*RESOLVED* Help from The Smart Guys®

*RESOLVED* -- tldr: it was the PSU.


Hey DS Family... I have an issue that maybe one of you can help resolve.

I have a 7 year old DS machine that still runs fabulously for what I need. I have replaced/upgraded and made additions to as I have needed with never an problem because of the folks here on these forums. All of you that know so much more than I do and have always been willing to lend an ear with advice.

Currently, my GPU is a EVGA GTX GeForce 1070, I have 16Gb Ram, 4 storage devices and 3 monitors. (Yes, the 1070 can support 3 monitors with no issues under ideal circumstances.) The monitors I have are my main monitor (Large) at 32" LG Ultragear 2560x1440@120Hz, secondary (Medium) 24" Viewsonic 1920x1080@60Hz and on top is the Small 20"LG 1600x900@60hz. I gave all that information for a reason.

Over the last couple of weeks, periodically and usually in the middle of a game (either RDR2 or Path of Exile) all monitors will just shut off suddenly and without warning and I have to hard reset, or power off and on, to get them back.

First I thought it was thermal. So I deep cleaned everything made sure all the little dust bunnies were gone and there was no chance at overheating. Of all my temperature checking program the only oddity is that HWMonitor lists my mainboard @ 120°C. We know that's wrong because it's been like that since the beginning and it's that temp even on first start up in the morning. So, no worries there. All other temp programs list all temps nominal.

Still happens.

Checked Google.. was told to use Win+Shft+Ctrl+B to reset the graphics driver while running. No change.

Uploaded ALL current drivers system wide. No change.

Even tried a "WhoCrashed" report, but as there is no dump file there is no report.

Downloaded Intel's Intel Processor Diagnostic Tool® and had intel check the chip, CPU and GPU with no problems and still... no change.

Now, getting frustrated I circled back to the Thermal idea just to make certain. I ran the latest GPU Stress tests. Unigen's Heaven & Valley on the large monitor while concurrently running FurMark on the medium monitor with the small monitor displaying any of 4 different temp programs (Speccy, HWMonitor, RealTemp, and EVGA Precision X16 from NVidia).

I ran those two programs, at the same time, for half an hour. My GPU never broke 60°C with the fan curve I've set up. There was no shut down in anything I've done except when I'm in those two games. Now, if it were only RDR2 I could probably trace the problem to something new and current, but Path of Exile has been around for what, 5 or 6 years and the engine is older and the game demands much less from a machine.

So, now I'm completely flummoxed and can really think of nothing else I can do.

Any suggestions?

Thanks for reading all this.   

Edited by Snaike - 07 Dec 2021 at 12:39pm

You've already done more than I could have. Best I can offer is perhaps you don't run all three mons during those games? That would certainly change the power demand, and it seems to be a transient power overdraw, doesn't it?

Have you checked

https://outervision.com/power-supply-calculator ?

Follow up...

Power Supply Calculator confirms enough headroom to run the entire system on existing power plant.

1 or 2 monitors off while gaming - crashed.

Small monitor plugged into mobo during crash never got a signal.

That's strikes 1 through 3. I'm open to any other troubleshooting or ideas.

Thanks again.

Originally posted by hoserator Do you have an old gpu you can install and see if the blackout still happens?

Somewhere in this mess is the old PNY 770 that came with the machine. I'll have to dig around and see if I can find it. That may be the next course of action.

I do have other feelers out in other forums to see if I can find more information. I'm thinking it may be time for a new DS machine but COVID and the attendant economy slowdown have pretty much depleted any and all resources that I could have drawn on... maybe I'll send a note to Alex to see if he's finished my Envy Green Aventum that we talked about a few years ago... that one he was making me gratis for all my work here in the forums.   

@MrCheetah - Thank you for chiming in as well. So far I have tested the problem with 1 and 2 monitors completely disconnected and/or plugged into the mobo video connection. Nothing there made a difference.

I have watch the first video with Jayz and have followed his instructions for a DDU enhanced complete driver wipe and install. I have hopes that this may be the answer as he was describing some of the issues that I have recently encountered. Ah well, time will tell if it was successful.

and for @hoserator as well.. I am not in a position to be able to add/remove parts while the machine is in operation. It would take way too long for me to explain and it would just be a waste of space... But I will keep that in mind. Removing the 1070 during a crash and forcing the VGA seems kind of drastic but, hey, at this point who knows what we're going to end up trying. lol

Thank you very much.

I'm out of my league here, but I want you to know I'm pulling for a quick solution for you. In the meantime, I hope it confines itself to just those two games. Best wishes!    

Originally posted by Snaike YouTube vids and general browsing

This makes me think you should:

How to Test Your Computer’s RAM for Problems | How To Geek

Update.... There is no joy in Mudville.®

After waiting a couple of days with zero crashes I went ahead and started PoE and within 10 minutes the system crashed.

Further digging on the EVGA sites directed me to a fantastic little monitoring/debugging/testing program called OCCT. I'm not sure if I've ever seen that one mentioned here but it was extremely thorough, even testing my GPU fan speed up to 2500rpm while the GPU stayed at a relatively cool 58°C.

But the stress tests all showed everything within nominal ranges with no issues. Ran it for a full hour pushing everything from the individual CPU cores to the over all power draw and nothing showed any signs of being defective.

Then the universe decided to kick me while I was down... as I ended the program and before I could even get to another program, the machine crashed.

I don't know what to do. I am out of ideas. I do, however, recommend to any one testing their machines to take a look at OCCT. It's quite the little workhorse.

More later if anything changes. No news means it is still borked.   

Interesting. I'm not a user of Steam; so, I don't have any additional insight.

Originally posted by fwfdfireman Are all the games causing you problems run through Steam? If so, you might have a corrupt installation of Steam. If it was just a game file validation thing, I don't think and have never heard of several games becoming corrupt at the same time. But..... might try validating game files on each game, IF they are through Steam. Just a thought.

Rocket League is through steam and verified and its working
Valhiem is through steam and verified and crashes
RDR2 is through steam and verified and crashes
Valhiem offline is working.

Yes, local file integrity is one of the first 'fixes' that I had tried with the steam games.

I am leaning toward a reinstall of Steam but I need to run some more tests. If I can play RDR2 and a few other problem children while "offline" and they continue to crash "online" then that will be the next project. (Can't play PoE 'offline'.)


Edited by Snaike - 23 Nov 2021 at 5:28pm

Reinstall of Steam is a good idea.
I know you have searched and might have already found this, but here is a link to a good trouble shooting plan for YOUR issue with your Steam games.
https://e-methodstechnologies.com/fix-steam-games-keep-crashing/
Their last solution is to reinstall Steam. lol
Hope this helps, nothing more frustrating than trying to run down a problem like that. Good Luck!!

Thank you gentlemen. I'll take a look.

For those following along... I did get a sentence of wisdom from the Nvidia GeForce Discord... one of the mods finally posted back and said that the latest drivers have been having issues with dual screens and I should roll back to the driver released in Sept.

So, that's what I've done, but it's 0500 here and I really need to get to bed before I can test it thoroughly. I was able to play an hour or so on Shadow of Mordor (Love that game!) with no issues, but we'll see later today.

Happy Thanksgiving to you and yours.

In further discussions with the great folks over in the NVidia tech support forums I was informed that my Win10 was over 2 years old (Oct '19) and that was the issue. So, before testing the driver rollback I went ahead and updated to the most current version of Win10 and the most current NVidia driver which I was assured would 'most probably' fix any problems that I was having.

It didn't. Within 5 minutes of Rocket League this morning BLANK, it crashed. (If you're not playing it you're missing out on their 'grid-iron limited time' games which are hilarious... football with cars!)

So, this morning I have rolled back the driver again, because as I said above I haven't had a chance to really test this part.

I am at the point now, I have realized, that should I encounter another crash it is most likely a hardware problem in either the GPU (75% chance) or PSU (24% chance) with the balance going to the dreaded "unknown".

I am losing hope as new components are not in my future due to circumstances beyond my control.

Cross your fingers and, if so inclined, say a little prayer.

Thanks for the help.

I have my fingers and toes crossed for luck.

Originally posted by Nomad1970 A question, do you still have your 770 GPU? Could you use it it see if the GPU is the problem?

I do not. The replacement was in Jan '17 (Documented for the forums on this thread). As much of a packrat and hoarder that I am, "she who must be obeyed" is not, therefore 7 year old GPUs that were replaced almost 5 years ago will not be saved.