[SOLVED] 3 monitors randomly losing signal simultaneously

[muchdsign]

SOLUTION : The GPU seemed firmly seated, but apparently it wasn't. Simply had to reseat it and it fixed it, for now

Hey !

I've been using my new RTX3080 (This one) since november 2020 and it worked perfectly until very recently.

I've got 3 monitors plugged into this gpu :

1. 1920x1080 60hz
2. 2560x1440 144hz
3. 2560x1440 60hz

For some reason that I can't seem to figure out on my own, they all lose signal together at random times. Sometimes while I play games, sometimes when I work or even when I'm browsing stuff online. Can't pinpoint a specific trigger...
I've been assuming that it might have to do with the higher temps that have come with the summer, increasing my pc heat a bit, but nothing never goes above the expected temps.

The audio usually keeps running while the signal is lost, so I guess that the system isn't freezing. The only way I've found to resolve this is to power shutdown the computer, holding down the power button.
It happenend a few times now, 3 or 4 times across a period of ~50 days I'd say.

Here are my specs :

• Motherboard : Aorus z390 Ultra-CF
• Intel i9 9900K (stock clock)
• AORUS GeForce RTX™ 3080 MASTER 10G (stock clock)
• 32Gb RAM 2x Corsair Vengeance 16Gb DDR4 3200Mhz
• CPU Cooler : NZXT Kraken X72
• PSU Corsair RM750x 80PLUS Gold

Alright. That's about it, I can provide stuff if needed, I'll gladly take any help :')

Edit : I previously owned a 2080ti which I got refunded because it ran at abnormaly high temperatures, with fans spinning fast all the time. That makes me wonder if the issue might be coming from my motherboard... Don't know.

Thanks !

 

[Ralston18]

Have you tried reinstalling the GPU drivers?

Manually download via the manufacturer's website. Reinstall and reconfigure. (No third party driver installation software.)

Look in Reliability History and Event Viewer for error codes, warnings, and even informational events that match with signal losses.

Reliability History has a timeline format so look back ~ 50 days to see if something changed or began then.

How old is that Corsair PSU? Heavy use for gaming, video editing, even bit-mining? Was the PSU new, refurbished, previously used?

Check that the GPU is fully and firmly seated. Likewise for all other cards, cables, RAM, and jumpers.

Clean out dust and debris to help keep the system cool.

 

[muchdsign]

Have you tried reinstalling the GPU drivers?

Manually download via the manufacturer's website. Reinstall and reconfigure. (No third party driver installation software.)

Look in Reliability History and Event Viewer for error codes, warnings, and even informational events that match with signal losses.

Reliability History has a timeline format so look back ~ 50 days to see if something changed or began then.

How old is that Corsair PSU? Heavy use for gaming, video editing, even bit-mining? Was the PSU new, refurbished, previously used?

Check that the GPU is fully and firmly seated. Likewise for all other cards, cables, RAM, and jumpers.

Clean out dust and debris to help keep the system cool.

I'll try all that, thanks a lot.

Regarding the PSU, it's fairly old, around 2-3 years old. The PC is often under load, since I do quite a bit of gaming and I work as a 3D artist, the pc is often rendering large images for long sessions.
Everything should be firmly seated, but I'll check again. And I do take care of the dust filters at least once every month, and I usually clean everything inside out once a year.

 

[Ralston18]

My thought is that that PSU is likely nearing its' designed in EOL (End of Life) and not always able to meet system power demands - especially if there is some "perfect storm" of power requirements with respect to the GPU and its' trying to support 3 monitors.

And the PSU trying to support other components as well....

If you have a multi-meter and known how to use use it you can do some limited testing on the PSU. (Get a friend to help if needed.)

Reference:

https://www.lifewire.com/how-to-manually-test-a-power-supply-with-a-multimeter-2626158

Not a full test as the PSU is not under load. However any voltages out of tolerance or getting close to being so would make the PSU suspect.

As a side note: You mentioned 3D images and rendering.

With possible PSU problems be sure to make regular backups of your work/data. At least 2 x copies located off of the current host system and the copies proven to be recoverable and readable.

If the PSU is indeed showing its age it could go at any time. Likely file corruption and loss of data will result.

 

[muchdsign]

With possible PSU problems be sure to make regular backups of your work/data. At least 2 x copies located off of the current host system and the copies proven to be recoverable and readable.

If the PSU is indeed showing its age it could go at any time. Likely file corruption and loss of data will result.

Yup, thanks for the heads up, backups are already in place

I was aware that the PSU would be barely able to support these new GPUs, but it did for a good while. Is it possible that running it at its maximum potential for this time has damaged it or anything of the sort ?

 

[muchdsign]

Also, I'm currently reading multiple threads on other forums regarding similar issues, one of which in particular focused on display ports and how they can cause such problems.
Is that a possibility ? Two of my monitors are plugged in via DP and the last one is hdmi.

Does this thing have anything to do with my current situation ?

 

[Ralston18]

Re: Your post (#5)

Yes. Most products are designed to work at some expected level with occasional "overloads" or other excessive (likely subjective) use.

Running at full max will, sooner or later, lead to an early doom.

Many cars can go 100 mph (or better). Still if you were able to run that same car at 100 mph continuously for some amount of time then that car would likely catastrophically fail in some manner: engine, transmission.....

Even race cars designed for running at high speeds for long periods of time can breakdown.

Even the tiniest component that overheats, overloads, or goes out of spec in any manner can bring it all down.

= = = =

Your post (#6) - literally almost overtaken by events....

No harm I think in determining if the firmware update is needed.

If the firmware update solves the immediate problem all the better.

However, I would not leave PSU problems to chance.

 

[muchdsign]

I will definitely upgrade my PSU as soon as possible, can't right now.. money's running low :')

I'll leave this thread unsolved for now, but I did reseat the GPU after reading your first post, no more signal loss for now. Maybe that did something I'll make sure to close the thread after a day or two.

 

[rcald2000]

Hi muchdsign. Is it possible that one of the monitor 's DisplayPort cables is the culprit? I simply wonder.

A PSU being 2-3 years old is actually young. Your RM750x actually has a 10 year warranty. If you really suspect the power supply then you can reach out to Corsair and do an "advanced RMA", which would allow you to get a replacement before needing to send in the old unit. Personally, I'm not absolutely convinced that this is a power supply issue.