[TRACKING] FW13 AMD USB-A expanson card powers down (?) and does not wake up

I didn’t say it wasn’t a problem, just that it doesn’t appear to be related to the main issue discussed in this thread, which was confirmed by a mainboard swap which resolved the main issue.

I don’t know how to fix the error message, but it I also don’t see any side effects or issues caused by it. It might be a bug/quirk in the EC or the UCSI kernel module. Since the error message is easily reproducible on “supported” distros such as Fedora by simply unplugging and plugging the power cable, I’m sure the Framework team will have seen this at some point.

@Matt_Hartley as per my previous update, it looks like this was a mobo issue. Nearly 2 months later now with the new mobo and it all works perfectly with no issues. Do you think it’s worth marking this as resolved?

As for the ucsi_acpi error messages, just to give a final answer to @Nimrod_Hajaj as well, they appear to have been unrelated and have been resolved with the BIOS 3.05 update (also included in the ChangeLog). I’ve not seen them since I updated. In fact, my dmesg output is devoid of any errors, unlike my old ThinkPad which the FW13 replaced :smiley:

Finally someone who is (was) experiencing the exact same issue and even provides confirmation on the solution. @0xc0ffee thank you! I thought it was a software problem under Linux. Have opened up a support ticket with a link to this discussion.

2 Likes

I also have issues with the supplied 2.5g FW NIC adapter ; would be interested to see someone else who has it replicate

@Markus12 Glad this has helped, hope your issue gets resolved soon.

@jwp Could you provide a bit more detail on your issue and what makes the behaviour similar to what has been outlined in this thread?

So after almost a month of providing info, logs, photos, and videos of this issue occurring and answering some questions multiple times, Framework support was finally convinced that I was suffering from the same hardware issue as @0xc0ffee . They sent a replacement board which I installed and finally this issue appears to be gone.

P.S.: After doing a mainboard change, if the internal screen turns off shortly after showing the Framework logo during boot but can otherwise display the BIOS screens just fine, double-check the screen connector is properly seated on the board. Apparently I didn’t connect it fully and the screen was only able to display low resolution content. This really scared me at first into thinking the new board was broken as I would have expected the screen to fully work or not at all.

5 Likes

TL;DR: I had the exact same issue and it is now also resolved after I got a new mainboard. Support also requested some videos, photos and logs, but that’s all reasonable. I had a replacement mainboard in about two weeks and most of the delays were caused by myself being busy with other things and not have time to do the requested steps. So overall I’m pretty happy with the support process and how quick the responses were. During the tests the support was asking for, I noticed some new details, which aren’t mentioned in this thread yet, so I’ll write my new learnings down.


I already noticed this problem when I installed the Laptop for the first time, but back then it still had the old BIOS and I didn’t pay too much attention to it. And it then kinda “disappeared” and I forgot. It sometimes happened again, but I never found out what was triggering it. I thought it might have something to do with powersaving stuff, but couldn’t find anything and it was hard to test as it still usually didn’t happen when I was looking for it, and then it wasn’t important enough again and I kinda forgot again.

Skip forward to a few weeks ago, where I needed the USB-A port more often than usual, and suddenly it also happened every single time I wanted to connect something, and sometimes it even stopped working while a USB-stick was still plugged in. I was so confused, why it was suddenly so much worse than usual (usually when I re-connected the USB-A card it kept working for the rest or the day (or at least long enough, that I never noticed when exactly it stopped working). So it became very clear that it can’t continue like this anymore, I can’t use a laptop where the USB-A port stops working while I use it. But at least it was suddenly very reproducible.

I then also did some testing with a friend who owns a FW16, and we exchanged the USB-A card, to check if just my card is broken. But my card worked on his Laptop but his card showed the same broken behavior on mine, so it wasn’t related to the card. That’s when I also when I did some more searching in the Forum again and found this thread (which I somehow didn’t find the last time I searched). And I learned here, that apparently it’s only happening on the left side, so I tried connecting the USB-A card on the right, and indeed, it worked without breaking there. I never had the idea to try the other side, because the front-left port seems like the perfect port for an USB-A card, since it’s the only port that doesn’t support display-output (I have USB-C in both back ports, so I can charge from both sides, and had a HDMI card in the front-right … so permanently switching the sides of both front cards wasn’t an option).

I also tested with a live-Fedora and live-Ubuntu to check if the problem is reproducible there, or if it’s something in my system. And while it was reproducible on Fedora/Ubuntu (so not related to my installation), I noticed (while rebooting a lot of times) that the USB-A card doesn’t stop working if I was sitting in the BIOS, or on the bootloader-menu … it only broke as soon as I booted an actual operating system (I didn’t need to fully boot and login, loading the kernel/initrd was enough). I can only test with Linux, I don’t know about Windows? I don’t have a Windows-copy available, so I can’t test that. It’s possible that it only breaks when booting a Linux-OS and windows works (the same as the BIOS and bootloader don’t break it)? But it doesn’t matter which Linux-OS you boot, and since Fedora/Ubuntu are officially supported, and it’s broken there, it’s still broken. But this was my first new learning about this issue.

By the new learning that I can reproduce the problem on both left ports, but the card works perfectly on both right ports, I was convinced that I have the exact same problem, and opened a support request (mentioning this thread here) and asking for help for what steps are required to get a replacement board. They first asked, if the problem also happens if the charger is connected. I did the tests with the friend with the FW16 on battery, but in the week before, where I had the problem very often, I was connected to the power most of the time, so I knew it also happens with power connected. But I still did some more testing and that’s when I found the first new thing: At home, I usually have power connected on the left, and when I tested that, the problem disappeared again (that explains why I only had the problem “sometimes”, because it doesn’t happen when power is connected on the left). But in the other week, I was at a different place, and I connected power on the right (that’s the whole point of having two USB-C ports, so you can connect on whichever side is closer), and then the problem does happen (that’s why I suddenly had the problem so often). Connecting power on the left doesn’t bring the port back to life if it’s already dead, but it keeps it alive after the USB-A card was re-plugged.

So the second new learning was:

  • USB-A cards stop working on the left side, if either working on battery or if power is connected on the right.
  • USB-A cards keep working (if it already is in working state) on the left side, while power stays connected on the left side. It doesn’t matter if power is back or front, so even when the USB-A card is in the back-left slot, and power is front-left port, it keeps working, and breaks shortly after disconnecting power.

So it matters on which side power is connected if it breaks or if it keeps working.

The support then also asked, if I also have a similar issue with the HDMI-card, which so far I never used on the left side (where the problematic USB-Ports are). Since the front-left port doesn’t support display output, I tested the HDMI card on the back-left port, to see if it also stops working after some time. It didn’t … but what I noticed when testing that was, that the USB-A port (which was still in the front-left port and I re-plugged in the process of switching ports around) also didn’t stop working anymore. So for some reason, having the HDMI card plugged in into the other left-port, also keeps the USB-A card working, the same as if power is connected. So I also started testing other cards on the left, combined with an USB-A card.

And this is what I found out:

  • USB-C doesn’t do anything (which is what I had all the time anyway), as it is just a dumb pass-through
  • HDMI card, Network card and the Micro-SD card on the back-left port kept the USB-A card working on the front-left port
  • A second USB-A card on the left (so both of them on the left) also keeps them working, so they somehow keep each other working … but as soon as you remove one of them, the other one stops working after a short time.
  • The new big SD-Card expansion-card was special. If only the card-reader was plugged in, the USB-A card still stopped working, but if there was also a card sitting in the card reader, the USB-A port kept working. But it looks like the big SD-card reader is completely passive when nothing is in it (to save power?), it also only shows up in the USB-devices list, if an SD-card gets inserted and disappears again when the SD-card gets removed.

So the third learning was: It depends on if another active device is connected on the other left port. The empty USB-C card and the empty big SD-Card reader are both passive, so the USB-A card stops working … but the HDMI, Network, Mirco-SD, or even just a charger, are active, and if connected also on the left side, the USB-A card keeps working in the other left port.

All these new learnings looked more like a software problem (powersaving?) to me than a hardware problem that can be solved by a new board … but I already had the latest BIOS, and the support then suggested to send me a replacement board, so I thought lets just try that (as the problem also got fixed for others in this thread by a new board, there was at least a chance). And indeed, the problem is completely gone with the new board :partying_face:

So maybe these new findings help somebody else when debugging this, or it might help finding the root cause why this even happens at all. And if you can’t get a new board, it shows some possible workarounds to prevent the problem from happening.