Pc shutting down
Posted by Diligent_Cut_3808@reddit | buildapc | View on Reddit | 7 comments
So my pc keeps randomly shutting down and turning back on for some reason and I'm just looking for some suggestions on what it might be.
postsshortcomments@reddit
If you recently replaced your GPU or you've always had this problem, what is your PSU wattage and GPU model? It can be due to a PSU model that is too small. Sorry, but you'll need to check your receipt or physical PSU.
You may want to check your temps. I still use HWMonitor because I think it's easier for beginners, but it's a bit out of date and sometimes has issues with modern equipment censors. HWInfo64 seems to be more preferred these days, but I've personally never had an issue. Check for any equipment running over 90-95C (some gpus like 9070 XTs can run hotter) and report back if you see any. Shutdowns usually do not occur unless it's an extreme cooling issue, but if you have few fans in your case, broken fans, a stock cooler, etc., it's certainly possible. Usually issues like these compound. For instance, let's say you never crash while just browsing, but when you play a game you crash after 20 minutes. Then when you boot it back up, you crash again after just a few minutes because the toaster is already hot.
RAM overclocks can (XMP profiles) cause occasional crashes. Sometimes they are very occasional, in other cases it may almost instantly or several times in several hours. My recommendation is disabling it completely and seeing if it's stable, then enabling lower overclocks.
If you are running 4-sticks of RAM, especially on AM4 or with AMD CPUs, this sometimes causes instability. If this started occurring after a 4-stick upgrade, it's very likely this. Again, if you have a RAM overclocked enabled, sometimes a lower overclock can make it stable on AM4.
Try a BIOs update. This can sometimes solve it. Be careful not to power-off during a BIOs update, even if you think it's hanging. Interrupting a BIOs update is much like interrupting other electronics during crucial updates and sometimes they are not easy to fix without specialized equipment (especially newer motherboards often have the ability to do CPU-less BIOs updates or BIOs updates from just flash drives.)
Is this happening in specific games or anywhere including while just browsing? If it's just happening in specific games, it can be something on the developers side.
Updating GPU drivers (especially if above) and chipset drivers can solve a lot of crash issues. Chipset drivers are available on your motherboards support page and include chipset drivers, network, and sound driver (often RealTek). Sometimes they also include bloatware or non-essential drivers like RAID, so you may want to check during the install.
If you have an Intel 13th/14th gen CPU and you are running it or ran it on a motherboard with a pre-microcode hot fix BIOs, you may have a CPU with CPU degrade.
A Windows reinstall is preferred, but you can try doing a repair using Rufus first. Unless you have an extra SSD sitting around, I would probably try doing the Windows reinstall after we've diagnosed other hardware. If you have an extra, old SSD sitting around you can remove your current storage to "save the Windows install." If the old SSD has windows on it, you can usually go under utilities in your BIOs to format the drive. If your windows install is really important, you can find some $15-20 128GB-256GB SSDs on ebay pretty easily.
If you've gotten this far and still haven't figured it out, let's start hardware testing.
CMOS reset by removing the battery, if you feel comfortable reconfiguring any BIOs settings. Basically, this will revert most BIOs settings back to factory defaults (including overclocks). If you're not fully comfortable doing this, the main three issues I am trying to clear here are: CPU overclock settings (possibly from a prior owner) and RAM overclocks (possibly from a prior owner). You can often save BIOs settings to a flash drive if you are worried about not knowing how to get them back.
By default RAM should go in first A2, then B2 slot (unless otherwise indicated by the motherboard). If you have two sticks of RAM. Remove one RAM stick and test in A2. Test. Try the other RAM stick in A2. Test. Next switch slots (A1) with one stick. Test. Then the other stick. Usually RAM issues are single
Fetch your CPU model and check if it has something called an "iGPU" or provide me with the model. If it's an AM4 CPU without a g at the end (like the 3200g or 5700g) it probably doesn't. Most AM5 CPUs have them, besides the 7500f. Most intel CPUs have them, unless they end with an F.
If you have a CPU with an iGPU, we can plug your [usually HDMI] cable into the motherboard I/O and not the GPU and then remove the GPU (with ASUS boards especially, the tab to disconnect the GPU from the PCIe slot sometimes need to be pulled parallel to the motherboard, not pushed down). Unplug GPU power, remove GPU, plug monitor into iGPU. If you just see a black screen, you probably don't have an iGPU or it's disabled in the BIOs (a CMOS reset typically defaults this to enabled).
If you do not have a CPU with an iGPU and you have an old GPU, you can usually test with that. If it's a GPU model older than 2012, you may want to provide a model or research if it has any issues on modern equipment.
From here we're at a bit of troubleshooting hell if it hasn't been resolved. Your best lead is to look into your Event Logs and see if you can get a bit more data on it. A Youtube video should explain how, but they are a bit hard to decipher as there can be a lot of events.
We're left with PSU, motherboard, and CPU [and GPU if you couldn't test it with an iGPU or an old GPU model]. It's most often the PSU or motherboard.
On PSU models: If you plan on doing a $1300+ upgrade, a used GPU upgrade, or a $500+ GPU upgrade any time in the next 5 years and want a cheap PSU.. you can snag a speculative B-tier SAMA GT750BK for $67 or SAMA GT850BK for $72 to open your GPU options. Better A-tier options like a Montech Century to are usually $80-90. Alternatively, a $50-55 C-tier or better 600W/650W.. but that somewhat restricts used GPU options. Make sure you refer to a community PSU tier list and don't just throw in a random model, as failures of certain cheap models can take out equipment with it.
The following information would be useful: CPU model to check for any difficulties you may have with motherboard replacement, PSU wattage, or GPU model.
Diligent_Cut_3808@reddit (OP)
Everything is brand new, the GPU, CPU, and mobo are 100% working, it's not the PSU because it is a good make, the bios are up to date, I'm not planning on upgrading for at least the next few years, and nothing is over clocked/ undervolted
postsshortcomments@reddit
Also figuring out if it's dumping an Event Viewer log (eventvwr) may be a fast track. I'd call it a moderate to advance research/learning curve and it's a massive information overload so it's not my favorite to suggest and get someone 'stuck on' early. Usually those who know it have already done it and the learning curve can be longer than an early-list brute force fix. But those logs can sometimes work wonders.
postsshortcomments@reddit
That's really all I have and covers every issue that I can personally think of. If it's only occurring in specific titles, it could just be a known GPU issue that isn't patched yet.
Gotta start from the top and work your way and exclude whatever isn't (even if "yes i just did this, it should work," in case it doesn't because whatever is causing it is probably there. Even on new builds, parts can arrive broken on arrival (including PSUs and especially malfunctioning motherboards/GPUs). We see that constantly. Sometimes it's a motherboard damage from even someone else bumped on a standoff or dropped and resulted in a bad controller/solder (motherboards often aren't sealed and if it doesn't appear open box, it may be being returned as new when it shouldn't be).
If you have old parts that are known working, they may be able to assist you.
A BIOs rollback (you may want to research your current BIOs version and see if others are having crashes with your BIOs version) is by no means unheard of.
Make sure you don't have XMP/RAM profiles enabled above default. If it's AM5, I believe that some motherboards RAM training can sometimes automatically overclock to an unstable voltage. This is the most likely culprit if it's not damaged hardware and is a fresh windows install. Memory training is the only thing I'd change with AM5 - it's nice to have RAM oc's automatically tested and is easier/sometimes user friendly, but there can be a lot of issues with it that may only pop up after 3 hours.
Also make sure you installed chipset drivers from your motherboard's support page. Sometimes Windows Update will fetch the drivers and I don't know if it's a huge issue anymore, but I've had motherboards in the past that threw in either generics or out of date ones.
postsshortcomments@reddit
Also could be a home wiring issue (especially if you have an old home or have a history where other devices have intermittent electrical issues). You may want to try the build in a different room.
Diligent_Cut_3808@reddit (OP)
The wiring was redone not too long ago
postsshortcomments@reddit
May be something with that. I'd try it in another room, for sure, that you don't have any other electronics in. Especially if you have a ton of other electronics like a Window AC/hair dryer on the same breaker. I know someone with a fairly new house who has always had troubles with a single wall having issues.