New build starts and runs briefly, but crashes randomly

[easythezengod]

Hey guys! So im in the process of completely overhauling my old PC- it starts up and runs well but crashes/freezes/BSODs at seemingly random times. I'll occasionally get the stopcodes "WHEA Uncorrectable Error" and "Critical Process Died", but sometimes the machine just stops and restarts with no indication of what went wrong. So far I've updated the bios, installed the CPU cooler (in case this was possibly a heat related issue), reseated the whole system, updated all drivers, and installed a new SSD- none of which have yielded any results.

Additionally, I'll occasionally get stuck in a bios loop upon restart, and recently noticed that when this happens the machine won't recognize either the disc drive or SDD to boot from. This is typically fixed with a good ol' fashioned restart, but I thought it might be important info for you folks who are much more knowledgable than me.

One last, possibly very important note- im building this around an old 730w Kingwin power supply, as it never gave my old build any problems and met my wattage requirement, but if this seems like it could be a PSU issue then that might be the place to start. Unfortunately, since it's my old unit, I don't have another one on hand to do a 1-1 comparison to.

Thanks in advance everyone, hope this turns out to be an easy fix


Edit- adding specs below!

OS- Windows 10 Home (64 bit)
MOBO- Gigabyte B250M-DS3H
Processor- Intel i7-7700K @4.20GHz
Graphics Card- Geforce GTX 550Ti
SSD- PNY 240GB Solid State
RAM- 2 8GB sticks of Ripjaw DDR4
PSU- Kingwin ABT 730MM

 

[animemangamer]

I'm no expert OP but I do recall encountering a similar issue.

I suggest you start with the standard paper clip test to see if your PSU is functioning properly:

[video="https://www.youtube.com/watch?v=5FWXgQSokF4"][/video]

If the PSU is okay then it must be a motherboard thing (you might've gotten a bad lemon - return/replace it)

Also, that Gigabyte B250M-DS3H seems really inferior/low end for a high-end CPU. May I suggest you to get a Z270 chipset motherboard instead as I think you'll appreciated the better performance/features offered.

 

[BobCharlie]

What video card is in this new build? I used to run a cheaper 650w with a GTX 550ti and never had issues, even when I upgraded the board and chip to an 8350. However, when I stepped up to a GTX 970, I noticed erratic benchmark testing, which I narrowed down the issue to the power supply not actually being able to provide a reliable continuous current i.e. it's rated "cont. amperage" was too LOW with no head room.

While shopping for a more modern PSU, virtually every manufacturer seemingly hides this value, while relying solely on the "wattage" number or worse, "MAX" amp numbers if you look at their output spec sheets. Even PSU within the same "wattage" ranges (where documentation existed) had wildly fluctuating "MAX" values (like 10 amp+ variation). "MAX" *should* be a limited burst. "Continuous" should be 24/7 365.

Because of this nonsense , I went with overkill, and got a gold rated 1000w EVGA super nova G2. Graphics Benchmarking was solid with it.

FWIW, the old 650w PSU listed a 12v continuous rating ON the actual unit. BUT, newer versions of the same PSU switched to the "MAX" ratings in online spec sheets, which were nearly double the continuous rating IIRC. Not sure what grey-area standard they were abusing, but it's seemingly rampant between companies or the units are extremely inefficient. If I had to guess, they aren't forced to list continuous ratings, similar to how car audio isn't forced to list actual wattage the unit realistically does, so they post inflated "MAX wattage" numbers to bump up sales, though home electronis is heavily regulated and must conform.

That "1000w" EVGA PSU for example, lists 83A as "MAX". Using the math for the junky PSU it replaced, it's cont. rating is somewhere between 50-65A cont. depending on how efficient it actually is. Those are very rough numbers, it might be a higher/lower cont. amp, though I'd need to SLI to find out it it falls flat or not. I don't have the hardware to check these things.

If you are still running that old PSU, consider updating it for reliability, and try future-proofing by going as large as your budget permits for whenever you update the GPU. It'll only use what it needs if going larger, i.e. it won't pull 50 amp every second of the day if it's only needing 30. If the problems go away, great. If not, double-check everything you did. Make sure you have the power correctly going to the GPU (think it took a 2-in 1 connector).

 

[easythezengod]

What video card is in this new build? I used to run a cheaper 650w with a GTX 550ti and never had issues, even when I upgraded the board and chip to an 8350. However, when I stepped up to a GTX 970, I noticed erratic benchmark testing, which I narrowed down the issue to the power supply not actually being able to provide a reliable continuous current i.e. it's rated "cont. amperage" was too LOW with no head room.

While shopping for a more modern PSU, virtually every manufacturer seemingly hides this value, while relying solely on the "wattage" number or worse, "MAX" amp numbers if you look at their output spec sheets. Even PSU within the same "wattage" ranges (where documentation existed) had wildly fluctuating "MAX" values (like 10 amp+ variation). "MAX" *should* be a limited burst. "Continuous" should be 24/7 365.

Because of this nonsense , I went with overkill, and got a gold rated 1000w EVGA super nova G2. Graphics Benchmarking was solid with it.

FWIW, the old 650w PSU listed a 12v continuous rating ON the actual unit. BUT, newer versions of the same PSU switched to the "MAX" ratings in online spec sheets, which were nearly double the continuous rating IIRC. Not sure what grey-area standard they were abusing, but it's seemingly rampant between companies or the units are extremely inefficient. If I had to guess, they aren't forced to list continuous ratings, similar to how car audio isn't forced to list actual wattage the unit realistically does, so they post inflated "MAX wattage" numbers to bump up sales, though home electronis is heavily regulated and must conform.

That "1000w" EVGA PSU for example, lists 83A as "MAX". Using the math for the junky PSU it replaced, it's cont. rating is somewhere between 50-65A cont. depending on how efficient it actually is. Those are very rough numbers, it might be a higher/lower cont. amp, though I'd need to SLI to find out it it falls flat or not. I don't have the hardware to check these things.

If you are still running that old PSU, consider updating it for reliability, and try future-proofing by going as large as your budget permits for whenever you update the GPU. It'll only use what it needs if going larger, i.e. it won't pull 50 amp every second of the day if it's only needing 30. If the problems go away, great. If not, double-check everything you did. Make sure you have the power correctly going to the GPU (think it took a 2-in 1 connector).

Thanks for your replies guys! With another day of troubleshooting I came to find that after removing the GPU and sending to the monitor through the motherboard's VGA slot, it resolved the crashing issue. Strangely, however, this came with additional problem of my screen going blank for about 1-2 seconds every 5-10 seconds, and has me wondering if the card itself may be fine and its actually the inputs/PCI slot on my motherboard that are bunk? (The same screen flash happens when I run through HDMI) I tried flashing the BIOS and updating the drivers but that didn't seem to fix it. I'm about to grab a graphics card from Best Buy with the intention of returning it this afternoon if it doesn't yield any results.

Also, I bought a new PSU in hopes of resolving the issue. While it didn't fix it, its nice not to have that old dinosaur PSU waiting to erupt in my computer lolol