cancel
Showing results for 
Search instead for 
Did you mean: 

General Discussions

Rubbler
Journeyman III

New AMD build constantly crashing

Hi everyone - I've just recently built my first new PC in almost a decade. I've got a lot of past experience building but am currently stumped on this one.

I picked up a Ryzen 9 7900X3D, Asus Prime B650-Plus MB, 32GB (2x16GB) Team Force 6400MHz RAM, a Thermalright AIO liquid cooler and am currently using a 6650 XT GPU until a new 7900 GRE is delivered. Also have a Corsair 850e PSU.

Build went smooth, all parts recognized in BIOS, no worries.  When I installed Windows 11 I began to have random crashing, through the install process, once it was installed, etc.  The PC would crash between 10s-60s after bootup, consistently.  When this happened there was nothing in the logs, Event viewer captured nothing other than an error after bootup saying the PC had restarted from an unexpected shutdown, but I could not figure out what was going on.  Eventually I got to the point that having only 1 stick of ram in the system would keep it stable, I never had any crashes doing this.  Now, I tried both sticks, in slots 2/4 on the MB and they both worked perfectly fine on their own, in either slot.  As soon as I'd add the 2nd stick the system would go through the cycle of crashing.

I figured it might be the MB since each RAM stick was fine on its own so I RMA'd the board.  While waiting for the new one to show up I decided to try installing Win10 and it seemed to be much more stable than Win11, I thought it was fixing the problem, but alas it was only delaying the crashes up to 10-15 mins.  However same result as Win11, no error logs anywhere I could find.  I did once get a BSOD and it mentioned a heap problem, which from my Googling indicated a memory error, so I ordered a new set of T-Force RAM, just to compare against the old set in case it ended up being a RAM issue.  Got the new RAM, installed it and tried installing Win11, and same issue as before, so doesn't seem to be a RAM problem.  On a lark I thought since I have 64GB of RAM I may as well install it - and amazingly the system is much more stable.  It still crashes, but seems much more random than before and I can use the PC for a while without it happening.

I have the RMA board now from Amazon, but haven't installed it yet because now I'm wondering if it's a board issue with it actually being busted.

The PC is still shutting off and rebooting randomly, and while browsing MB's for fun it dawned on me the cheap (relatively) Asus board I have only has 1 8-pin atx 12V power socket.  Almost all other MB's I see have 2 8's or an 8 and a 4-pin.  It made me realize my crashes aren't like standard Windows crashes but more like the PC just shuts off for a second then reboots itself, with no errors as it's instantaneous.

 

So I'm left here wondering what to do next - does the 7900X3D require more than 1 8-pin power?  The MB supports the CPU so I would assume it was capable of running the chip, even if it spikes in power or whatever but I'm starting to wonder if that's my problem?  

Should I swap the MB's anyway and see if the issue goes away?  Is there anything I can do to validate my CPU is fine and not causing the crashes?  I'm just not sure where to go.  This build cost me a ton of money and it's really discouraging I can't use the machine for more than a few mins at a time.  I just realized the potential power issue 5-mins before writing this post but maybe it's me grasping at straws.

Anything anyone could offer to help would be super appreciated.

 

Thanks very much.

0 Likes
6 Replies
FunkZ
Paragon

A single 8-pin CPU power connector is rated for up to 384W so only having one even for a 7900X3D is no problem. The entire line of ASUS Prime motherboards are very skimpy on VRM components and power phases which can lead to other issues like low boost clocks.

However given your problems go away with a single stick of memory installed likely means it's memory-related, not power. Make sure you are installing the DIMMs in the correct two slots for dual channel operation and leave XMP/EXPO disabled in the BIOS until you can verify stable operation. Also check your motherboard memory QVL list to see if ASUS has tested and certified that Team Force memory.

Ryzen R7 5700X | B550 Gaming X | 2x16GB G.Skill 3600 | Radeon RX 7900XT

I've validated the RAM is on the QVL list and also tried non DOCP and same issue. That's why I got the second set of memory but same issue. 
I'll try the memory context restore feature mentioned below and report back. Thanks for the help

0 Likes
johnnyenglish
Grandmaster

You can try this if you haven't

 

Memory Context Restore to disabled instead of auto or enabled and if you use all 4 sticks, its really hard to make it stable at the kits rated speed.

The Englishman
0 Likes
ThreeDee
Paragon

For starters .. not a fan of ASUS anymore .. but I digress (ASRock has been my go-to since early AM4)

Is your motherboard's BIOS up to date?

Do  you crash with EXPO/XMP/DOCP disabled when RAM sticks are installed in slots A2/B2 (2nd and 4th slots away from the CPU socket)?

How are your CPU temps?

Are you running any cable extenders INSIDE your case? .. PCIe riser cable .. decorative power cable extenders .. etc

How is the airflow thru your case? You have to keep your VRM's and RAM cool too .. 

 

It could be that you just have a dud IMC (Integrated Memory Controller) on your 7900X3D 

 


ThreeDee PC specs
0 Likes

I was always a fan of MSI and Gigabyte boards years ago but considered Asus pretty reputable. Seems they may all have fallen downhill based on reviews...
First thing I did was update to the latest bios from April which had no effect. Whether Expo/DOCP enabled or not I have the same crashing. And yes was using A2/B2 with both sets of RAM. PC wouldn't even boot using A1/B1. 
Cpu hovers in the low 40's for temp. It will spike at 81 if I can get it to stay on enough to run a stress test but then goes back to 40 within seconds after I end the test. So temps seem good. No extensions or anything else in the case and airflow is great. 

I was worried about the ICM... I got the cpu on a steal of a sale on Amazon and now it looks like they're cleared out so no replacements. That's said, is there a way for me to check to rule it out or in? It's obviously intermittent if it is having issues but any dumps or logs I can check or tests I could run to verify if that's it?

0 Likes

any memtest type utility .. but maybe try lowering your FCLK to under 2000. Check SOC voltage that the motherboard isn't inadvertently running at 1.3v or higher (My 6400 RAM kit's XMP sets SOC to 1.3v .. I run 4 different ASRock AM5 setups in my home)

 

You could also reseat your CPU and check for any discoloring (burn marks) on bottom of CPU... also some have reported issues from overtightening their AIO's on their CPU's on AM5


ThreeDee PC specs
0 Likes