Hi there, hoping to get some idea’s / guidance on how to proceed with this issue.
Built a new Ryzen 3900x based system last year, dual booting linux and windows with linux being my daily driver. A few months after I initially built the PC it just wouldn’t turn on at some point… something shorted somewhere and despite being pretty familiar with PC’s neither my sysadmin buddy or myself could figure out what happened, but the PSU and motherboard were dead. No obvious signs of shorting anywhere and everything is sitting behind a UPS.
RMA’d both (gigabytes RMA process is a headache compared to corsair btw) and things seemed to be fine for several months again until yesterday and same exact thing, system doesn’t boot at all and no signs of life from either PSU or mobo… still in the process of pulling everything apart to figure out what happened but I am strongly suspecting I’ll get the same results.
The first time around I was messing with memory overclocks and the usual but nothing crazy or even serious beyond plugging in numbers from DRAM Calculator if I recall correctly. Second time around I’ve left everything untouched except turning on virtualization and default XMP.
The board I got back from Gigabyte the first time was a refurb (different board, serial numbers were different) and had some sort of film around the socket, it brushed off and I cleaned it up with alcohol, I think it was just a really sloppy thermal paste job when they were doing whatever tests on it before shipping it to me.
My questions are thus:
The obvious one, what could be happening here that I’m missing?
Should I just sell whatever they give me back from RMA (assuming this is the issue) and get a different manufacturer?
I haven’t really been keeping up with hardware, is Zen 3 having these issues or am I likely just an outlier? I do have two friends with Zen 3 that haven’t had any issues at all so I’m sure I’m just unlucky…
Am I being unreasonable for wanting to just ditch AMD and hop back on the Intel train? I know with the new Zen architecture all the manufacturers do their own implementations which has caused some to be (at least as of Ryzen 3 launch) significantly more reliable than others… I thought I had basically the widely regarded best motherboard for Ryzen 3, but clearly not. I have also seen various content creators suggest the motherboard manufacturers voltages are not always accurate (I forget what the report was exactly) and honoring what the CPU is requesting or something? Basically overvolting slightly for additional performance?
TLDR: Intel is a tried and true platform and while I love that AMD is back in the game in a big way and I want to support them, I don’t want to keep having this shadow over my head. Maybe there is some weirdness going on with my specific peripherals or just the motherboards themselves that the battle-hardened Intel 14nm would handle more gracefully.
The system details:
Bios was F11 second time around.
Type
Item
CPU
[AMD Ryzen 9 3900X 3.8 GHz 12-Core Processor]
CPU Cooler
[Noctua NH-D15 82.5 CFM CPU Cooler]
Motherboard
[Gigabyte X570 AORUS MASTER ATX AM4 Motherboard]
Memory
[Crucial Ballistix Sport LT 32 GB (2 x 16 GB) DDR4-3200 CL16 Memory]
Storage
[Sabrent Rocket 1 TB M.2-2280 NVME Solid State Drive]
Video Card
[Gigabyte GeForce RTX 2070 SUPER 8 GB GAMING OC WHITE Video Card]
Case
[Corsair Crystal 570X RGB ATX Mid Tower Case]
Power Supply
[Corsair RMx (2018) 750 W 80+ Gold Certified Fully Modular ATX Power Supply]
(long time lurker and fan of the youtube channel, sorry this is my first post!)
I did swap the cables the first time it died since it took out both PSU and mobo and i was concerned about faulty cables or something… though i did read a gamers nexus article yesterday about modular cables not being 1:1 like you might expect which i wasnt aware of, guessing that what youre alluding
I have double checked with my wifes computer and PSU, my PSU is actually fine this time around but motherboard is definitely dead.
If nothing else, which motherboard manufacturers have good support?. EVGA and Corsair are the only ones I am aware of but they don’t do motherboards and like i said i dont pay attention to hardware and things like that when im not actively building. I will say corsair was excellent on the first RMA.
Wow, sounds as if you have had a whacked streak of luck - it’s got to be frustrating in the least.
Here’s my thought and I’m just a cat on the inter-webs so take it for what its worth. To ditch AMD and migrate to Intel you’re going to need to buy yet another CPU and motherboard - kinda spendy - sure you may recoup some cash selling the 3900x used and the rma’d board (whenever you get it back) but that could take a while unless you got a buyer lined up - and I’d guess too, that you might not re-coupe your full original investment in those parts.
You are already invested in the AMD platform, and the 3900X is a great CPU. Yes, it has been a bit of a ride but not sure the platform itself is entirely to blame - there are faulty parts produced on every production line - it just a thing that is.
I have been super pleased with the x570 Master (I have 2 of them in service, one since the Ryzen 3000 launch). I have been equally pleased with my x570 Unify’s (Have 2 of them as well). I get that you may be frustrated with Gigabyte - I mean what you described getting back sounds a bit odd. So, my suggestion to you then, would be the x570 Unify, based from my experiences.
Motherboards are like cars its sorta a subjective subject - Some like Fords others Chevy, and others Audi - So there will always be someone offering up another opinion - but it’s you who will need to make a choice and get your own experiences.
Geez - OK TLDR;
Keep your CPU, get a different board, switch brands if Gigabyte has lost your trust.
Also, RMA has been a pain with every board manufacture I’ve ever dealt with. Though, I share your experience with Corsair, their service and process was top notch the one time I needed them.
Okay good that you also replaced to cables as well.
Because faulty power cables was actually my first thought as well.
But the motherboard is completely dead?
No leds or fans spinning up?
The yeah the board is likely dead.
I assume you also tried a different psu as well to rule that out.
The Master is actually a great board one of the better boards out there for x570.
So that two of them failed on you in a relative short time period is strange.
It could be allot of things of course, the first thing i would really double check,
is the case, and then i mean things like the front usb ports etc,
to be sure that non of them are damaged or causing shorts.
Or any other unwanted contacts.
Of course it could be just bad luck.
If you get a new board from Gigabyte, i would firstly try everything outside the case,
on a cardboard box or something.
Just because you don’t mention it, did you try resetting the board by popping the CMOS battery? I have a Gigabyte mobo that occasionally dies completely, no lights at all, and I have to take out the battery for a while to revive it.
Haha wow physically removing the CMOS fixed it, the hell kind of black magic is this? Why would that the issue and why would it fix it? Some sort of BIOS memory corruption?
It booted up like a charm even without the battery in the case, put the battery back in and same. Still need to put everything back together and test but seems like its fine.
Thank you so much for the suggestion, strange reseting it doesn’t do the same thing.
I’ve got the same issue with my Gigabyte Designare EX Threadripper Board.
Sometimes it just wont start and i need to remove the BIOS battery and reinsert it to start the board again (CMOS reset button wont do anything either in that case)
The only thing i found may be in relation with the behavior is whether i shutdown the whole system using the power button (which triggers a clean shutdown of the main os) or if i shutdown all my vms before shutting down the main os (i’ve got a host os without gui and several vms with passed through hardware for each).
I suspect that one of the os-es leaves the system in unclean state which causes the non-start behavior of the board.
I do dual boot windows and linux (daily driver) and I’ve noticed I have to completely shut down to go from windows -> linux; otherwise it seems windows doesn’t let go of all the drivers correctly. Something to do with their fastboot I think but I’ve tried disabling everything I can related to that and couldn’t get it to cleanly work (mostly I’ve noticed my sound stops working, but there are probably other issues).
I do also run a few VM’s as well but they’re not always on and I always shut them off before the PC.
I vaguely recall messing around in my windows boot with Ryzen Master and Typhoon Booster/DRAM Calculator a couple boots before it died and I think similar’ish the first time stuff actually died. I didn’t actually change anything from the apps but I wonder if the process of reading the information or maybe windows doesn’t ship with the latest chipset drivers or something and because I only use it infrequently to game all of it in concert is seizing up the bios. All pretty wild speculation and anecdotal but I’ll try and see if I can identify a pattern if (hopefully not…) and when it happens again.
Well I’ve had it happen two or three additional times since then with seemingly increasing frequency; I’ve ordered a replacement CMOS battery though it seems unlikely a new board would ship with a faulty battery. I also moved the the power cable on the UPS off battery on the exceedingly remote chance the UPS is doing something (why moving from battery to just surge would affect anything, who knows, grasping at straws).
If and when it happens with the new CMOS battery I’ll probably just RMA this and get an MSI and sell the RMA.
I wouldn’t mind having to pop out the battery if it was accessible at all but the clearance between this Noctua cooler and my GPU is non-existent and it’s a huge hassle trying to get the GPU out every time.
