I ran NixOS 24.11 mainly, but recently tried Ubuntu 24.04 and Fedora 42 on my FW16 (batch 19) with 7940 and 7700 dGPU.
I used the laptop for almost a year now, running rock solid for the first 6 months until i installed Bios 3.05. At first there were performance issues with battery draining on AC, then i installed the bios update and the system started to freeze then reset and sometimes hanging, forcing me to hard reset. I suspected a driver issue, so i tried with updated mesa drivers and the amd vulkan drivers. The freezes became more and more frequent in the recent days sometimes even freezing while booting. It also sometimes froze while i was in the uefi firmware settings. Or the keyboard wouldn’t work in there. However i could at least boot into my OS somehow. After trying out different distros, all showing the same freezes, i completely wiped all my storage and set up nixos again, and it froze again. So i created a ticket, just a few hours ago. But even worse, after that it would not even boot from usb anymore, freezing and hanging everytime in the process, same for booting from SSD. Now it wont even let me pick a boot device anymore (F12 right after power on). It will boot to some extent after it is unplugged from AC for a longer period but that is rather rare in comparison. When it hangs, most of the time the fans start to spin and the device gets hot, i am hesitant to wait that out.
Is there any way to recover at this point? I don’t want to void my warranty with custom EC images, but at the same time i really need the device for UE5 development and i dont have a spare computer for that. The project i am working on is really important and that all happens at a really bad time for me. Which is why tried so eager to get it running still. I will give it some rest now, but any suggestions are very welcome.
You should contact support as soon as possible, even if you are not sure of the issue or its cause. Framework Support takes some time to respond, so if you file one now, it’ll let you both wait for a response and get time to diagnose the problem.
You can first try the BIOS reset via chassis intrusion button. Or the top lid button.
Plug it in with the midplate removed. The LEDs should blink red. Then press the button. Then release. Then press.
Press this button at least 10x. But you must press it slowly. So push down, blink, release, blink; push down, blink …
Then attempt boot. It’s ok to not have stuff like the midplate installed, we want to make sure the core system functions properly. Maybe even remove the expansion bay (completely). and the modules as well. The system won’t be too mad if the expansion bay (the fans) don’t exist.
Thanks for the suggestions. I swapped both dimms and then tried each slot with just one module, with the midplate removed. It did not get to post, with red-blue-off led blinking pattern. With both RAM modules installed, plus midplate and keyboard, it got to post once with leds indicating it got to 0x8B PEI_GO_TO_DXE_CORE
@Xavier_Jiang using the intrusion detection switch made it get past the bootloader, like it does when ac disconnected for 90 seconds, hanging in the os boot process again, complaining about xhci not found/dead, but it did not show any post led action (or i missed it every time i tried) After that the screen stays black again, when turning it on again. i had everything removed including the expansion shell, except one usb-c expansion card to connect the charger. I will have to try again later
I am the author of one of the custom EC images. I am not suggesting you try it, but just that I now understand a lot more about what the EC does.
First thing, the problem you are seeing is a hardware issue, and not a software one. I have concluded this from your mention of it happening while in the BIOS.
This is why a don’t want you to try my custom EC image, because it will not help at all.
The next step is then to try and diagnose which device is faulty.
Try removing all the following:
Battery
NVME SSD
Wifi card.
Leave only one RAM chip installed. (It won’t reach BIOS without a RAM chip)
If it still FTH/R, then my best guess is the main-board is faulty.
Hopefully FW support will agree.
My guess at the cause is that one of the Power rails has gone bad, failed capacitor, something like that.
hey you are getting somewhere.
I would try to make sure the SSD and stuff actually have a working OS (among other things, like if the SSD work at all).
That is just .. not a thing. BIOS will do its thing regardless of any of that. At worst you don’t boot into a OS, or your OS gets confused by the wifi card.
There’s two things to check for:
Does it POST? (go to BIOS). Is everything attached detected (no defective connector/wires on board)?
Does it BOOT to OS?
If it doesn’t POST, or go to BIOS, that’s a firmware/Framework laptop issue. Contact support or whatever. If it doesn’t BOOT, that’s on whoever that did the system install.
Have you ever tried plugging a be200 into an amd board XD. Pretty much anything can mess with the POST process or crash it, especially if it’s on the pcie bus or talking to the ec.
Okay, i removed everything you said plus the expansion bay and also tried both my dimms separately, but still just darkness on my screen. It did turn on the display on a few attempts but it did not show anything. I had not missed the post leds before, as it turns out my device needs around 60 seconds to output the post led sequence. It showed 0x0B each time, which i cant even find in the insydeH20 postcode sheet. Pushing the power button again did not turn it off, i had to hold it for 10 seconds each time again. Recorded three attempts with my phone. This behavior remains now after fully assembling it again. Probably irrelevant, but waiting for some time will cause the fans to spin up and then down again, so it is still doing something… anyway, water under the bridge.
I was wondering if the hard resets were somehow affecting the EC firmware as well, but now i don’t think i should attempt any flashing
Also, could it be that the 240W charger somehow caused this to appear?
I already created a ticket before it got worse. That was three days ago and i have not heard anything back. My description in there was quite lackluster as well, should i create a new ticket, since the situation has changed? I have not yet got a ticket number to add these new developments.
That may actually be a good thing, sounds like it could actually be at least trying to do memory training which would mean the cpu isn’t entirely dead.