Issues and sudden failures with certain ASRock boards and Ryzen 9800x3D pair.

gogogozillla

gogogozillla

Beginner
Came across this reddit post today where a user has been tracking a number of failures on ASRock motherboards paired with a 9800x3D. He links a number of such posts for evidence and ASRock supposedly has been refusing to acknowledge the issue and denying the issues as isolated incidents.

Sharing this here so ASRock users can be aware and report the problem immediately and future buyers can reconsider other options till ASRock offers a fix. If a fix is offered or ASRock offers any other solution such as RMA/refund, I will share an update on this post/thread.

Link:
https://www.reddit.com/r/hardware/comments/1jmqajy/dead_9800x3ds_in_asrock_boards
So as of today, there are 108 cases and counting. It is not limited to ASRock motherboards, other brands are also affected. It is likely to be an x3D processor issue, because some 9950x3D users have also faced this sudden failure. [Link]

There is a web form, which is being used to collect case information, to hopefully identify the faulty batches causing this issue. If you are affected by this problem, please fill the form: [Link]
 
Last edited:
  • Like
Reactions: Adwaith

Attachments

  • 487739947_24376844388582425_6801188083346069328_n.jpg
    487739947_24376844388582425_6801188083346069328_n.jpg
    14 KB · Views: 30
  • Like
  • Wow
Reactions: DBot, PotBelliedGarfield, bssunilreddy and 1 other person
Darth Vader said:
Even during that fiasco, many stats from sources like Pugetsystems pointed out high failure on AMD CPUs more than the flawed Intel ones.
Click to expand...
Puget says its Intel CPU failure rate is lower (2% in 13th- and 14th-gen) than AMD Ryzen failures (4% in 5000 and 7000 chips) — system builder releases failure rate data, cites conservative power settings read more here

Some Ryzen 7 9800X3D CPUs are allegedly failing prematurely — over 100 cases documented based on user feedback read more here

Most of the Ryzen 7 9800X3D problems seemed to have occurred on ASRock motherboards.
Up to 82% of the incidents were on an ASRock motherboard.
Asus, MSI, and Gigabyte rank second, third, and fourth, respectively, with participation rates of 13%, 4%, and 1%.
 
Emperor said:
Up to 82% of the incidents were on an ASRock motherboard.
Asus, MSI, and Gigabyte rank second, third, and fourth, respectively, with participation rates of 13%, 4%, and 1%.
Click to expand...
82% is def concerning, even 9950X3D CPUs are dying too as reports popping up on reddit across multiple brand boards. For now, avoid Asrock and if possible avoid this gen (AMD/Nvidia/Intel) altogether lol. No one offered anything of value this time.
 
Emperor said:
Puget says its Intel CPU failure rate is lower (2% in 13th- and 14th-gen) than AMD Ryzen failures (4% in 5000 and 7000 chips) — system builder releases failure rate data, cites conservative power settings read more here
Click to expand...
You seem to have left out some further interesting bits in your paraphrasing from the article:

  1. Aside from the CPUs you have mentioned, note the article also mentions that Intel's 11th-gen chips had the highest failure rate - "Another interesting tidbit that Puget revealed is that Intel’s 11th-generation Intel Core processors had the most failures among all recent processors, with a recorded rate of over 7%. The issues with 11th-Gen Intel Core processors haven’t been widely reported, so we’re unsure about their cause and how (or if) Intel resolved them."
  2. "As with all failure rate data that doesn't come directly from the chipmaker, Puget's data should be taken with a grain of salt."

The 13th and 14th-gen chips had ACTUAL CPU defects due to a messed-up microcode implementation. With the 7000 series it was due to excessive default SoC voltage from the motherboard. This was fixed with a simple voltage limit cap. The affected 13th and 14th gen chips were damaged permanently. Regarding the 9800x3d and (mostly) ASrock motherboards, let's await more concrete info on the cause(s) behind the recorded incidents.

You can be sure GN is already dissecting this like a frog in a biology lab :hilarious:
 
There is more if you dig deeper. Some of these could be because of user errors.

Improper installation example here:

Sample of user error

There are other possible explanations of the issue. I would suggest going through the gamernexus extensive investigation and make your own conclusions.
 
smaratha said:
There are other possible explanations of the issue. I would suggest going through the gamernexus extensive investigation and make your own conclusions.
Click to expand...
Didn't this guy concluded the whole 12 VHPWR burn issue on RTX4090 as "User error" two years ago ? And no one questioned him/investigated the issue properly.
With 5090, the story was very different and he as usual joined in the end. I like his case reviews but not these ill-informed investigation.
 
Darth Vader said:
Didn't this guy concluded the whole 12 VHPWR burn issue on RTX4090 as "User error" two years ago ? And no one questioned him/investigated the issue properly.
With 5090, the story was very different and he as usual joined in the end. I like his case reviews but not these ill-informed investigation.
Click to expand...
He never concluded that it was solely user error. It was clearly mentioned that it was a combination of user error + design flaws.

Also on the 40 series, unlike the 50 series, the cable was EXTREMELY irritating to lock in, with very little wiggle room for errors. So yes, for the 40-series, user error induced by a hard-to-secure cable was a valid reason to deduce at the time, but not the sole one.

We also have to keep in mind that this was during the early days of the melting-connector saga. Not only GN, but most tech analysts never delved deep into what was causing it, unlike the 50 series, where more research was done collectively.
 
It's just best to go with the cheapest B850/B850M motherboards if you really want PCIe 5.0 and NVME Gen 5 or go with the somewhat expensive B650/B650M mobos from Gigabyte & MSI which have PCIe 5.0 and NVME Gen 5 because these tried and tested boards. Only go with the above mentioned brands that's all.
 
bssunilreddy said:
It's just best to go with the cheapest B850/B850M motherboards if you really want PCIe 5.0 and NVME Gen 5 or go with the somewhat expensive B650/B650M mobos from Gigabyte & MSI which have PCIe 5.0 and NVME Gen 5 because these tried and tested boards. Only go with the above mentioned brands that's all.
Click to expand...
B850 is garbage for the price you have to currently pay for it. They are a 1-1 rebrand of B650, with mandatory 5.0 support for the primary NVMe slot, not even the primary x16 slot. That feature is optional, the same as a good B650. B850 boards start at 17-19k. For 19K, you can get a MUCH better B650.
 
  • Like
Reactions: bssunilreddy
Top Bottom