Aurora R6, hardware error, Nvidia GTX 1070

Question

After 5 years of no problems ... about 2-3 times per week lately, monitor goes black for a few seconds. Then, the image re-appears and it runs fine (for a day or so). Within a week, the multiple days of critical errors stair-stepped my rating down to a 1 or 2.
Every time it happened, I was getting some "Hardware Error" critical errors in the Windows-11 Reliability History Report.

Before - Error

So, the above screenshot is after I reset the Reliability Monitor report. For some reason, I can't find the original screenshot with all the errors. But imagine that error above, every few days, stair-stepping my rating down to a 1.

Detail.

Before - Error (Detail)

Blew-out whole Aurora-R6 with air-compressor (including fans and radiator).
Removed Dell-OEM Nvidia GTX-1070 Back-blower (in 5-years, it's never been removed before).
Tried not to touch circuit board or gold-contacts on slot edge.
There is a light-blue-plastic keeper-lever on the far-side of the PCIe-slot that might need some help in both directions.
While removed, carefully cleaned card (and its fan) with small brush and more air (40psi from all angles).
Reseated GPU-PWR connector (on motherboard) and some others cables.
Checked that all components and connectors are secure and fully plugged-in (including modular PowerSupply).
Inspected GTX-1070 card under magnification. Looks fine. Uses one 8-pin PCIe-Power-Cable.
Cleaned GTX-1070 PCIe edge-connector with clean-cloth .
Re-installed and re-Seated Nvidia GTX-1070. Inserted-back Dell's black-plastic brace toward last 2mm final placement down into slot.
Re-assembled Aurora-R6. Helps if you reconnect all USB cables in same ports as before (I took some pics first).
Boots and seems fine. Device-manager is correct and clean.
Did a Reset on Reliability History Report.
After 14-days, never again saw the error-event, or had any problems with video or video-card.
- Re-seating video-card worked.

That's all I did (I didn't reload the drivers or anything else). It's still working perfectly without a single error.

After- Working fine since

Iain K Mackay · Answer

Windows and Windows 11 in particular likes to tell you your issue is hardware related. I had black screens on a 12900KF and a 3090GTX, after the last big Windows 11update, solved by rolling back..

First thing I would be doing is a clean install of Windows. Maybe your card is fried, but I would not be relying on Windows to tell you it is. Windows is without doubt the last place I would look. The settings menu is now a morph of at least 4 Windows versions, with added keyboard presses, just in case you didn't get the memo a new version of Windows is out.

Tesla1856 · Answer

@Iain K Mackay wrote:

Windows and Windows 11 in particular likes to tell you your issue is hardware related.

Windows is without doubt the last place I would look.

I don't know ... maybe I have a fundamental misunderstanding of how hardware, software and Operating Systems work together.

Let me ask you ... when YOU observe a possible hardware problem with a computer, where do you YOU start looking for problems (ie Trouble-Shooting) ?

Tesla1856 · Answer

@Iain K Mackay wrote:

1. Windows and Windows 11 in particular likes to tell you your issue is hardware related.

2. I had black screens on a 12900KF and a 3090GTX, after the last big Windows 11update, solved by rolling back..

3. First thing I would be doing is a clean install of Windows.

4. Maybe your card is fried, but I would not be relying on Windows to tell you it is. Windows is without doubt the last place I would look.

5. The settings menu is now a morph of at least 4 Windows versions, with added keyboard presses,

6. just in case you didn't get the memo a new version of Windows is out.

1. My experience is that Windows only says that when some hardware is not quite right (which is rare). I only support about 30 PCs now-days.

All those connections in the PCIe slot ... it's just little pieces of metal pushed-up against other little pieces of metal (contacts). Could have been a bit of oxidation, heat expansion, dust ... any number of possibilities. It only takes one poor connection to cause a problem.

2. Yeah, it was like (that old problem) when the Nvidia driver crashes (and recovers) ... seen that before (on a different PC once). But no, this was different (see above).

3. Yes, I clean-installed Windows-11 Pro (64 bit) back in April when I upgraded this machine from Windows-10 . Worked great and has been rock-solid since then (except for those weeks when it wasn't). I think that is why it was so easy to tell it was not 100%.

4. No, it's not fried. It has been working perfectly since I did above. Fallout-76 sessions a couple of days a week and all the rest. Not sure why you are so down on Windows ... maybe you should try Linux.

5. Yeah, well ... they want it to work on (touch) tablets. I use (legacy) Control-Panel and Settings as required.

6. Right. I heard Start-Menu Folders are back. I think it's being offered on my laptop, but haven't had a chance to mess with it.

Vanadiel · Answer

I start looking at the sensors from HWinfo64. Specifically look for temperature and voltages. That's how I noticed one time the 12 volt rail was dipping out of spec when under load. Changing the PSU solved the random reboots I was experiencing.

It's another reason why it's important not to OEM too much on a motherboard, because otherwise you cannot read out the voltages properly. (like that was the case on my R10).

Tesla1856 · Answer

@Vanadiel wrote:

I start looking at the sensors from HWinfo64. Specifically look for temperature and voltages. That's how I noticed one time the 12 volt rail was dipping out of spec when under load. Changing the PSU solved the random reboots I was experiencing.

Good troubleshooting there.

Yes, after the first couple of "screen goes black events" ... I did start running CPUID's HW-Monitor (a similar utility program) while I was at idle desktop and running games. Everything worked and looked fine. Even though I was using the computer normally/heavily ... it still only cropped-up every couple of days.

When I discovered the errors in Reliability History report, coupled with the screen going black ... it was pretty obvious to me that the video-card could possibly be involved somehow.

I was just hoping this troubleshooting could be helpful to someone one-day. We often tell people to "re-seat your video card". Sometimes it helps and sometimes not (it's actually a bad video-card, PCIe slot, or even motherboard). However, I can't say that I've ever seen it work on one of my personal systems until now. It definitely fixed it. The Aurora-R6 lives on!

Tesla1856 · Answer

Just a final update. Everything is fine with my Aurora-R6 since I re-seated the Dell-OEM Nvidia GTX-1070 video card- My monitor does not go black for a few seconds any more (every few days)- No more red/critical Hardware Errors recorded in the Reliability Monitor any more I use the computer every day. For the usual stuff and games also (like Fallout-76 on Steam). There was also a 'Patch Tuesday' where a bunch of updates installed. This computer is still running Windows-11 Pro (64 bit) 21H2. And in case you are wondering, that one red-mark on 9-29-2022 is unrelated. It was just the QuickBooks-2018 Updater not exiting properly after it did an update-check. However, looks like my rating is back at 10 now.   Aurora-R6 with Windows-11 Pro

Alienware Desktops

Aurora R6, hardware error, Nvidia GTX 1070

Was this post helpful?