Pacific and the circuit my room was on was toasted visually. That circuit was fried first then it started frying both the top, bottom and the circuits directly across top and bottom. That's what caused the immediate response when the whole house was getting knocked out for 4-5 hours while it cooled.

If you have a Federal Pacific in America I highly suggest replacing it folks. It is not worth the issues it has caused this household and many others.

Once more thanks for your time and patience dealing with my ignorance on this matter and calming me down a bit lol.
 
Please read what was posted. Best 'study up' is a meter, requested instructions, one minute of labor, and post numbers for a reply that answers all in layman terms.

Otherwise take a few semesters of Electcrical Engineering. Nobody on the internet is going to teach you that. Again, otherwise you have silenced your best assistance.
 
If you have a Federal Pacific in America I highly suggest replacing it folks. It is not worth the issues it has caused this household and many others.
That type failure would not nothing harmful to the computer.

But again, everyone reading this should be trooping to their breaker box. Since that major human safety defect is 100% the responsibility of the homeowner.

When Schneider Electric bought Federal Pacific, they discovered outright fraud. And informed government agencies of this widespread human safety defect. Of course, some governments are run by people who fear responsible action. So many homeowners are stuck with the problem.

This failure would not damage any appliances. The fire might.

Another company has done same. Schneider bought APC. Then discovered some APC surge protectors are so dangerous as to be removed immediately without delay. But again, how many still recommend APC surge protectors (and equivalent products from Belkin, Panamax, Triplite, Monster, etc) without even understanding this well understood threat or how electricity works.

From the Washington post of 13 Jan 2007:
Federal Pacific "Stab-Lok" Electrical Panels

These electrical panels, widely used by builders in the 1960s and 1970s, are considered unsafe. The problem is usually described as a latent hazard -- that is, ... the breakers often will not trip when they should. That can lead to overheating and fires.

In extensive testing by the Consumer Product Safety Commission, the circuit breakers had a failure rate of approximately 25 percent.
 
A degree in electrical engineering is a bit excessive, though a few college courses (or a short lecture from someone in the know) on the subject wouldn't hurt.

Since nobody wants to provide any direction except "do this--if you don't know how, you're beyond help," I will.

Here's the ATX 24pin connector's pinout (a list of pins and what voltage/signal they should carry, a troubleshooter's best friend and an EE's best kept secret). Touch one probe to any of the ground pins, and then touch the other one to each of the other pins and record the values as you go. If they are inverted (should be getting 12v but you're getting -12v), then you have the wrong probe on ground...or there's something seriously wrong with your PSU. You won't get much of anything as far as voltage unless you have pin 16 shorted to ground. Good luck. ;)
 
Haha thanks for that map. I finally just YouTubed it and winged the paper clip and testing needles based on his. I wasn't sure if the layout was standardized or not and these all black connectors and cables *sigh*. I did not know you needed to apply a load (i.e. used a case fan connected to an 8 pin) and apply a paper clip to simulate the MOBO which was what we were missing from the start.

It turns on and is testable. Might have just wasted $150 on a new PSU if it isn't bad now. Oh well.

These are the 6 different numbers I got

1.68
5.04
.12
7.23
17.37
3.36
They all reverse to the same value too. Will go back over and do it the right wave to make sure the pins are getting the exact + or - they should be, but do those numbers look "OK" at least?
 
Last edited:
Don't know where you're getting 17.37 from...that seems way off. Should be 12, -12, 5, -5, 3.3. If you can list the pin number with each value, that'd also be helpful.
 
Don't know where you're getting 17.37 from...that seems way off. Should be 12, -12, 5, -5, 3.3. If you can list the pin number with each value, that'd also be helpful.

He will have measured -5V to +12V.


Yup I winged it from the YouTube video. Now with the map we got something more, uggh, reasonable? lol

Pin 1/2 - 3.36
Pin 4/6 - 5.04
Pin 8 - 4.92
Pin 9 - 5.15
Pin 10/11 - 12.29
Pin 12 - 3.36
Pin 13 - 3.36
Pin 14 - -12.35
Pin 21-23 - 5.04

So did I just waste money buying a power supply prematurely or should I start with ordering a new motherboard and cpu?
 
PSU seems ok on a basic level at least.
it could have been just a motherboard failure.

Be very careful testing motherboards.
If you bend a pin in the CPU socket, its not covered under warranty unless they are very nice to you.
I had a motherboard denied warranty by Asus like this.
Gigabyte replaced a motherboard with bent pins.
 
Those numbers do not report what is would create an informed reply. One item was listed many times and ignored. You Tube definitely was not on the list.

Power supply must be connected to everything. Nothing in a computer can be disconnected. Otherwise a power system (more than just a PSU) may not report its defect. Pin 16 was also important.

Testing did not find any defect. And would not identify some defects. IOW measuring that way cannot say a power 'system' is good; can only identify some defects..
 
@westom: you're the only one so far who's suggested he hook the PSU up to everything while testing it, and you haven't said how to test it while it's all hooked up--it's hard to read the voltages off of a pin if it's connected and shielded by plastic, correct? Now, I'm sure there are ways, like taking a disconnected 4-pin molex and reading the voltages off of it, but you have to have some method/procedure or the results may be useless. For example, if you measure the voltages but don't know what rail they're from, you couldn't tell whether the measured voltage drop is normal or not. I'm sure there are other peculiarities that should be considered when measuring and recording results, so that they are structured in a way that has meaning--it would be helpful to know this before he just starts randomly poking his multimeter around and then you get upset because he doesn't give you the information that you wanted.
 
I'm not plugging it back into the motherboard when it refused to start last time it was fully plugged into everything. That seems silly risky when it burned twice while plugged in and the PSU started. I couldn't test Pin 16 because that's what was used via the paper clip, which now conflicting info is being said was a useless test.

At this point I'm basically getting that I should keep the new PSU coming in tomorrow and order a cheap $350 mobo/cpu combo or waste more time trial and erroring on a single $300 mobo and possibly a $500 cpu. $800 which I cannot afford and can't really afford what its costing me now. I appreciate all the help, I really do, but time is actually crucial else I would not have overnighted a new PSU in the first place.

What I was hoping to get from all this was whether or not its worth downgrading for a cheaper $350 mobo/cpu combo or whether to try replacing either the CPU or MOBO because I couldn't afford both. The PSU was a side issue, which regardless is solved with a new one in-route. I could still try plugging that into the system when it comes but then people tell me things are going to explode. I'm being pulled in every direction lol.
 
I would RMA the PSU if you can afford the wait, and get a new MB. Since you've already ordered a PSU, you can either sell that one or the one that came back via RMA. I'm not sure I would trust the old MB, and I wouldn't feel right selling it like that, but you could probably get a few bucks on ebay (just be sure to mention in the description that it may have been damaged by a faulty PSU or some other cause)--plenty of people will buy used/damaged electronics just for the parts on them.

I don't know whether a damaged CPU can damage a MB or not--anyone else know?
 
Any damaged component holds the possibility it could damage what it is connected to.
You rely on luck when something dies badly and want to recover hardware (unless you have diagnostic gear).
But if the PSU is ok, the odds are that it was a local failure and nothing else will be damaged.
ie the weakest point died when the mains went tits up.
 
I took the plunge and next day ordered the new Asus Sabertooth X79 hoping that it's just localized and didn't fry the RAM or anything else connected to it. If it turns out the CPU is gone too I'm screwed. With my bad luck this entire year most likely I should have just ordered a cheap $350 high-end mobo/cpu combo.

Once it gets up and running I'll deal with Asus, Corsair, and hopefully NOT Intel and see if I can get some kind of check to reimburse me. Taking chances gents. At least with the same hardware getting up and running shouldn't be a problem (backups) if I had to.

Hopefully next-day with Newegg means it comes tomorrow.
 
He will have measured -5V to +12V.

The presence of -5v is a problem in and of itself. It was banned from the ATX12V Power spec as of version 1.3, which was in 2003 (12 years ago.)

So his PSU is either over 12 years old, or it's a terrible Asia-X brand that still uses that ancient specification. Quality units don't even have the -5v wire in the Minifit Jr. ATX motherboard power connector anymore.
 
I know wikipedia isn't always right, but it never says in any version that -5v was banned, only "Guidance for the −5 V rail was removed (but it was not prohibited)." Still, don't see many modern PSUs which have -5V, and he listed 21, 22, and 23 as -5, but they should be +5.

I think it's more likely that he measured +5v to +12v, which would both be present on a modern PSU. ;)
Edit:...-5 to +12 would equal (+/-)7, whichever leads you used to measure it, anyway. o_o
Edit2: no, nvm, reading voltage differences with a multimeter is weird. lol
Still, more likely he measured +5 to -12, since those are actually present on modern PSUs.
 
Last edited:
I'm not plugging it back into the motherboard when it refused to start last time it was fully plugged into everything. That seems silly risky when it burned twice while plugged in and the PSU started. I couldn't test Pin 16 because that's what was used via the paper clip, which now conflicting info is being said was a useless test..
Appreciate what has happened here. How many designed power supplies? Apparently only one. As made obvious by so many posts even without numbers. Some fundamental facts apply that would only be known to the fewest who know how this stuff works.

Anyone who knows how power supplies work also knows why defective supplies act perfectly good when not connected to the read of its power 'system'.

First, if damage could happen, it has already occurred. In the time it takes for you decide to move a finger, damage is already done.

Second, no properly designed aupply can damage a motherboard. An no motherboard failure can harm a power supply. That standard existed long before PCs existed and long before the ATX standard said same.

For example, short all PSU outputs together. Then power on a supply. Many 'fear' that will damage a supply. Total nonsense made obvious from knowledge of how a PSU works. In fact Intel even defined how thick that shorting wire must be to short all power supply outputs together - and have no damage. How many know this ... and why with numbers.

Having removed even one wire has already complicated a solution. How many days and posts ... without even one fact. Everything still remains speculation ... "it could be ..."

Defined with four specific requirements was a useful reply. Have said so repeatedly. And still so many do not read with sufficient care. Another has just posted same because he id not read with care. The expression included, again, "requested instructions". Posted specifically because a useful reply means one cannot vary from a good solution. For example, that YouTube instruction does not provide sufficient information. A problem often found with computer techs who do not know how a PSU works, its many functions, and how electricity works. Which is why a majority can only recommend shotgunning.

At least ten different reasons have been provided that says which posts to pay attention to and which are only speculation. But again, one even designed this stuff even before PCs existed. Why then does one say something completely different from others? Same person also says recommendation without numbers is best considered bogus. That explains so many days and so many posts ... without one definitive answer. Even a new PSU is only speculation.

-5 volts is known why it exists and why it is not relevant to your situation. No definitive answer on -5 volts is due to not knowing what the power 'system' (PSU is only one part) does.

Four items were listed to have a complete and useful response. It is, without doubt and by far, your best recommendation. Each fear or concern was addressed and deflated by facts from one who would even cite numbers from datasheets. How many others learned how power supplies work by even reading manufacturer datasheets? Just another reason why one who knows this stuff says something completely different from so many others who did not design electronics.

Posted repeatedly was the only solution that eliminates all those speculations in all those posts. Including the latest discussion about something obviously irrelevant - if one knows how digital electronics work.

Why -5 volts is irrelevant also becomes obvious if carefully following (and not varying from) requested instructions. Notice the You Tube instructions did not make it obvious. Just another reason to separate speculation from a definitive solution.
 
I know wikipedia isn't always right, but it never says in any version that -5v was banned, only "Guidance for the −5 V rail was removed (but it was not prohibited)." Still, don't see many modern PSUs which have -5V, and he listed 21, 22, and 23 as -5, but they should be +5.

The only thing that used -5v was the ISA bus, and even then it was only really old ISA cards that used it. The spec that I read stated -5v was forbidden/banned on either 1.3 or 2.0 of the ATX12V spec.

-5v used to be where pin 18 is, which is N/C now. In any case, if a PSU still has -5v, it's immediately indicative of an extremely old PSU design, usually a terrible Asia-X brand if it's brand new. Though in recent years I've seen Asia-X brand PSUs not connect pin 18, but they still have -5v available on the PCB.

Edit:...-5 to +12 would equal (+/-)7, whichever leads you used to measure it, anyway. o_o
Edit2: no, nvm, reading voltage differences with a multimeter is weird. lol
Still, more likely he measured +5 to -12, since those are actually present on modern PSUs.

Going from -5v to +12v will yield ~17v taking into account the negative rails don't have to be as tightly regulated as positive rails (10% vs 5%). If he indeed registered 17v from somewhere then the PSU has a -5v rail. If he measured from -12v to +12v, it'd be around ~24v instead.
 
@westom: you're obviously an electrical engineer...because your writing is really difficult to understand (aside from a few things here and there). No offense intended, of course--my dad was very engineering-oriented (more mechanical, though), and he doesn't write very clearly either. I would really appreciate it if you pointed out those instructions you keep mentioning, because I've gone through this thread a number of times and couldn't find them. I'm also dyslexic, so that may have something to do with it.

@GiGaBiTe: Couldn't he have measured across +5 and -12 and had the same result (difference of +5 and -12 is 17)? Whether the sign is positive or negative depends mostly on which lead was on which pin.
 
I'd just like to keep you updated. Got the new MOBO and PSU in and it still wasn't turning on. Pulled the parts out one by one and found out it was the F@H GPU I had cracked 24/7 for the past 1 1/2 years that shot out, which explains why the spark/smoke was between the CPU and top GPU. It never occurred to the GPU would have prevented the starting up, but it makes sense with how resilient hardware is now built.

It's still within the 3 year mark to RMA with EVGA I'm told and I can return the MOBO to Newegg which I'll call them tomorrow to deal with so the only money I'll be losing is on the PSU, which is never a bad thing to have a backup.

Thanks all for your support. My impatience and fear bit me on this one. If I happened to annoy some of you I apologize as it was probably my frustration of things not making sense. I've never had a GPU burn out to know what would happen to a system and now I know lol. Your time was well used assisting me.
 
Appreciate what has happened here. How many designed power supplies? Apparently only one.
How did you know I have designed several power supplies? :}


Second, no properly designed aupply can damage a motherboard. An no motherboard failure can harm a power supply. That standard existed long before PCs existed and long before the ATX standard said same.
Properly designed and built and not damaged being key and many aren't even from the big names. I have seen several cases where the power supply took out the motherboard and at times other components in the real world.
There have also been a lot of cases where both were dead and there was no evidence which took out which.
 
For example, short all PSU outputs together. Then power on a supply. Many 'fear' that will damage a supply. Total nonsense made obvious from knowledge of how a PSU works. In fact Intel even defined how thick that shorting wire must be to short all power supply outputs together - and have no damage. How many know this ... and why with numbers.
No Dont.
Do that in the real world and I would expect at least 25% of them to be dead afterwards.
And yes we used to take pulled power supplies and do that to them as well as other cruel things for fun.
 
just an fyi we usually tested power supplies with a ~10A load on the 12v and a ~5A load on the 5v and ~1A on 3.3v then measured the voltages if the supply was rated that high.
At no load at times they would be fine but under load they would either quit or drop below spec.
 
@westom: you're obviously an electrical engineer...because your writing is really difficult to understand (aside from a few things here and there). No offense intended, of course--my dad was very engineering-oriented (more mechanical, though), and he doesn't write very clearly either.
If something is new, then it is never understood in a first or second reading. Eventually, some learn that concept in school. It was an important theme in the movie Paper Chase.

Obviously what I wrote was completely new to most all. Everyone (including me) never see everything in the first reading. Those instructions were defined specifically. "Requested instructions". An important point I keep making. Many did not see the word 'requested'. Therefore no one has requested those instructions.

Appreciate why so many complain it is too long or too hard to read. That is often the first indication that it is actually saying what happens (is not a soundbyte half truth). Or that it is saying something completely new.

Also demonstrated by this thread is why shotgunning is a poor diagnostic procedure. The OP replaced how many good parts? Spent all that time. So many posts included speculation. And none (other than his original notation of a spark and how it would not start) provided useful information.

Concept was well defined in a series called CSI. "Follow the evidence". Rereading this thread demonstrates how many did not understand the concept. Many simply replied with wild speculation identified by expressions such as "try this ... " or "maybe .... has failed". "I feel this must be the only reason for failure."

Better techs would not yet be done with the OP's system. Requested instructions also identify the power 'system' (more than just a PSU) as good. Normal is for defective components to still boot and run a system. And then fail days or even a year later. Those instructions not only teach how a computer works. And why things are done. But also identify a defect long before its warranty expires.

Why would you know any of this is true? Other that what the OP just posted, you have no reason to know. Because better diagnostics procedures is learned from experience (and mistakes). For example, if one learned from this thread, then one knows the difference between shotgunning and "follow the evidence".

Honestly, most people never learn good diagnostic procedure (how to think through a problem) even though we teach those concepts in junior high science. From history, we learn examples such as the many who just knew cigarettes are smoked to increase health, health store supplements (that do not even contain what they claim to provide), the many who just 'knew' Saddam had WMDs, and even magic ingredients found in yogurt.

So yes, much of what I posted will be hard because the concepts (ie answers must have numbers) is foreign to so many.
 
Last edited:
Back
Top