Random Reboots

[TheeRadioDJ]

I just recently bought a new Gigabyte AMD mobo, with the B850 chipset, and I put an AMD Ryzen 7 7800X3D CPU on it. The Ryzen CPU is also brand new. I had been running an RTX 4070 Ti Super GPU on the system, and it was very stable, as far as I could tell. I decided to move the entire system into a different case that I already had used in the past. I was in a hurry to fire it up and test it in the new case, so I tried the 7800X3D's onboard graphics before dropping in a GPU and getting carried away with all the cable management. Things went downhill from there. When running the 7800X3D's onboard Radeon graphics, the system started rebooting or freezing several times in just a few minutes, even when I wasn't stressing the system. I didn't change anything else on the system, so I'm at a loss as to why it became unstable on the integrated graphics. I haven't otherwise had any issues with the Gigabyte mobo, and I updated the BIOS to the second to most recent version. I haven't updated to the very latest because I wasn't sure if it was a beta or not.

Power supply is an NZXT 1000w unit, and I'm running Windows 11 Pro 24H2. I really don't have a clue if this is a borked Windows installation, or the Ryzen 7 7800X3D just has flaky integrated graphics? Any thoughts?

I forgot to mention that once I dropped the RTX 4070 Ti Super back in, the system went back to being stable...as stable as WIndows 11 ever is.

[Grogan]

Hi GK!

It's possible the onboard graphics adapter has a fault, or it could be the drivers in Windows. It also could be shifting hardware resources when you took out the Nvidia card.

Either way, it's good that your system is stable again with the Nvidia card. I don't imagine you want to screw around removing it again and testing, but you could try booting a live image of some Linux distro to see if the problem is in Windows.

[TheeRadioDJ]

Hey Grogan,

I didn't even think of trying a live Linux distro. I should've done that. I guess I still can.

My luck with computer hardware is generally pretty shitty, not helped at all by my even more shitty 60 year-old eyesight. I somehow managed to bend a couple of pins in the USB header on the new Gigabyte mobo - the header that connects to the USB type A ports on the case. Against all odds, I was able to bend them back into close enough alignment that the plug went in. I figured the plug would make sure they were all back in alignment. I guess I was lucky, because the ports work fine.

Just when I thought I could chalk up a win, I got an ASRock RX 6600 Radeon card today that was DOA, new in box. With the PCIe cable connected to the card, the computer wouldn't even power on. Just to make sure I didn't suddenly have a dead PSU or bad PCIe cable, I tried the card in a completely different system. Same story, would not power on with the PCIe cable connected to the card.

I've had two ASRock mobos die suddenly on me, and now a DOA ASRock GPU. I should've known better. I won't ever make the mistake of buying ASRock again.

[Grogan]

Yeah, I'd avoid Assrock. I had an Assrock video card (RX 6700 XT) that failed in 31 days, just outside of the period where I could have returned it to Newegg. Assrock's return policies and behaviour are horrible, and I'm ashamed to admit that I just gave up early and didn't bother trying again for a response.

So I've got a $469 video card sitting in a box. Part of it is, I don't really care to have another one the same. It had horrible behaviour from the start, like being ultra sensitive to ESD. Touch the computer case (which was properly grounded, I tried two cases even) and the screen goes black and the computer was halted.

I understand what you mean when you say shitty luck with computer hardware. I don't like messing around either and would probably leave well enough alone.

I broke plastic on the data connector of a SATA drive once. Like you say, against all odds I was able to straighten the pins and STILL stick a SATA cable on it and get the pins in those little holes. I then squirted Crazy Glue around the connector (and between the data cable and the adjacent power connector lol) and fixed it in place, permanently. It never moved once, it lasted years as the boot drive (only drive) in a tower. To my knowledge it should still be fine (though I have that tower sitting in a corner disconnected and haven't booted it up in a year or so)

[TheeRadioDJ]

once I dropped the RTX 4070 Ti Super back in, the system went back to being stable...as stable as WIndows 11 ever is.

That didn't last long.
The random reboots and freezes I was getting with the Gigabyte mobo and Ryzen 7 7800X3D led me to believe that one of them was borked, as the system barfing started up again after I put the RTX 4070 GPU back in. Having neither another AM5 CPU to try in the Gigabyte mobo, nor another AM5 mobo to try with the CPU, I realized I had no way to test them separately. The Gigabyte mobo was past the return window, but the Ryzen 7 is still returnable. What I decided to do was buy another mobo to test with the CPU. Let's spend more money. Yay!

The new(er) mobo is an MSI, which I've always had good luck with. Using the same CPU, DDR5 RAM, SSD, etc, the MSI ran solid. I ran that Unigine_Valley-1.0 benchmark for over an hour, and had zero problems.
I realized there was ONE more possible culprit. On the Gigabyte mobo, I removed the stock CPU retention system right out of the box and replaced it with one of those third party CPU frames, because I'm gullible and think those are actually necessary. I put the stock retention system back in place, and tried the Gigabyte mobo one more time. Guess what? It ran fine, with no issues with the same benchmark software.

I'm still a little scared that the other shoe will drop anytime now, but it appears that I caused my own problem by using the CPU frame, instead of the factory standard. My guess is the CPU wasn't making full contact with the socket at all times.

The Assrock DOA doorstop GPU is on its way back to Amazon, and they've already refunded my money. I'm happy about that.

The new MSI mobo I ordered? The morons at Amazon couldn't be bothered to even put it in an Amazon box. They slapped a shipping label on the retail box, and decided that was good enough. Even the Amazon driver thought that was dumb. Thankfully, it seems fine, and there weren't any missing parts. Their packing has really sucked of late. If you do specify that you want Amazon packaging, they usually just throw the retail box in a larger Amazon box with NO packing stuffs to fill the empty space.

[Grogan]

At the time I bought that Assrock card, even at Newegg they didn't have the usual refund or replace options, only replacement. At the time, Amazon didn't carry their products at all (presumably because of that). They must have had to change that policy.

So the CPU wasn't mounted properly and during heating/cooling, expansion/contraction it would lose connectivity. That's a pretty good diagnosis, it would have anybody chasing their tails

[Zema Bus]

Glad you got it figured out. And now that you have another board you can start working on a 2nd AM5 system

I've had Amazon do that sometimes, ship stuff in just the retail box. I never had anything arrive damaged as a result, maybe they aren't as rough as UPS and FedEx.