Other Dedicated GPU issue or PSU or.. ?

[ignsvn]

Hi all, so I have this problem with my PC, with symptoms as follow:

For a while, when my PC goes to sleep, sometimes my dedicated GPU (a Zotac GTX 1660s) does not seem to wake up my monitor. Clicking a button or a key on my keyboard will wake the PC (you'll hear the usual HDD spins etc), but the monitor stays off.

When that happens, I usually turn off my PC, switch the HDMI cable to use my iGPU, and turn on the PC again using the iGPU. Checking my Device Manager (Win 10), the GTX 1660s is hidden. I need to make it appear by enabling the "Show hidden devices" in the menu. If I checked the 1660s' properties, it shows this message "Currently, this hardware device is not connected to the computer. (Code 45)"

One remedy is to turn off the PC, unplug the power from the wall for a while (say 30 mins?) and then re-plug & turn on the PC again. After that, usually I can use the 1660s again. And yes I was able to play games, such as Metro Exodus etc.

However, this doesn't always work. And when it doesn't work, I can only revert to my iGPU again.

For context, my PC is a mix of old & new-ish hardware. Here's my overall spec:

• CPU: Intel i5-4670
• Mobo: Asus H87i-plus
• RAM: Kingston DDR3 LP 8GB
• GPU: Zotac GTX 1660 Super
• Storage: Samsung Evo 860 SSD + WD Black HDD
• PSU: Corsair SF450 Gold
• Case: CCD MI-6

The PSU is never used to its limits. The most power hungry component it ever served is that Zotac GTX 1660s.

Anyone encountered something like this? What are the possible culprits? Something with my PSU, or an old & leaky capacitor in the motherboard, or perhaps something with CMOS battery, or riser cable, or..?

 

[msystems]

Similar has happened to me. I never figured out what the cause was but just tried not to leave unsaved work on screen and it was just occasionally frustrating.

As a workaround you could try disabling suspend to ram, and enable hibernation only. Or just disabling low power states entirely.

 

[ASCii]

Clearing the keys in the bios helped me with a similar issue when I ran Windows.

 

[ignsvn]

Similar has happened to me. I never figured out what the cause was but just tried not to leave unsaved work on screen and it was just occasionally frustrating.

As a workaround you could try disabling suspend to ram, and enable hibernation only. Or just disabling low power states entirely.

I already disabled both sleep & hibernation, and yes practically I set my PC to never sleep.

this lower power states you mentioned - is it the one in Window's settings or BIOS?

 

[ignsvn]

Clearing the keys in the bios helped me with a similar issue when I ran Windows.

which keys? can you elaborate more please?

 

[tinyitx]

I vaguely remember, from years ago, that this might have something to do with the Nvidia driver (I imagine you have already updated to the latest). If so, try disabling anything related to Nvidia HD Audio drivers. Or, if you do not need them, uninstall the Nvidia audio drivers altogether. No hard to try.

 

[ignsvn]

I vaguely remember, from years ago, that this might have something to do with the Nvidia driver (I imagine you have already updated to the latest). If so, try disabling anything related to Nvidia HD Audio drivers. Or, if you do not need them, uninstall the Nvidia audio drivers altogether. No hard to try.

oh. I haven't tried this. will try & see.

 

[ASCii]

which keys? can you elaborate more please?

Secure boot keys

On further consideration it's probably best not to from what I read about windows these days.

In a nut shell with my OS install & UEFI bios memorized various components of my install and tied it to keys, in my circumstance it recorded a particular video output, it would boot and display windows just fine, however accessing the bios, and various sleep states the screen would not display. this persisted through multiple installs, occasionally I could fix it by swapping display outputs, once I cleared the keys and performed a new install all was well.

If you wish to try this back up any data you don't wish to lose.

https://dlcdnets.asus.com/pub/ASUS/mb/LGA1150/H87I-PLUS/E8210_H87I-PLUS.pdf 2-39

 

[ignsvn]

Secure boot keys

On further consideration it's probably best not to from what I read about windows these days.

In a nut shell with my OS install & UEFI bios memorized various components of my install and tied it to keys, in my circumstance it recorded a particular video output, it would boot and display windows just fine, however accessing the bios, and various sleep states the screen would not display. this persisted through multiple installs, occasionally I could fix it by swapping display outputs, once I cleared the keys and performed a new install all was well.

If you wish to try this back up any data you don't wish to lose.

https://dlcdnets.asus.com/pub/ASUS/mb/LGA1150/H87I-PLUS/E8210_H87I-PLUS.pdf 2-39

Thank you @ASCii (and everyone else)

Will try the suggestions as soon as I have free time.

 

[AlexTSG]

I had this issue a few years ago with an Alienware X51 R2 and a Dell U2414H monitor.

Thanks to Alienware's excellent support they eventually persuaded Dell to exchange the monitor, and the problem was resolved. This was after they changed the graphics card twice, the motherboard, and the external power brick.

It turned out to be an issue with the DDC controller in the monitor. You may want to see if the monitor has an option to disable DDC and see if that helps.

During all the troubleshooting, and component swapping the issue did occur much less once they replaced the GeForce card with an equivalent Radeon, but it only went away completely after the monitor swap.

 

[ignsvn]

I had this issue a few years ago with an Alienware X51 R2 and a Dell U2414H monitor.

Thanks to Alienware's excellent support they eventually persuaded Dell to exchange the monitor, and the problem was resolved. This was after they changed the graphics card twice, the motherboard, and the external power brick.

It turned out to be an issue with the DDC controller in the monitor. You may want to see if the monitor has an option to disable DDC and see if that helps.

During all the troubleshooting, and component swapping the issue did occur much less once they replaced the GeForce card with an equivalent Radeon, but it only went away completely after the monitor swap.

I can confirm it's not DDC as it is always disabled

 

[Choidebu]

If it did not always do that in the past, then I'd strongly suggest looking into your PSU. Maybe get a spare or borrow from a friend, try it for a few days see if it does that again.

I had same issues for months until I replaced my dc-atx plugin board.

one more thing that comes to mind is riser card. MI-6 uses riser right? Can't remember

 

[ignsvn]

If it did not always do that in the past, then I'd strongly suggest looking into your PSU. Maybe get a spare or borrow from a friend, try it for a few days see if it does that again.

I had same issues for months until I replaced my dc-atx plugin board.

one more thing that comes to mind is riser card. MI-6 uses riser right? Can't remember

Yes, the PSU and the MI-6's riser cable are suspects as well haha.

It's actually the time for me to upgrade my PC as well, so probably I'll start by buying a PSU.

Btw, if I want to go with AM4 + B550 platform, am I right to say that Gigabyte's B550i Aorus is currently the best option?

 

[AlexTSG]

Btw, if I want to go with AM4 + B550 platform, am I right to say that Gigabyte's B550i Aorus is currently the best option?

The Gigabyte, MSI, and ASUS are all good boards with VRMs that are capable of easily supporting an overclocked 5950X.

Your choice should probably come down to features.

The MSI (which I have), supports front panel USB C, and is the only one with optical output for audio, but it has a non-standard backplate which limits your choice of CPU coolers. The active cooling is for the M.2 drive, not the VRM and can be turned off (although I've never heard mine over the CPU/GPU fans).
The non-MAX version has a terrible location for BIOS reset which requires you to remove the rear IO shield, which is screwed into the underside of the board.

The ASUS also supports front panel USB C, but has fairly limited rear IO compared to the other options. It's the only B550 with a thermal sensor header, which is great if you want to go for water-cooling. It has active VRM cooling. If the X570 ITX board isn't too much more, it's a better buy.

The Gigabyte doesn't support front panel USB C (which I wanted), and I wasn't very impressed with it's M.2 cooling solution (under very heavy loads), but I built 2 systems with this board last year, and they're both rock solid. I think it also achieved the highest sustained clock speeds for the CPU when Hardware Unboxed tested all the B550 ITX boards, although this could change with BIOS updates.

 

[ignsvn]

The Gigabyte, MSI, and ASUS are all good boards with VRMs that are capable of easily supporting an overclocked 5950X.

Your choice should probably come down to features.

The MSI (which I have), supports front panel USB C, and is the only one with optical output for audio, but it has a non-standard backplate which limits your choice of CPU coolers. The active cooling is for the M.2 drive, not the VRM and can be turned off (although I've never heard mine over the CPU/GPU fans).
The non-MAX version has a terrible location for BIOS reset which requires you to remove the rear IO shield, which is screwed into the underside of the board.

The ASUS also supports front panel USB C, but has fairly limited rear IO compared to the other options. It's the only B550 with a thermal sensor header, which is great if you want to go for water-cooling. It has active VRM cooling. If the X570 ITX board isn't too much more, it's a better buy.

The Gigabyte doesn't support front panel USB C (which I wanted), and I wasn't very impressed with it's M.2 cooling solution (under very heavy loads), but I built 2 systems with this board last year, and they're both rock solid. I think it also achieved the highest sustained clock speeds for the CPU when Hardware Unboxed tested all the B550 ITX boards, although this could change with BIOS updates.

Thank you. Those pretty much confirm my opinion.

I don't really overclock so any reasonably decent VRM will do. I'd love to have optical audio output, but unfortunately MSI screwed up with their non-standard backplate. That will pretty much leave me with either Gigabyte or Asus. I'm slightly leaning to Gigabyte for it's non-active VRM cooling - less potential point of failure is always good.

If I remember correctly, there should be an adapter for the front panel USB-C right?

 

[AlexTSG]

If I remember correctly, there should be an adapter for the front panel USB-C right?

Yes, if you don't have need for USB-A 3.1 ports, then you can get an adapter like this one to convert the header to USB C.

 

[ignsvn]

So it seems that @Choidebu is right - it's something with the riser cable, specifically the area where the cable is joined to the PCB that is connected to the motherboard.

It seems that the joint between the cable and the piece of PCB is getting weaker & no longer able to withstand the stress the comes with the bending.

I press the cable to the PCB by hand and it works for now, hopefully it'll take a while until the problem re-appears.

The good news is at least my PSU and other components are fine

 

[Meekyuu]

So it seems that @Choidebu is right - it's something with the riser cable, specifically the area where the cable is joined to the PCB that is connected to the motherboard.

It seems that the joint between the cable and the piece of PCB is getting weaker & no longer able to withstand the stress the comes with the bending.

I press the cable to the PCB by hand and it works for now, hopefully it'll take a while until the problem re-appears.

The good news is at least my PSU and other components are fine

Have you been fine since? Swapped out two 19cm linkup 4.0 risers, two motherboards, and three graphics cards and still have the issue.

 

[ignsvn]

Have you been fine since? Swapped out two 19cm linkup 4.0 risers, two motherboards, and three graphics cards and still have the issue.

Yes I've been fine (at least for now). If it happens again I will try to manually press/push the cable again toward the riser PCB haha. If the riser finally gives up then it means time to get a new one.

In your case, are you sure the issue is on the riser? Can you try to test with a different PSU?