Installed a new video card, now my computer constantly restarts on its own

Title. I’ve been working on upgrading my computer with assistance from this thread I made on another forum, if you want to see the story there. However now I’ve run into another issue I can’t figure out how to solve.

Whenever I load up any game, I can play it for a few minutes before the computer decides to randomly restart. It might be an overheating issue, or a power issue, from that I’ve been told. But I honestly don’t know where to begin fixing that. Can anyone help?

Start with the obvious: put the old graphics card back in. It’s entirely possible that the new one is faulty straight out of the box (though not at all common it can happen). This should rule out the possibility of something else having gone wrong during the transition.

A power issue also a possibility. Your old card was an entry level card with no auxiliary power to speak of. The 1660 Super you have now is going to have at least an 8-pin PEG connector, which allows for it to use up to 3x as much power. Without knowing what power supply you’re using it’s just speculation, but from what I can see it comes factory fitted with a 460W unit - which uses multiple 12v rails. If you’ve got anything sharing with that PEG connector’s 12v rail, it could be getting sagging as that’s going to be very close to saturated.

1 Like

Sadly this is impossible. After the first day of it running awesomely I genuinely did not think I’d need it anymore. I stupidly threw it out.

Beyond that though is there anything I can do to salvage this situation?

Well, chalk the disposal of the old graphics card up to experience. Probably a good idea to hang onto old parts for at least 2 weeks, just in case.

As for options? It’s going to come down a bit to throwing things at the wall to see what sticks, if you take my meaning.

Start by re-seating the new graphics card and PEG power connector. It could be that it’s just not making a good connection, so refit it and also double-check that everything is fully - and evenly - plugged in. Also, if the power supply has it, try using a different PEG power connector - make sure you’re using an 8-pin cable for an 8-pin socket (even if it’s one of those ones that just loops two of the wires back from elsewhere).

NB There’s likely a “P4” connector on the motherboard with 8 pins. These have a different pin configuration to the PEG-8 connector and aren’t interchangeable. You shouldn’t physically be able to connect the wrong one without using excessive force anyway, but it’s worth knowing.

Next would be a software sweep. You’ve changed GPU architectures, even if not manufacturers, so it’s possible the drivers have some things working a little wonky. You can run the driver installer, select Custom (Advanced) and check the Perform a clean installation option. Or, and this is generally preferable, use a tool by Wagnard called the Display Driver Uninstaller (DDU) from Safe Mode to completely remove the old drivers before re-installing them.

After that? Well, the easiest thing to do would be to watch temperatures. TechPowerUp’s GPU-z can monitor them, as well as log to a file if you can’t watch it yourself while waiting for the problem to occur. Unless it shoots up particularly high (90°c+) than you might have to get it to reboot a few times in case it happens at a particular temperature.

There are other steps which could be taken, but I’m not 100% familiar with the software necessary to do it, or if it even exists anymore. It would require some kind of test suite which checks over the VRAM and portions of the GPU, but I haven’t seen a purpose-made program to do that in a very, very long time (mostly it’s just stress tests and benchmarks these days).

2 Likes

Alright, thank you. I’m going to try my best to follow your instructions. Im going to admit that this all sounds daunting though and I went into this endeavor as a total idiot. I just wanted to get better performance for shadowlands but I knew next to nothing about this sort of thing.

I was kinda hoping the solution would be something as simple as buying a little side fan.

I just want this nightmare to be over…

Hey Shardling,

The advice from Asterchades is solid. Usually system reboots like this indicate a hardware or power issue. Since it started happening after installing the new GPU it could be an issue with the GPU, or an issue with the power draw from the power supply or motherboard. It could be that your old power supply is not able to keep up with the new GPU for example.

Blizzard is not able to diagnose hardware issues since that is outside of our scope of support (we can’t exactly put hands on the system :slightly_smiling_face:), but hopefully with the info you already have you can get things narrowed down and cleared up. Best of luck!

3 Likes

It is solid, just kind of beyond the scope of my understanding. Is there a way to make it so that my power supply can be made to keep up? Do I just buy a new power supply? Can I buy a new power supply?

I know my request isn’t specifically a game thing (though my primary motivation is trying to upgrade precisely because of Shadowlands), just figured I’d try the only way I know how, by these forums.

Thanks anyhow.

We all start somewhere. Only difference is that I began my learning with regards to computers back in the '80s, before I was even a teenager.

Absolutely this is possible. I’m just loathed to recommend a first course of action which has a not-insubstantial cost attached to it without being reasonably certain that’s the problem. I’d rather exhaust the free options first.

From what I can tell, that particular Dell uses a standard ATX power supply, so a replacement will be easily sourced. Myself, I would opt for a SeaSonic-branded unit, somewhere around the 550W to 600W range. This is a tad excessive, but I prefer to run my power supply at around 50%-80% for efficiency and in case I change something later. It costs a little more than some other brands though even their cheaper units are extremely good.

There are other brands that are quite good as well, but that tends to be particular series more than the entire line-up. I’ve had good experiences with Antec in the past, though generally anything “Gold” rated or better will likely be good enough.

1 Like

No worries. I’m open to all the free options too, but I’m willing to do what it takes. My biggest concern at the time before I ordered the video card was that it wouldn’t even take. I know nothing about video cards and figured there was a process or something. And when I put in the video card, I almost thought I was right; the thing was a monster compared to the 645 and I was scared it wouldn’t even fit in the tower. Miraculously though I managed to get it in. I had to have people in that thread point me to precisely what I needed to buy, like the video card and that adapter plug that would be needed.

So if it is just a heating issue, it wouldn’t be helped if I just bought, say, a tiny desk fan to place right in front of the tower?

Can your power supply handle the load of the new GPU? You didn’t post any hardware specs so just throwing that out there.

Also check to see if the fans on the GPU work. A lot of modern GPUs will have fan profiles where the fans won’t even spin up until the GPU reaches a certain temperature. In other words the fans won’t spin up if you’re in the OS doing normal stuff, but as soon as you start gaming and GPU is under load, the fans should start. It’s possible the fans aren’t spinning up as expected causing overheating, or because of some other technical issue. You can download a free application called MSI afterburner that will let you set a custom fan curve (IE you can set how fast the fan spins at certain temperatures, then have the fan spin faster and faster as the GPU gets hotter and hotter, etc. You can use this to test if the fans work at all).

Placing a box fan in front of your computer won’t do much to remediate hardware overheating issues. These components operate at temperatures well over 100 degrees, and require component (heatsink, etc) cooling to be effective. Besides, that’s addressing a symptom, not the actual issue at hand. This is not a good practice with computer hardware issues. You wouldn’t put a band-aid on a gushing head wound that requires stitches, would you?

Finally, make sure your power cable is snug in the back of your case. I had an issue where my PC was shutting down randomly and unexpectedly. Turned out my power cable was loose and my foot was brushing against it, causing it to turn off. Double check the simple stuff before you dive into the more technical stuff.

Isn’t that what the DxDiag gives? My apologies.

I recommend a 750w PSU. I have one that’s been running solid for 6 years.

1 Like

Something like this?

https://www.amazon.com/dp/B07SL9NFVS/ref=dp_cerb_1

I want to be sure that the problem is a lack of power first though. Can I also be sure that my computer would even take something like this?

ATX power supplies tend to be limited only by the physical size, usually because they’re too long. If the pictures I’m seeing of your machine (it is a Dell XPS 8700, as per the DXDiag, isn’t it?) are accurate than you’d have no worries at all fitting even a unit with an output 4 digits long.

This is a point of concern. I probably should have looked over that thread earlier (I see you already ran DDU) but power adaption cables are always going to be a “use with caution” deal. There’s a possibility that adaptor is linking two of your three 12v rails together, which has a very real chance of overloading one of them. This could indeed cause a shutdown/restart situation.

I’ve seen exactly that kind of overload when trying to use two 7905 regulators to get more amperage on a -5v line. One of them winds up taking all of the load with the other sitting idle, and the one that gets loaded up overheats and can’t maintain output.

I can’t say it with absolute certainty, but it’s sounding more like switching the PSU would solve the problem - not at least because of the ability to retire that adaptor cable. The unit that you linked should be more than adequate, and since it [the 750w] seems to be the same price as the 650w there’s no real reason not to go up a step.

1 Like

Ah okay. I honestly had no idea why I bought the cable, just that I was told to. I don’t know what it’s really there for. I could attempt to give it a go after removing it. Would that help?

I’ll do what it takes to get this setup back on track. I just need a bit of guidance. Tell me what to do and I’ll do it. xD

My understanding, based on what I saw on that other thread, is that you bought the cable because the PSU you have doesn’t have the right connector. And it’s not, in and of itself, terrible advice to point you in that direction - there’s every chance the system may not boot if you try to use a 6-pin connector (which would and should fit) instead.

That having been said, there shouldn’t be any harm in trying - it just might bring up a warning, either before reaching Windows or after, saying that something is wrong (the same warning it would give if you left it unplugged entirely). The 6-pin plug will fit into the 8-pin socket when the correct two pins are left open. If you look from the top of the socket on the graphics card, with the clip pointed down, the open pins should be on the left. If it doesn’t fit as easily as the adaptor, don’t force it - try leaving the opposite end open!

ed: In the event it does work, and doesn’t give an error, test it with caution. Since you’re now almost certainly using just one rail for that connector, there’s still every chance of an overload occurring just from the graphics load. Strictly speaking it should do the same (reboot/restart) or trigger an internal breaker, but keep an eye - and a nose (you’ll smell it) - out for problems.

Adaptors like that can’t account for every situation, unfortunately. There was a period a few years ago where multi-rail power supplies were all the rage. Before that were single-rail, monolithic designs which would work such an adaptor without even the possibility of a problem. That’s just one of the reasons they’ve gone back to monolithic designs.

ed2: There’s also a chance it might work if you just disconnect one of the 6-pin lines from the adaptor and just leave the rest plugged in. Since it’s all black cabling and sheathed it’s impossible to tell from pictures how it’s wired up, but if it just splits some of the 12v lines while linking everything else together than it’s no different to most of the hybrid 6-/8-pin connectors. Most PSUs are going to be at least 18awg, which should be able to handle the extra draw over such a short distance without a problem.

1 Like

Awesome, thank you. I’m trying to understand though.

Is a 6-pin plug a…
-Something that was already inside my computer that I have to utilize?
-Something that the adapter cable I bought is capable of and I just have to attach things a different way? Or…
-Something totally different I don’t have that I’ll need to buy?

Ah! My bad.

The 6-pin plug is the one you have on your power supply already. You should have two of them plugged into that adaptor to make an 8-pin which then fits into the graphics card.

Also, check ed2 on my previous post - might be easier to test, though you may need to try both inputs on the adaptor.

1 Like

Ah! Gotcha, thanks so much. xD

Sounds good in that case. I have a couple things to try out when I wake up later. I’ll keep this thread posted.

1 Like

No luck thus far. This is Shardling btw.

I learned of something called “Event Viewer” from a guildy last night. After it restarted yet again I took a look and saw this.

“The local adapter does not support an important Low Energy controller state to support peripheral mode. The minimum required supported state mask is 0x2491f7fffff, got 0x1fffffff. Low Energy peripheral role functionality will not be available.”

Would this potentially tie into what I’m experiencing?