I think Im experiencing a bad PSU, please come in

newls1

Supreme [H]ardness
Joined
Sep 8, 2003
Messages
4,607
Been using a EVGA 1300 PSU that I bought "B-Stock" about 6 years ago. Been through no less then 6-7 highend builds (3950x, 5950x, 10900K, 9900K, another 10900K, etc) Been on SLi'd r290's, 2080Ti, Titan XP, SLi GTX 980Ti's..... PSU has been rode very hard. Today, i just built another 10900K OC's and a OC'd 6900XT.. Was running Timespy, got 2 runs in trying some different OC's then finally on 3rd run, pc hard resets. Dialed back OC on gpu, restart timespy, insta reboot. well shit...... took OC completely off gpu, and timespy ran for about 45 seconds this time then pc reboot. My CPU oc is very stable, but took PC to completely stock settings for everything, then started timespy again, and timespy ran for 30-45sec, then pc reboot again! Would we all agree this PSU lived its last leg? Since i was pretty sure, i ordered a nice highend unit (Super Flower Leadex III 1300 white) hope i made the right call!
 
Replacing the power supply seems like a good call given the old units age. Worst case, the problem ends up being a different component and now you have a shiney new PSU that will provide many years of good performance.
 
It does sound like your PSU is on its last legs. Only way to tell for sure is to replace it.
 
I would really love for one of the PSU reviewers around here to take a high end but several year old PSU and see how the performance has changed over time. I also have some old PSU's that work.........I think.
 
I would really love for one of the PSU reviewers around here to take a high end but several year old PSU and see how the performance has changed over time. I also have some old PSU's that work.........I think.

I have some very recent personal experience with this. I just dropped a Core i9-10900K and an RTX 3090 FE in my rig. Everything is watercooled. Playing any other game or performing any other task and the system was fine. Playing Cyberpunk 2077, the system would just shut off. I used GPU-Z logging to see what might be going on and sure enough, anytime the board power exceeded 340 watts the system would shut off. My power supply was a Corsair AX1200 that I have had for about five years or so. It's powered everything from a Core i7 5960X to a Core i9-10980XE and finally a Ryzen 9 3950X and the 10900K. The 10900K and RTX 3090 FE proved to be too much for it. I swapped in a Seasonic 1200w Platinum and everything's running fine now.

That being said, when a power supply fails it normally presents the way I experienced it. That is, you get hard shut downs. You may also experience cold-boot issues. Random reboots can occur due to failing units, but its not what you normally see in my experience. In any case you can run GPU-Z or anything that can measure power consumption of some of your more demanding hardware components and figure out an application that makes the problem repeatable. Go over your data and it should be relatively easy to confirm. You can also put a multi-meter to the unit and look at your 12v readings. They should be within +/- 10% of 12v. If the numbers aren't, you need to replace the unit.
 
I have some very recent personal experience with this. I just dropped a Core i9-10900K and an RTX 3090 FE in my rig. Everything is watercooled. Playing any other game or performing any other task and the system was fine. Playing Cyberpunk 2077, the system would just shut off. I used GPU-Z logging to see what might be going on and sure enough, anytime the board power exceeded 340 watts the system would shut off. My power supply was a Corsair AX1200 that I have had for about five years or so. It's powered everything from a Core i7 5960X to a Core i9-10980XE and finally a Ryzen 9 3950X and the 10900K. The 10900K and RTX 3090 FE proved to be too much for it. I swapped in a Seasonic 1200w Platinum and everything's running fine now.

That being said, when a power supply fails it normally presents the way I experienced it. That is, you get hard shut downs. You may also experience cold-boot issues. Random reboots can occur due to failing units, but its not what you normally see in my experience. In any case you can run GPU-Z or anything that can measure power consumption of some of your more demanding hardware components and figure out an application that makes the problem repeatable. Go over your data and it should be relatively easy to confirm. You can also put a multi-meter to the unit and look at your 12v readings. They should be within +/- 10% of 12v. If the numbers aren't, you need to replace the unit.
I never thought i'd be able to kill this PSU... These super flower based psu's have done me well for years and years that ive been buying them (In evga form rebranded) guess anything electronic will die in due time and this 6900XT made her bleed out. Thanks Dan for your feedback, and your right.... I should have taken out my DMM and read the load on the lines while a game was going, but didnt think about that before I already removed PSU of PC, however tho, I always had GPUZ in background while running timespy runs and never was able to complete a run before pc hard resets, so i was never able to see load voltage/wattage. Anyways.... My new super flower 1300 will be here in a few days and am crossing my fingers this will be the fix. Almost has to be tho, i have no OC on this system to diagnose the issue, so my hopes are high!
 
I never thought i'd be able to kill this PSU... These super flower based psu's have done me well for years and years that ive been buying them (In evga form rebranded) guess anything electronic will die in due time and this 6900XT made her bleed out. Thanks Dan for your feedback, and your right.... I should have taken out my DMM and read the load on the lines while a game was going, but didnt think about that before I already removed PSU of PC, however tho, I always had GPUZ in background while running timespy runs and never was able to complete a run before pc hard resets, so i was never able to see load voltage/wattage. Anyways.... My new super flower 1300 will be here in a few days and am crossing my fingers this will be the fix. Almost has to be tho, i have no OC on this system to diagnose the issue, so my hopes are high!

You can set GPU-Z to log its data to a file and view it later. It wouldn't be much good without that feature. It can also pull some CPU data as well. When I reviewed the data on my machine I could clearly see that my temperatures were fine for everything, but anytime the graphics card's board power showed anything over 340w, the system died. It only did that running Cyberpunk 2077, which made sense since I am at 4K, using the RT ultra preset. Its rather demanding on both CPU and GPU. The problem also got worse rapidly. The system was fine for two weeks and then one day it happened once, and then from then on out I couldn't play the game for more than 10 minutes without the system shutting down. In contrast, I could play Destiny 2 for an hour or more and never had trouble. Looking at the data, the board power never went as high while playing other games.
 
I would really love for one of the PSU reviewers around here to take a high end but several year old PSU and see how the performance has changed over time. I also have some old PSU's that work.........I think.
I'm pretty sure one of the last HardOCP article articles was exactly that. Not sure how to access that content now though :/
 
Had 3 1000watt and above psus fail in the last year all in different systems and all over 4 years old except the ax 1200 2 years. Two failed to start after shutdown and one crashing system while gaming as soon as it loaded the 12v rail down to 11.8v 2020 can kiss my ass and gth. Good riddance
 
Been through no less then 6-7 highend builds (3950x, 5950x, 10900K, 9900K, another 10900K, etc) Been on SLi'd r290's, 2080Ti, Titan XP, SLi GTX 980Ti's..... PSU has been rode very hard.

Were you actually exceeding the capabilities of the PSU in any of those computers? Did you never clean it out? Do you have bad power at your house? Do you not use an AC during the summer and have very high ambient temperatures? I'm trying to understand why you think the PSU has been "rode hard". Sounds more like it's simply been "used". Nothing wrong with that.

Today, i just built another 10900K OC's and a OC'd 6900XT.. Was running Timespy, got 2 runs in trying some different OC's then finally on 3rd run, pc hard resets.

So you just built a new system, changing just about every component in the system. Sounds like there could be a lot of things that might be the cause. It worked fine with whatever system you had the day before, correct? Do you believe the previous system consumed significantly less power? Why do you think that the PSU would automatically start having issues? Because it's a handful of years old?

Dialed back OC on gpu, restart timespy, insta reboot. well shit...... took OC completely off gpu, and timespy ran for about 45 seconds this time then pc reboot. My CPU oc is very stable, but took PC to completely stock settings for everything, then started timespy again, and timespy ran for 30-45sec, then pc reboot again!

So you significantly decreased the power consumption of your computer, yet the problem persisted. If anything, I would take that as evidence that it's not a PSU issue.

Would we all agree this PSU lived its last leg? Since i was pretty sure, i ordered a nice highend unit (Super Flower Leadex III 1300 white) hope i made the right call!

Agree? Last leg? You have not provided any evidence or even a reasonable theory as to why you think the PSU might be having issues. Your post is basically "I changed everything in my system at the same time. Something doesn't work. Ummm..... PSU?"

Of course it *could* be the PSU, although it does seem like a bit of a coincidence that the PSU would fail at the exact same time that you changed almost every other major component in your system. Your best bet would be to start swapping components for diagnostic purposes. You really should always have at least one spare PSU, or at least be able to temporarily swap a PSU out of another system. Try putting in a less power hungry GPU (for testing purposes) and see if that changes anything. If it doesn't, that would at least eliminate the GPU from the equation as well as open the door to possibly swapping out a smaller PSU (maybe you have a spare 750w or similar laying around?).

6 years really isn't that old. Many PSUs these days even come with 10 year warranties. The PSUs in my main and backup system are both 1000w units from 2007 and have run both Triple-SLI and Quad-Crossfire at various points in their lives. There have been a few times over the years where I had unexplained reboots or similar failures. I've even swapped the PSUs between the systems on numerous occasions in an attempt to diagnose the issues. In the end, it was always something else that was causing the issue. It's not normal for a good quality PSU that is well cared for to die after only 6 years unless there was some sort of design flaw or it was seriously abused (and no, use =/= abuse). There are too many people out there that get caught up in this sort of odd obsessive paranoia that any PSU more than a few years old is some sort of ticking time-bomb and must be replaced ASAP. Don't fall into the trap. By all means, swap it out, test it, but don't assume it's a foregone conclusion. Good electronics that are well cared for can and do last multiple decades.
 
Only one of those systems was really close to full power capacity. The SR2 with dual 3.6 xeons and 2 vega 64's killed the ax1200 in just over 2 years.
 
I have some very recent personal experience with this. I just dropped a Core i9-10900K and an RTX 3090 FE in my rig. Everything is watercooled. Playing any other game or performing any other task and the system was fine. Playing Cyberpunk 2077, the system would just shut off. I used GPU-Z logging to see what might be going on and sure enough, anytime the board power exceeded 340 watts the system would shut off. My power supply was a Corsair AX1200 that I have had for about five years or so. It's powered everything from a Core i7 5960X to a Core i9-10980XE and finally a Ryzen 9 3950X and the 10900K. The 10900K and RTX 3090 FE proved to be too much for it. I swapped in a Seasonic 1200w Platinum and everything's running fine now.

That being said, when a power supply fails it normally presents the way I experienced it. That is, you get hard shut downs. You may also experience cold-boot issues. Random reboots can occur due to failing units, but its not what you normally see in my experience. In any case you can run GPU-Z or anything that can measure power consumption of some of your more demanding hardware components and figure out an application that makes the problem repeatable. Go over your data and it should be relatively easy to confirm. You can also put a multi-meter to the unit and look at your 12v readings. They should be within +/- 10% of 12v. If the numbers aren't, you need to replace the unit.

Unfortunate.

My close to 10 year old Corsair AX750 (Seasonic X-750) was running 24/7 for 8+ years in my gaming machine with a 1080ti. It was put back in its box about a year ago.
Last year I bought a 10700K + high end 3090 and thought it worth seeing if it was still in good working order.
So far it hasnt put a foot wrong, even when overclocked.
With a watt meter on the mains it averaged 620W sustained without issue.
A recent article on Igors Lab shows power peaks can be crazily higher so this PSU has held up remarkably.

ie on this page using a 5900X + 6800XT:
https://www.igorslab.de/en/when-the...red-and-evaluated-together-basics-practice/6/
Average load 365W, peak output 631W !! (20mS interval)

Overclocked:
Average load 400W, peak output 673W.

Both have max peaks around 70% higher than average load.
This would put my AX750 max peak load potentially around 1KW!
 
Last edited:
Unfortunate.

My close to 10 year old Corsair AX750 (Seasonic X-750) was running 24/7 for 8+ years in my gaming machine with a 1080ti. It was put back in its box about a year ago.
Last year I bought a 10700K + high end 3090 and thought it worth seeing if it was still in good working order.
So far it hasnt put a foot wrong, even when overclocked.
With a watt meter on the mains it averaged 620W sustained without issue.
A recent article on Igors Lab shows power peaks can be crazily higher so this PSU has held up remarkably.

ie on this page using a 5900X + 6800XT:
https://www.igorslab.de/en/when-the...red-and-evaluated-together-basics-practice/6/
Average load 365W, peak output 631W !! (20mS interval)

Overclocked:
Average load 400W, peak output 673W.

Both have max peaks around 70% higher than average load.
This would put my AX750 max peak load potentially around 1KW!

Well, mine was fine using the 3950X and the 2080 Ti which was factory overclocked. It was one of the higher clocked factory cards out there too. However, the RTX 3090 and a 10900K are an entirely different matter. Even then, the system was perfectly stable outside of Cyberpunk 2077 which is far more demanding than the other games I have tried. It hits the CPU unusually hard for a game.
 
Well, mine was fine using the 3950X and the 2080 Ti which was factory overclocked. It was one of the higher clocked factory cards out there too. However, the RTX 3090 and a 10900K are an entirely different matter. Even then, the system was perfectly stable outside of Cyberpunk 2077 which is far more demanding than the other games I have tried. It hits the CPU unusually hard for a game.
No doubt. I hold what you find in high regard.
My surprise is how well mine held up.
A fair amount of variance is expected between PSUs over such a long period (more with hugely different max power) but I wonder if there is mileage in keeping a PSUs caps at a more or less constant temp for most of its life ...
My PC has been mostly idle and at the same temp, and pretty cool as far as PCs go.
 
Last edited:
No doubt. I hold what you find in high regard.
My surprise is how well mine held up.
A fair amount of variance is expected between PSUs over such a long period (more with hugely different max power) but I wonder if there is mileage in keeping a PSUs caps at a more or less constant temp for most of its life ...

My Thermaltake Toughpower made it 10 years before I started having issues with it. I had a PC Power & Cooling 750 that lasted forever as well. My PC Power & Cooling 1Kw SR lasted 7 or 8 years easily as well. This Corsair AX1200 only lasting 5 is somewhat disappointing but honestly, I still think I got my money's worth out of it. The thing has had to power a Threadripper system and a Core i9-10980XE overclocked to 4.8GHz all core. Combine that with 1080 Ti's in SLI and then later a 2080 Ti, watercooling, etc. and it's had a hard life. When I ran the 10980XE in it, my configuration pulled about 1100w under load. I haven't been kind to any of these units. My configurations have always been more complex than average. Multiple GPU's, higher power draw CPU's, HEDT hardware etc. have been common in these setups. I change them quite frequently as well.

Good thing is I got this Seasonic 1200w Platinum unit for free from a friend. He hadn't used it long and he started having trouble. The 24 pin cable was bad but he had bought another unit for diagnostic purposes. So he replaced the unit and decided not to swap it back out. So, I had to buy all new cables for it. Effectively, this thing was $100. So, I can't complain.
 
  • Like
Reactions: Nenu
like this
just to update this thread as I forgot too, Replacing the eVGA 1300 with my new SuperFlower 1300 fixed this issue 100%... Also to note, this PSU is beautiful LOL
 
Back
Top