POST troubles (CPU+RAM Error LED) ASRock X570M Pro4 + Ryzen 7 3800X + 64GB Crucial RAM

D

DeaffyMan

n00b
Joined
Nov 20, 2006
Messages
18
Hello [H],

I built a new rig based on:

Mainboard: ASRock X570M Pro4 mATX
CPU: Ryzen 7 3800X
CPU Cooler: Noctua NH-L12S in High Clearance Configuration (fan on top of cooler)
RAM: 2 x 32GB Crucial DDR4-3200 (CT32G4DFD832A)
Storage: Intel SSD 660p NVMe

PSU: Silverstone Strider 600W
GPI: GeForce GTX 970
Case: Jonsbo V4 mATX Case, extra 80mm fan at the back.

Initially, everything went well, the system booted without any problem. I just went into the BIOS to set the boot drive, but changed nothing else. Memory was fully recognized (slots A2+B2). I installed Windows, pulled updates, multiple reboots, and generally worked on the system for about 2 hours. All temperatures were well in hand, I'd say never in excess of 60°C on the CPU. Suddenly, the system shut down and wouldn't turn on any more.

I had to pull the power cord from the PSU and let it sit for a while. When I opened up the case and tried to reconnect power, the system spun up for a few seconds, but shut itself down without POST or boot. After I let it sit overnight with power cord disconnected I was able to boot on the next morning as if nothing had happened but the system shut itself down just after a few minutes. Again, nothing suspicious on the temperatures.
I opened up the case and tried to get it to boot, but since then I've pretty much had no luck:
When the system turns on, fans start spinning, but the CPU and RAM power on status check (PSC) LEDs are on, and the system turns off after a few seconds again.
Sometime it stays on for maybe 20seconds, and then it might be that the PSC LEDs switch to VGA and briefly BIOS before turning off again. If i let it sit, this game will continue a few times, until it finally shuts off and remains off.

I made sure that everything is properly connected, tried to boot with only a single stick of RAM (the RAM is on the QVL at ASRock in it's DDR4-2666 variant, and on the crucial website, they promise compatibility for my exact sticks if you buy from them). Also tried to clear BIOS using the Clear CMOS jumper.
I even reseated the CPU cooler since I checked the CPU for bent pins, but nothing seems to be out of order. And I did countless trials of reseating the RAM, booting with just one of the RAM stick in slot A2 and so on.

I also used a different PSU to make sure that isn't the problem. The PSU and the GPU are from my old build, and the continue to run in that one as well (I checked to make sure they weren't damaged).

Barring the fact that the CPU was somehow damaged by heat (which I would really rule out, since I watched the temperatures like a hawk), it looks to me as if the mainboard is faulty. Maybe a regulator died? Any of you have any other ideas? I am basically asking the vendor to replace it for the same model, but I am worried what to do if this happens with the new board as well.

Cheers,

Tom
 
Double check you didn't plug 8-pin connection for you into the motherboard and that the 24-pin is fully seated (I doubt it's either of theee, but easy check). Run with a single stick of ram to test? re you running xmp or base speed for memory? You didn't mention when the GPU you're running. Also, where are you located, maybe someone local you can mix/match to check individual parts (I'm located @ 29856).

My vote is ram isn't making good contact someway.
https://amp.reddit.com/r/buildapc/comments/dvdxgm/asrock_x570_pro4_cpu_and_led_psc_lights/
 
Yes, did a lot of checking and replugging of the connectors and the RAM as well. I run the memory on stock BIOS (i.e. didn't change anything in the BIOS with respect to memory timing). The GPU is always present since the CPU doesn't have a GPU included, so it wouldn't POST without one.
Thanks for the offer to check parts, but I am actually overseas in Germany :D.

Cheers,

Tom
 
I'm reading recommendations to actually pull the BIOS battery out for a few minutes to fully reset the board and begin from there if it won't POST.
 
DeaffyMan said:
I'm reading recommendations to actually pull the BIOS battery out for a few minutes to fully reset the board and begin from there if it won't POST.
Click to expand...
Try this and I would try to reseat the memory again, even if it feels fine. Memory can be finicky in these boards.
 
A short,maybe got bumped just enough. Pull all out and reasemble on card board to take out a possible grounding. Start with the usual cpu,gpu, 1 stick of ram and go from there
 
I decided to return the board and ask for a 1:1 replacement, which arrived yesterday. And the board is doing the exact same POST error routine.

When I got the board, i installed it on a cardboard and attached all the peripherals. It booted straight away into windows and I worked on it for a few hours, this install software, I did a BIOS update to the latest 2.80 version. Generally, the system worked completely fawless, with everything looking 100% normal, i.e. RAM was being recognized at full capacity and with correct timing. Again, the temperatures were low and uncritical, especially since it was still sitting outside the case.

So, when I was done in the evening I shut it down properly. I came back to it this morning, expecting to be able to install it in its case and continue flying happily. What I found however was that the system had turned itself on during the night and was performing the POST error bootups as described above, again with the CPU and RAM PSC LED turning on most of the time and shutting off shortly after. Again, this happened without anyone touching the system going from perfectly fine to completely broken with nothin in between. So, once again, I think i can be 100% sure that there is nothing wrong with the other hardware, but I strongly suspect that the board design by ASRock must be faulty in some respect. My biggest suspicion are the VRMs, since the initial PSC sequence seem identical as with successful POSTs, where the board is initially turning on those two LEDs and doing multiple runs of LEDs showing up, before finally POSTing. The difference is, that the power turns off in the middle of this POST sequence and resets the process with no chance of completing. To me, this looks like a problem in the power delivery on the MB.

I have now disconnected everything and pulled the CMOS battery to let every capacitor drain and will try again in a few hours. I fully expect to return this board for good and get a different X570 board. The sad thing is that this is the only µATX or mITX board on the market which has a PCI-E x4 slot next to its PCI-E x16 so that I can fit my ASUS 10G NIC. This means that I will now need to look for a bigger ATX board and a new case :/

Cheers,

Tom
 
Hmm, I just got it to boot without any problem again. In my previous tries, I had the power button of my case connected to the board and was using that to switch it on.

I got suspicious after noticing that the PSC sequence looked exactly the same as with successfull POSTs just with the power turning off in the middle of it. Ruling out the PSU since that worked in my old system without problems, I turned to the case. I basically disconnected the system panel header on the board and briefly touched a jumper cap across the two pins on the board for the power switch and it booted without any issue.

I used to think that all today's power switches are basically just buttons that short the connection for as long as you press them. Maybe this one is different and it actually toggles the connection between open and closed whenever pressed. In this case the MB would think that the user is continuously pressing the power button, which usually means that the user wants the system to shut down immediately without waiting for the OS to shutdown. This is strange and needs some more investigating.


Edit: Used an Ohmmeter and checked that the power button is doing what you would expect it to, the only difference being that in the unpressed state my Ohmmeter is showing something like 2.9 Mega-Ohm and doesn't display the connection as fully open/infinite resistance. Maybe the board doesn't like that.

Cheers,

Tom
 
Last edited:
DeaffyMan said:
Edit: Used an Ohmmeter and checked that the power button is doing what you would expect it to, the only difference being that in the unpressed state my Ohmmeter is showing something like 2.9 Mega-Ohm and doesn't display the connection as fully open/infinite resistance. Maybe the board doesn't like that.
Click to expand...

Nice job troubleshooting that. What a weird issue.
 
You must log in or register to reply here.
Back
Top