I have noticed significant battery drain whether this laptop is either asleep or off since I bought this laptop and have been trying to fix it for a few months. I’ve checked that it’s capable of deep sleep, and made a point to not use USB-A cards in the rear peripheral ports. I’ve made sure I’m on the most updated version of the BIOS, because from what I gather this was a known issue that was fixed in a semi-recent release of the Intel BIOS firmware.
I’m here to ask if anyone else has encountered something similar and has a solution (as forum posts normally go lol) but also to hopefully see if this is on a roadmap to be fixed on the AMD version of the BIOS firmware. This may be a separate issue, but I am trying to get a sample size greater than one.
This laptop has been my foray into learning Linux and more deeply about computers in general, so my solution may be much more obvious to another set of eyes.
Thank you in advance for anyone kind enough to respond!
I had some fast battery drain on Lubuntu with the AMD board, but that was mostly due to sleep not working so it’d shut down when trying to sleep (was too lazy to fix sleep issues), so my laptop was just locked when not powered down.
After switching over to Fedora version 41 sleeping works great and the battery drains maybe 1-3% while asleep for 8+ hours at work without extra configuration. Not sure if it’s an ubuntu thing or what, but no issues on Fedora and it’s been a much better linux experience than Lubuntu, though that’s to be expected since the distro is a lightweight one.
Although Fedora took a little bit of getting used to, I’ve actually enjoyed it more than any other linux distro so far (tried Mint, Arch, and others as well) and it’s been very stable on the Framework, minus some issues with USB-C functionality with external displays and other USB-C devices on the newer linux kernel.
I’d say give Fedora a try for a less hassle linux experience.
That all being said, there are some other threads reporting similar issues and it is being tracked as a known issue in the first thread:
I also found this article that speaks to how to change some settings to get it working properly as well, but as I haven’t had the issue I haven’t tried it.
I haven’t read fully through those linked posts yet, though I likely will when I have time, but I can say I do have this issue with drain when laptop is entirely shut down. I’m running Fedora Silverblue with latest kernel on BIOS 3.05.
I did have suspend issues (still have too much drain), but they are much better after some tweaks. The thing that irks me right now is that the AMD board completely drains the battery after 3 or 4 days of being completely shut down. This obviously is crazy, almost like it’s suspended instead of off. But the issue is consistent every time. It also will, after a few days of being drained lose the BIOS/EFI settings. I may email support about this because it is a pretty significant issue for me.
It seems some others still have the issue too, so you’re not alone, but if you do find a solution, I’d love to know! I own 3 different Framework laptops for my company (I really love the mission and execution in general of Framework) and 2 of them have unusual battery drain issues (after 20 years of owning or working on scores of laptops from different companies I’ve never seen these issues), but the Framework AMD is the worst so far with the battery drain.
As such there are some kernel options modified for hardening purposes and some processes that aren’t enabled by default, but the it is Silverblue 41. Definitely let me know if you are finding a different experience, because I love to know if I could change some option that might ameliorate my situation
Your numbers sound about right from my experience too when my suspend works correctly. I wonder if the times when the suspend doesnt might be related to the IRQ1 issue patched in the latest BIOS beta: Framework Laptop 13 Ryzen 7040 BIOS 3.06 Release BETA
Still those wouldn’t fix the power drain while entirely shut down that I get frequently, But I will test further this week and see what happens when I record the exact numbers.
Update on the battery drain issue! I have been testing different things over the past month: I tried removing the expansion cards, booting different Operating Systems, updating the BIOS, and recording exact numbers and times to get a better picture of the issue.
Support recently suggested that I reset the BIOS to default settings , since they say Linux users tend to modify their BIOS settings (probably true in general, and true in my case) so I did that and kept testing. For whatever reason, resetting the BIOS to factory defaults has fixed the battery drain issue completely for the past 2 weeks. I asked support what setting they think may be responsible for that but they don’t have any ideas at the moment.
I am very happy that a solution is found, though I don’t think any of the settings should cause that kind of behavior. I’m a bit concerned about changing settings now (like the GPU RAM for example), but perhaps I will try changing one at a time and seeing if the drain returns. Perhaps anyone else with this issue can try a similar path!
I’ve been following this thread since deciding to order. I have had my Framework 13 Ryzen 7040 with Linux Mint 22 aboard since Nov. 6. My battery drains at about 1%/hour (just a little more than 1% on average) while on standby, nothing connected to expansion ports. It does not seem to drain the battery when I shut down completely. Battery goes from 100% to 30% in about 4-5 hours browsing (not usually video intensive, just reading) or in my email/calendar program. I expected the battery to last longer under these circumstances (thought I saw 7+ hours touted watching videos when researching). I have not been leaving the power plugged in, only using it to “top up” when batt goes to ~30% or lower. I am interested in hearing more on other’s experiences. Thanks.
I definitely would love to have these sorts of things resolved too. Your experience at least sounds reasonable (and probably about what mine is), although not performing how I would expect either. I may start tracking battery life more closely to get real numbers now that my laptop doesn’t die while shut down.
I didn’t try to count but a quick search on the forums shows at least dozens of threads about battery life issues for the AMD chipset, as well as other chipsets, and most remain unresolved. Like many others, I love Framework for what they have already done, but the power issues are a real sore spot, and something that is a real caveat when I make recommendations for new users.
One thing worth noting as a long time Linux user is that different kernels and different distributions can have a real effect on things like power. You might consider downloading Fedora Workstation to a USB drive and running it Live for a few days to see if you have the same power issues on that distribution (one of the officially supported ones). Linux Mint is hugely popular and version 22 has a reasonably recent kernel, but I believe technically it is not “officially supported” by Framework so might be worth checking to see how Fedora runs Live.
Oh yes, that’s a fact! To expand on my original post, I did make a change within BIOS: stopped Secure boot based on several sources’ opinions that it’s actually pretty useless on home machines, and some changes to GRUB so I can see and exercise options on boot. So far, the laptop is behaving as I wanted, except for the battery drain. Frankly, so far, that’s a minimally annoying curiosity rather than a real issue to me.
One other potentially interesting observation: while I was installing LM 22 etc., and transferring files wholesale, I left the Ethernet expansion card in slot #3 with the ethernet cable attached with the laptop shut down (yes, “powered off”) over night, battery at full charge before shutdown was initiated. The battery was dead the next morning. The expansion card showed network activity (blinking) after shutdown, so I guess that’s what was using the power.
I also see increased power drain during sleep on a FW13 Ryzen 5 7040. As Matt Hartley suggested, I ran this script to check on s2idle:
As root: python amd_s2idle.py
Using 10 seconds / cycle, 10 seconds in between and 5 cycles in total
This results in the first cycle looking good, while all others note: Did not reach hardware sleep state which might explain the increased battery drain. Can someone reproduce?
Thanks for posting that script here! I went through that with my laptop and on SecureBlue, I needed to enable the thunderbolt module kernel option (it’s restricted by default in the “distribution”). I also needed to use run0 instead of sudo as that is now changed in SecureBlue
run0 python amd_s2idle.py
Once I did that, though, the script is working for me correctly on a cycle of 10 seconds, 4 seconds, 5 cycles and reports that my machine is entering hardware sleep state correctly.
I will test out real world sleep performance this week and report back
