AMD Framework 13 unreasonable battery loss when suspended

[Samuel_Ainsworth]

Which Linux distro are you using? NixOS

Which release version? 25.05.811497.33c6dca0c0cb (Warbler)

Which kernel are you using? 6.12.53

Which BIOS version are you using? 03.16

Which Framework Laptop 13 model are you using? AMD Ryzen™ 7040 Series

I’m finding battery life of the FW13 to be rough in general, but especially when suspended. Whereas my macbook can go weeks with the lid closed, the FW13 drains the battery as fast as faster as when I’m using it. Why could this be? What are the steps for debugging this?

I have run @Mario_Limonciello ‘s handy amd-debug-tools s2idle and here are the results: amd-s2idle-report-2025-11-15.md · GitHub . I have read them, but admittedly I am not sure entirely what I am looking for.

Has anyone else successfully gotten the suspend battery usage under control? How does one go about fixing this?

[Samuel_Ainsworth]

Additionally: this is a Ryzen 5 7640U chip and I have 4 official FW USB-C adapters attached. There are no other peripherals.

[Mario_Limonciello]

I think you’re missing some systemd libraries that help get the journal. Dmesg seems to have not worked to capture that log.

Anyways - run a longer cycle (like hours over night) and look at what the tool finds in the aggregate and others can tell you if they’re reasonable numbers.

If you haven’t already, make sure you’re not putting cards in the wrong slots leading to high power consumption (framework has an infographic)

If the slots are correct and people agree your consumption is what’s expected I suggest looking into enabling suspend-then-hibernate. It’s a systemd feature I upstreamed a few years ago that will put the system into hibernate after being in suspend for a configurable number of hours.

[Samuel_Ainsworth]

@Mario_Limonciello

I think you’re missing some systemd libraries that help get the journal. Dmesg seems to have not worked to capture that log.

Hmm, ooc what libraries could this be? I’m not familiar with dmesg and its dependencies.

@Mario_Limonciello

Anyways - run a longer cycle (like hours over night) and look at what the tool finds in the aggregate and others can tell you if they’re reasonable numbers.

Roger that, will do

@Mario_Limonciello

If you haven’t already, make sure you’re not putting cards in the wrong slots leading to high power consumption (framework has an infographic)

I found the infographic. I am only using USB-C cards, and IIUC they are cleared to work in every slot. I can try reproducing without any cards, I would expect it should work the same.

@Mario_Limonciello

If the slots are correct and people agree your consumption is what’s expected I suggest looking into enabling suspend-then-hibernate. It’s a systemd feature I upstreamed a few years ago that will put the system into hibernate after being in suspend for a configurable number of hours.

Oh cool! I’ve been using suspend-then-hibernate… didn’t realize that you contributed it. Huge thank you! It’s been great, and I use it all the time now.