really confusing power supply issue.

[_Grid21]

Hey all. I got a really really bizarre situation I am going to try to explain. Here we go. I have an AMD Ryzen 2700, an Asus ROG Strix X470-F, an EVGA 980 Ti FTW, and an EVGA G2 SuperNova Power supply. Also some extra storage drives, capture card, PCIe to USB expansion card, Corsair H115i AIO, and an internal USB hub.

 

Yesterday, I was at my PC, doing general things, when all of a sudden my computer blacked screened, then rebooted. I logged back into windows and checked the Event Viewer and I saw error code 41 Kernel-Power. I then went ahead and started running benchmarks to see if it would do it again and it did the same thing again, black screened, then rebooted. I then went and called EVGA and explained my issue, and told them that I saw in my bios that the 12v rail was showing a range between 12.2 and 12.1 on idle. Then they said that the 12v rail shouldn't be shifting like it was and said that it should be RMA-ed. Today I decided to do a test and I barrowed my dad's PSU from his gaming rig which is the exact same Power supply. I tested it for 30 minutes and I saw that the 12v rail on his PSU would dip to 12.1V running a CPU and GPU benchmark at the same time but the PC didn't black screen. Then I swapped back to the original power supply and I ran it for 35 minutes running Cinebench and Haven while watching a 4k video just to REALLY push my PC. My 980 Ti has 2 8-pin power connectors and during testing, 1 of the 12v plugs read as 12.1v and the second plug read as 12.2v and the main power plug for the PCIe motherboard 16x slot read as 12v. But it didn't Black screen after my 35 minute test. I am still going to have EVGA RMA my PSU, but I am very confused why the 2 power supplies would give 2 different readings between my dad's PSU and my PSU when their the same and further more, the 1st 8-pin power plug was pulling more watts then the second 8-pin power plug was showing. 

 

My PC has been running with it's original PSU for a few hours now and everything has been fine. What am I suppose to think? I don't know if I had a bios issue, running 5601 bios for my AMD Ryzen 2700, or if the motherboard got glitchy, or what? This isn't the first time I have seen my System do this, but this is the first time I have been able to do solid testing to monitor some real values. I am so beyond confused and VERY worried about it because I don't understand how to explain this or what is going on. Should I try updating my bios? I am not sure how stable 5603 bios is, or if this is Asus' fault? EVGA's fault? Further more, the fact that windows isn't even BSOD-ing, but just black screening with no real error reporting except for error 41 is pretty scary. 

Let me know and also let me know if you have the same system with similar issues or not. 

[aDoomGuy]

Hey bud. 

 

Well first of all error code 41 Kernel Power means that the system unexpectedly shut down, it doesn't give any reason. I assume you used software to measure your voltage and software doesn't do that very well. If you want to make sure what your voltages are the best to use is a digital voltmeter but you need to know how to use it. Unless you know someone that can help. As for what to do, I would either just replace the PSU or see how it goes. I would remove any overclock (including XMP/DOCP) to make sure it's not that (clear CMOS is recommended) and make sure I run the system completely stock until I am sure everything either works or not.

 

What I would not do is to update BIOS with possibly an unstable PSU, if it goes spastic and shuts down the PC during installation then you have a bricked motherboard on your hands.

[_Grid21]

1 minute ago, aDoomGuy said:

Hey bud. 

 

Well first of all error code 41 Kernel Power means that the system unexpectedly shut down, it doesn't give any reason. I assume you used software to measure your voltage and software doesn't do that very well. If you want to make sure what your voltages are the best to use is a digital voltmeter but you need to know how to use it. Unless you know someone that can help. As for what to do, I would either just replace the PSU or see how it goes. I would remove any overclock (including XMP/DOCP) to make sure it's not that (clear CMOS is recommended) and make sure I run the system completely stock until I am sure everything either works or not.

 

What I would not to is to update BIOS with possibly an unstable PSU, if it goes spastic and shuts down the PC during installation then you have a bricked motherboard on your hands.

Ok so I should just go ahead and let EVGA replace the Power Supply and for now, just keep using it? I do have a voltmeter, but you're right, I don't know how to use it on a PC. Is there a way to learn how to use it safely?

[aDoomGuy]

8 minutes ago, _Grid21 said:

Ok so I should just go ahead and let EVGA replace the Power Supply and for now, just keep using it? I do have a voltmeter, but you're right, I don't know how to use it on a PC. Is there a way to learn how to use it safely?

Yeah, I would probably just replace it but it depends on the inconvenience if not for anything else then it's a new PSU. I'm sure there are videos on Youtube that shows you how to do it. Have a gander at some of these.

 

https://www.youtube.com/results?search_query=voltmeter+pc+psu 

[_Grid21]

19 minutes ago, aDoomGuy said:

Yeah, I would probably just replace it but it depends on the inconvenience if not for anything else then it's a new PSU. I'm sure there are videos on Youtube that shows you how to do it. Have a gander at some of these.

 

https://www.youtube.com/results?search_query=voltmeter+pc+psu 

Do you think I took the right steps I did to test things? What else could cause a PC to randomly Black Screen, then reboot? Has AMD systems been known to have this issue? Is this a Windows issue? 

[Hackentosher]

it's possible there are transients on the 12v rail greater than the +/- 200mV you recorded that are occurring faster than whatever voltage monitor or voltmeter you're using can sample at. Larger transients may cause the hardware to freak out and shut down, but if you're reading fairly consistent 12V (usually +/- 200mV is fine for most electronics) with an external voltmeter I'm not so sure. To confirm, the power supply from your dad's computer passed the same stress test your power supply failed?

[aDoomGuy]

57 minutes ago, _Grid21 said:

Do you think I took the right steps I did to test things? What else could cause a PC to randomly Black Screen, then reboot? Has AMD systems been known to have this issue? Is this a Windows issue? 

It can be due to XMP. I had random reboots (just sudden black screen and reboot) until I manually set some XMP timings (notably TRFC including 2 and 4) as they wasn't loaded. Hence my recommendation earlier to disable and run fully stock.

[_Grid21]

13 minutes ago, aDoomGuy said:

It can be due to XMP. I had random reboots (just sudden black screen and reboot) until I manually set some XMP timings (notably TRFC including 2 and 4) as they wasn't loaded. Hence my recommendation earlier to disable and run fully stock.

 

1 hour ago, Hackentosher said:

it's possible there are transients on the 12v rail greater than the +/- 200mV you recorded that are occurring faster than whatever voltage monitor or voltmeter you're using can sample at. Larger transients may cause the hardware to freak out and shut down, but if you're reading fairly consistent 12V (usually +/- 200mV is fine for most electronics) with an external voltmeter I'm not so sure. To confirm, the power supply from your dad's computer passed the same stress test your power supply failed?

I do use DOCP, but this is a REALLY random event and rarely happens. I have 4 by 8GB Corsair Vengeance CMR16GX4m2C3000C15. It's clocked at 3000Mhz. But I don't really know how to manual OC something without breaking it so I just let the Motherboard figure it out. But because it is so random, I don't know how to narrow anything down which is why I felt I should just let EVGA replace my PSU since it's under the warranty still. 

[aDoomGuy]

2 minutes ago, _Grid21 said:

 

I do use DOCP, but this is a REALLY random event and rarely happens. I have 4 by 8GB Corsair Vengeance CMR16GX4m2C3000C15. It's clocked at 3000Mhz. But I don't really know how to manual OC something without breaking it so I just let the Motherboard figure it out. But because it is so random, I don't know how to narrow anything down which is why I felt I should just let EVGA replace my PSU since it's under the warranty still. 

My system did so randomly, by randomly I mean when gaming or browsing, watching video or even when left completely idle without user input or anything running except Windows and background tasks. It could happen anytime but not all the time. Maybe once or twice a day. It's been a year since it last happened since I fixed it but I'm sure there where days it didn't happen at all as well. If it is not inconvenient go ahead and replace your PSU, as I said if not for anything else you'll get a new PSU but if it's not the PSU causing it the problem will still remain. I would clear CMOS, run at stock and see if it happens again. If it does then replace the PSU and take it from there. By all means I would replace it but I would wait a few days to test the PC with stock BIOS settings and see if it happens again.

[_Grid21]

2 minutes ago, aDoomGuy said:

My system did so randomly, by randomly I mean when gaming or browsing, watching video or even when left completely idle without user input or anything running except Windows and background tasks. It could happen anytime but not all the time. Maybe once or twice a day. It's been a year since it last happened since I fixed it but I'm sure there where days it didn't happen at all as well. If it is not inconvenient go ahead and replace your PSU, as I said if not for anything else you'll get a new PSU but if it's not the PSU causing it the problem will still remain. I would clear CMOS, run at stock and see if it happens again. If it does then replace the PSU and take it from there. By all means I would replace it but I would wait a few days to test the PC with stock BIOS settings and see if it happens again.

That's fair. Is the Ryzen X470 Platform Known for this issue? Am I the only one to have this issue? I always feel like I have the problems no one else has and obviously I my motherboard, ram and CPU are NOT that old, I've barely had it for a year or 2 by now. 

[aDoomGuy]

3 minutes ago, _Grid21 said:

That's fair. Is the Ryzen X470 Platform Known for this issue? Am I the only one to have this issue? I always feel like I have the problems no one else has and obviously I my motherboard, ram and CPU are NOT that old, I've barely had it for a year or 2 by now. 

Well I had just built it when I had those problems.

 

I don't think it's a Ryzen or chipset (x470) issue (I had problem with B450) but I suspect it may be Asus issue. Not hard to fix manually though, to test if that is the problem just turn off DOCP and if the problem went away then head into BIOS look at SPD tab and take note of the timings listed for the (DOCP/XMP) profile you want to use (maybe take a picture). Then enable the desired profile and consult your picture or SPD tab and put in the timings that didn't change (if they have a value in SPD tab and value is set to auto or something else it probably needs to be changed).

[_Grid21]

4 minutes ago, Caroline said:

set it to VDC, connect the ground probe to any ground pin and the measuring probe to any live pin, in this case the 12V

you can also insert paperclips into the connector if your meter has clips instead of probes, just make sure you don't short anything and that the device is set to DC

 

about the blackscreen could be RAM

Ok, but how do I tell if it's the ram? I am only using ram that is verified for the motherboard and while I do have another set of DDR4 ram, it's not been verified for my motherboard.

[_Grid21]

2 hours ago, Caroline said:

what do you mean by verified? QVL? unless you're using RAM that's far off spec on voltage or speed any pair of sticks should work

Yeah I meant QVL, sorry I couldn't remember what the term was. I'll try again here. The Extra ram I have, isn't on Asus' QVL for my motherboard. I don't think using off QVL ram would be a good idea when I am trying to verify an issue.