AMD Investigating Reports Of USB Connectivity Issues on 500-Series Chipsets

[BigDamn]

Summary

 

AMD is investigating user complaints of USB devices briefly disconnecting on 500-Series motherboards when using a Ryzen 3000 or 5000 series CPU. It seems that the issue is exacerbated when using a PCI-E 4.0 GPU, so some users have rolled back to PCI-E 3.0 via a BIOS setting which has helped, but not eliminated, this issue. AMD is asking users impacted by this issue to gather information, such as steps to reproduce this issue, system logs, and relevant system info to help track this down.

 

Quotes

Quote

"Some AMD users have been posting on social media, including Reddit, about experiencing intermittent connectivity problems when using USB ports on their 500-series motherboards. In response, AMD has reached out through a Reddit post that it has been made aware of the connectivity issues, and is requesting more information from specific users with the issue to help them resolve it. Since this announcement, many users have been reaching out to AMD."

 

According to Anandtech, it appears that this impacts USB 2.0 more than 3.0/3.1:
 

Quote

"One possible workaround touted to reduce (but importantly not completely stop) the issues is users changing the motherboards' settings from PCIe 4.0 to PCIe 3.0. Users on Reddit have been reporting although it has reduced the frequency of dropouts, it hasn't eliminated them, which could point to an issue between the USB hub controller and other parts of the chipset. Some users are also reporting the issue to be primarily on USB 2.0 ports, and some have had success by switching devices to USB 3.1 ports. It is also recommended to update the firmware to the latest AGESA version on models, which could reduce issues. Although at this point, it's possible to presume that none of the workarounds could fix these problems users are experiencing with random and intermittent USB cutouts."

 

My thoughts

Although this seems to be impacting a small number of users (potentially due to PCI-E 4.0 not being widely adopted yet), I believe it's fair to say the Ryzen 5000 and 500-series chipset launches haven't been AMD's smoothest launch. First we had the delay in 500 series chipsets, then a high failure rate in Ryzen 5000 CPU's, and now this. I question whether AMD was able to perform a proper QA process on these chips with Covid limitations through most of last year. For those of you with 500 series chipsets, have you encountered issues like this? If you have, make sure to let AMD know on their Reddit post regarding this issue. Hopefully this is not a hardware level issue and will be fixed in later AGESA updates.

 

Sources

 https://www.anandtech.com/show/16498/amd-usb-connectivity

https://www.pcinvasion.com/x570-and-b550-motherboard-usb-issues/

AMD Reddit Post

 

[MageTank]

2 hours ago, BigDamn said:

First we had the delay in 500 series chipsets, then a high failure rate in Ryzen 5000 CPU's, and now this. I question whether AMD was able to perform a proper QA process on these chips with Covid limitations through most of last year. For those of you with 500 series chipsets, have you encountered issues like this? If you have, make sure to let AMD know on their Reddit post regarding this issue. Hopefully this is not a hardware level issue and will be fixed in later AGESA updates.

 

Sources

 https://www.anandtech.com/show/16498/amd-usb-connectivity

https://www.pcinvasion.com/x570-and-b550-motherboard-usb-issues/

AMD Reddit Post

 

Still waiting for a credible source on this issue. I remember that thread about the one random integrator that had an insane failure rate, but that was it. My employer has integrated with almost a thousand of these 5000 series CPU's since their launch in November and their failure rate has been absurdly low. I can count on one hand the number of chips that actually arrived DOA. Far more were killed by line techs bending pins by accident and trying to unbend them.

 

Most of the random quirks plaguing these processors is more related to the buggy AGESA firmware that keeps getting pushed as ready when its not. 1.1.0.0 absolutely wasn't done yet, and 1.2.0.0 is barely there for these CPU's assuming you don't mind broken FCLK straps when pushing 1933-2033 on some boards. Shockingly enough, 1.1.9.0 felt a tad more stable than 1.2.0.0 and 1.1.9.0 was supposed to be the pre-release of 1.2.0.0, lol.

 

Still, I can't say that I've experienced this USB issue (yet). Don't own a PCIe 4.0 card, but I've toyed with several PCIe 4.0 SSD's on my board and have not had this issue present itself. I've seen my fair share of complaints about it on various forum threads to know that it exists, I just hope something else doesn't get broken in AMD's quest to patch it.

[StDragon]

1 hour ago, MageTank said:

Most of the random quirks plaguing these processors is more related to the buggy AGESA firmware that keeps getting pushed as ready when its not. 1.1.0.0 absolutely wasn't done yet, and 1.2.0.0 is barely there for these CPU's assuming you don't mind broken FCLK straps when pushing 1933-2033 on some boards.

Why is this an issue? I know firmware have bugs and features added, but it seems that AMD's handling of AGESA has been a unmitigated dumpster fire! 

[Chiyawa]

Well, this post went silent... I think the issues is not as big as it looks?

 

Anyway, it could be a lot of factor that causes instability of the system. To pin point one out could take a lot of time.

[Avcn42]

1 hour ago, Chiyawa said:

Anyway, it could be a lot of factor that causes instability of the system. To pin point one out could take a lot of time.

Exactly. I have a decent amount of USB devices always plugged in and haven't had any problems with my system (AGESA 1.2.0.1). Would be interesting to see if others here have stumbled upon any of these issues. 

[thechinchinsong]

On 2/24/2021 at 9:31 AM, BigDamn said:

First we had the delay in 500 series chipsets, then a high failure rate in Ryzen 5000 CPU's, and now this.

I thought the high failure rates was something that was disproven.

[Bombastinator]

5 minutes ago, Nacht said:

these issues happen when you use pci-e 4.0 like having a nvme that supports pci-e 4.0 or gpu etc putting those devices in 3.0 resolves the issue, i dont have issues cos i dont have any pci-e 4.0 devices yet if i got a rtx 3080 now i probably put it in 3.0 mode altho i would test it at 4.0 just to confirm i have it or not.

Apparently it doesn’t completely though.

[trufret]

Hoping they fix it soon. I've had nothing but problems with my Reverb G2 on an X570 board. Doesn't work at all on USB ports on back of mobo either A or C ports.

Works some of the time on case ports connected to the USB 3 header but will regularly have audio crash and switch from headset to speakers and back once it recovers.

I've just installed an USB 3 PCIE card to try and eliminate the issue. seems to be working so far.

[ne0tic]

Thanks god they are publicly investigating this! I have had a lot of USB issues since I built my PC in December... I bought a 3.1 USB-hub and changed to PCIe 3.0 which almost solved the USB issues, but not completely so I'm really happy they now have said that they are looking into it. Hopefully it is not a hardware issue tho...

[Loote]

Would be nice to get some update on this. I hope they manage to fix it soon.

[avipars]

My Wifi/Bluetooth card have been causing issues with my B550... I think it's related to this issue but can't verify... What's surprising is the amount of time it took form release to acknowledge these issues