AMD comments on burning AM5 sockets — chipmaker blames motherboard vendors for not following official BIOS guidelines

(Image credit: Shutterstock)

AMD has finally responded to recent socket burnout issues plaguing some third-party AM5 motherboard vendors — including ASRock. Quasarzone had a Q&A session with AMD's David McAfee and Travis Kirsch, who explained the reason behind these burnout issues.

AMD claims the underlying problem behind the burnout issues is caused by some ODM BIOSes that do not adhere to AMD's recommended values. However, AMD has further clarified that the issue is complex and is working closely with its partners to resolve it. AMD also recommended users update their motherboard BIOS to the latest version.

AMD is inevitably discussing motherboard vendors who push AMD CPUs beyond their default specifications, even in minor ways, through modifications to voltage, power limits, and other tuning variables in their board's UEFI (BIOS). Some of Intel's board partners do the same thing. But it is also something AMD's board partners can do, albeit to a lesser extent. AMD's somewhat confusing wording also suggests that the burnout issues might go deeper than just overtuned BIOS settings.

Quasarzone AMD Q&amp;A regarding AM5 burnout issues — Text translated via Google Translate (Image credit: QuasarZone)

Earlier this year, an unusually high number of Ryzen 7 9800X3D failure reports were posted on Reddit, with the majority of the reports shared by users with ASRock motherboards. The abnormality was so significant that ASRock moderators created a Reddit megathread dedicated to the issue entirely in an attempt to figure out the root cause of the problem(s). Some of these failures were so severe that they caused the socket to "burn out," resulting in physical damage to the socket.

ASRock soon stepped in and made several attempts to fix the burnout/failure issues. Initially, ASRock blamed memory compatibility issues as the cause and attempted to resolve them with BIOS updates. However, that did not fully solve the problem. Its latest update was in late May, when ASRock confirmed that its latest BIOS version at the time, 3.25, altered a plethora of PBO settings to rectify the burnout issues. These included updates to EDC, TDC, and "shadow" voltages.

So far, this update has appeared to resolve the issue for most users, as the problem has significantly decreased in severity since ASRock's 3.25 BIOS update went live. Regardless, some Redditors claim the burnout issues are still present and affecting their ASRock AM5 boards.

Follow Tom's Hardware on Google News to get our up-to-date news, analysis, and reviews in your feeds. Make sure to click the Follow button.

See more CPUs News

Aaron Klotz is a contributing writer for Tom’s Hardware, covering news related to computer hardware such as CPUs, and graphics cards.

12 Comments Comment from the forums

hotaru251

AMD claims the underlying problem behind the burnout issues is caused by some ODM BIOSes that do not adhere to AMD's recommended values.

which is why there should be a warning specifically stating if a bios OC (which should NEVER be on by default) is within spec or out of spec of even amd's oc specs.
Reply
spoidz

I decided to stick with an ASROCK Nova for my newest build because of the PCI-E sharing.

I then found two specific videos showing that the BIOS/MB is already over spec at default but can be reset to stay under spec.

OcG1XANqRYQView: https://www.youtube.com/watch?v=OcG1XANqRYQ - 9800X3D

TofvZbLE3XQView: https://www.youtube.com/watch?v=TofvZbLE3XQ - 9800X3D

I haven't completed the build yet in order to check whether the latest 3.30 BIOS has completely returned to safe spec.
Reply
Jagar123

Doesn't really make me feel any better about having an ASRock board for my 9800x3d.
Reply
93QSD5

Jagar123 said:
Doesn't really make me feel any better about having an ASRock board for my 9800x3d.
Why not? If you're still under warranty, you're covered.
Reply
jwnm

93QSD5 said:
Why not? If you're still under warranty, you're covered.
Resolving a warranty issue is another level of suck.
Reply
93QSD5

jwnm said:
Resolving a warranty issue is another level of suck.
I agree but if you are already in this mess (purchased) and the fact that there are lots of folks out there with the same config without any issues (otherwise there would be a bankrupting roll-back), then honestly I would take the bullet and cozy up with your new friends over at AMD & Asrock if something ever happens.

It sucks but spending hundreds of bucks out of your own pocket for the manufacturers mistake is worse. Don'r reward bad behaviours.
Reply
cristiano99

ASRock is unreliable.. only their MB broke AMD processors, not Intel..
Only their video cards melt & take fire..
Reply
SomeoneElse23

cristiano99 said:
ASRock is unreliable.. only their MB broke AMD processors, not Intel..
Only their video cards melt & take fire..
I have three ASRock motherboards, two X570 and one 870E. All rock solid.

But I don't push the envelope.
Reply
King_V

nogaard777 said:
I have to question AMD's wisdom on this one throwing ASRock under the bus. We all know what a massive mess the raptor lake failures were but Intel never once tried to blame Asus when it's a fact that 99% of all the failures were on Asus boards pushing the chips to their limits out of the box.
...
Trust me I was watching this VERY closely
Links from reputable sources to confirm those numbers would go a long way in establishing the trust you're asking for.
Reply
cerata

nogaard777 said:
I have to question AMD's wisdom on this one throwing ASRock under the bus. We all know what a massive mess the raptor lake failures were but Intel never once tried to blame Asus when it's a fact that 99% of all the failures were on Asus boards pushing the chips to their limits out of the box.

There was a tiny smattering of raptor lake failures on gigabyte and I never saw a single report of a failure on an ASRock or MSI board. It was always Asus. Trust me I was watching this VERY closely because I built literally dozens of raptor lake systems the last several years and have to warranty them. I never had a hiccup reported from any of the buyers or my own family, but I also exclusively use MSI and ASRock because I have a bad history with Asus and Gigabyte over the last 30 years.

I'm not saying Intel was innocent of course, but it's pretty clear Asus' settings were heavily accelerating any flaw in RPL, and Intel never once tried to shove the blame onto Asus. This is literally the exact same issue with ASRock pushing Zen5 and it's PBO settings.
If 99% of failures were on ASUS motherboards (including laptop), then it's surprising we haven't seen more tech media coverage of that. ;-)

Raptor Lake degradation/failure was, in Intel's words, due to "a clock tree circuit within the IA core which is particularly vulnerable to reliability aging under elevated voltage and temperature." To clarify, IA is short for Intel Architecture, so "IA core" just means a CPU core on an Intel chip. They identified four scenarios that led to these elevated voltages/temperatures: motherboard vendors pushing power delivery settings beyond Intel guidelines, and three different bugs in Intel's microcode.

Say ASUS's desktop BIOS settings were indeed the most egregious, and Intel tried pinning all the blame on them (or on motherboard vendors in general). The issue wouldn't have gone away -- there's plenty of reports of Xeon servers suffering degradation & instability too, and servers generally aren't meant to be overclocked.

Of course, AMD could well have written buggy microcode that causes 9800X3Ds to burn up, and/or the similar problem seen early after the 7800X3D's launch. If the problem is due to some physical defect during fabbing & packaging, then TSMC might also responsible. But we don't know that yet. Until we do, this is not "literally the exact same issue" as with Intel.
Reply

Show more comments