[SOLVED] Question about blue screening restarts on games

[dauransx1]

I got a Cyberpowerpc with an i9 9900k, RX570, 16gb of RGB ram, 1tb m.2 with a 2 fan cpu water cooler, 650w gold thermaltake RGB psu and a Asrock phantom board and 7 RGB fans running at max RPM's assuming by how loud they are. Ordered it (5 or so months ago). My issue i'm having is that i upgraded it with a 1080ti Aorus, 16gb more ram of the same model and a 500gb ssd and when i run some games i get blue screens and the pc restarts in games like Subnautica below zero and planetside 2 but runs fine with Furmark and superpositio but bluescreens in 5 secs with Furmark CPU burn. Temps were fine with cpu (60s) and gpu (high 70s). Tech wants to blame my GPU with no data and the fact it runs 2 different GPU benchmarks fine and most games with no errors. Subnautica Below Zero dev says my 650w is below what i need for my setup but the Cyberpowerpc tech claims 650w is more then enough but before said is barely enough which is conflicting and confusing. On their site, you can build the same pc but with a 2080ti and the same stuff i added and also add the 30% CPU OC option and they still suggest 530w and the tech still redirects blame on my GPU or says "you will need to note that the PSU suggestion is just that,A suggestion" which i can see to a point but why not suggest a 750w or w/e and its misleading to suggest a 530w to people who knows 0 about pc's because they are there to buy a prebuilt pc because of that or in my case because i'm lazy and have basic knowledge.

My question really is, is 650w enough and or is the issue im having the cause of it being 650w? I did order a 850w of the same model but it wont be here till the 26th and ill test that and see if i bluescreen but im just confused because they tell me 650w is barely enough tho they suggest 530w and then also say 650w is more then enough and i see other questions on the net with answers saying you need least a 750w but 850w being the sweet spot for builds near mine. Unsure if a system refresh will fix this issue also because i did test 3 GPU OC programs, 2 being from the GPU brand i got and the other was afterburner testing the 125% power target but i did that one at a time but all 3 installed at once and tested each one one at a time and uninstalled afterburners and the other and left the main default program so maybe it did something?

Hopefully i made sense, not the best at typing and my words just wonder...

 

[Vic 40]

Test without the added ram. When you have removed the added ram do a clear cmos and try with the original sticks, see what happens.

 

[dauransx1]

Test without the added ram. When you have removed the added ram do a clear cmos and try with the original sticks, see what happens.

I just did that. Took out the 2 sticks i put in myself and reset the cmos and seems like it was fine with the two sticks they had in it in the same dims they had them in, played subnautica below zero and it was fine and then did furkmark cpu burn and it was fine and cpu-z stress. I then took the ram out and put mine in and in the same dims and seems like they work also. Put all 4 in now and seems like they are fine with same tests. Maybe its something that takes time and or is random?

 

[Vic 40]

Mixing ram even when the same make and model can create problems. Guess you were victim of that. Sometimes adding voltage can create stability. Either to the ram voltage itself or to the memory controller.

Did you also test all sticks with XMP enabled? Can test also with the next,
MemTest86 - Official Site of the x86 Memory Testing Tool
needs you to make a bootable usb. Let it run overnight. See if it comes up with anything.

Stil in the long run might there be stability problems.

Oh btw don't think that adding 200watts to the psu is going to make a huge difference. If you have a thoughpower should it be alright although not the absolute greatest choice. If you want better stability from the psu would a better quality one provide it. And 650watt should be fine.

 

[dauransx1]

Mixing ram even when the same make and model can create problems. Guess you were victim of that. Sometimes adding voltage can create stability. Either to the ram voltage itself or to the memory controller.

Did you also test all sticks with XMP enabled? Can test also with the next,
MemTest86 - Official Site of the x86 Memory Testing Tool
needs you to make a bootable usb. Let it run overnight. See if it comes up with anything.

Stil in the long run might there be stability problems.

Oh btw don't think that adding 200watts to the psu is going to make a huge difference. If you have a thoughpower should it be alright although not the absolute greatest choice. If you want better stability from the psu would a better quality one provide it. And 650watt should be fine.

PC seems to be running fine now because i think i had them slotted in part ways when i first installed them because i was having issues putting them in again and they felt like they were in ok but when i looked they were in at an angle barely but enough you could see the copper teeth more on one end but now the ram is seated just fine and been testing all day and not one blue screen. As for the memtest86, i do not have a usb drive right now, had one and lost it so that will need to wait and the ram that came in the pc was 2x 8gb 3200mhz 18-20-20-38 but the two sticks i added was a tad lower, like 16-18-18-34 or w/e but i dont recall and cant see it from here. Was told such ram was fine because the pc would just match to the lower one as long as the ram were the same make and model and had the same gb and mhz, least according to others. I dont know if this matters but the faster ones was in dim slot 2 and 4 but after swapping ram around, the 16-18-18-34 is now in dims 2 and 4 and the other two are now in dim 1 and 3.

Also canceled the order on the PSU after the pc seems to be stable now. Im not a OC'er and if 650w is enough im ok, least for now.

 

[Vic 40]

Yeah not properly seated will surely give problems and yes the tighter latency ram 16-18-18 will be matched to the 18-20-20 ram for timings. Can always check with cpu-z or hwinfo.
Maybe by setting ram voltage a litle higher (like 1.4V) could you gain the lower timings back, but performance wise will it not matter that much.