[TRACKING] High Battery Drain During Suspend

This is not a usable solution for those that want secure boot and luks

luks works fine with TPM autodecrypt with hibernate enabled. As for secure boot, unless you’re actually going through all the steps to enable UKI, it’s probably not doing much for you.

Hmm. I returned my first gen 1165g7 framework for this issue hoping it would be worked out after 1 or 2 gens at most. I was hoping to buy a new hx 370 framework 13. I see it’s been 4 years and this problem still exists? Wow.

1 Like

Don’t expect brand new hardware with unstable/unfinished drivers to perform better than old piece of hardware… And yes Framework still don’t communicate about this issue for monthes and let’s be realist, they might never solve this issue !

I’ll finish what you started.

@Matt_Hartley , It’s been 633 days. Are you continuing to focus on this?

That’s optimistic. This thread is going to outlive the both of us.

3 Likes

Looking at running w11 on 7640 or 125h. Any issues with direct to hibernate or suspend to hibernate? Has anyone tried this on intel and amd or on windows? Linux kernel issue interacting with drivers/bios vs. drivers/bios issue across the board.
Acknowledge, modern suspend sucks and uses too much power.

For Linux Mint 22.x users on a Laptop 13 Ryzen AI:

Besides battery drain issues while suspended, I also didn’t have screen brightness controls working. I saw another thread about installing linux-generic-hwe-24.04 to get onto 6.11 kernel.

After doing that I am able to control screen brightness and my battery isn’t draining dramatically while I’m suspended. I had my lid closed and laptop unplugged for 10 hours and I’m down around 15%. This is not as good as my 4 year old Thinkpad X1 but good enough that I don’t feel like I have to shutdown every time I’m done using the laptop.

I didn’t see anyone else mention that in this thread so figured someone using Linux Mint might find it helpful.

3 Likes

FW13 7640U Fedora 42 KDE - power drain during sleep seems to be worsening. @Matt_Hartley any updates? My optimism is waning.

1 Like

Yeah, honestly really disappointed in their handling of this. My dell XPS which is only 1 gen lower handles suspend with the same OS just fine.

Them pawning it off on Intel is a disingenuous excuse, even if there is potentially some truth to it; hard to give them the benefit of the doubt based on their silence.

Like said above, this is an enthusiast market with people who care about reducing waste, and actively ignoring a fairly critical function of a laptop like suspend without any attempt to remedy it, is inexcusable. They could at least acknowledged it and say ā€œhey, it seems updates can’t fix it, so we released new hardware with the issue fixed, here’s a discount to replace itā€. Like, anything. Nope, nothing. In fact, doesn’t seem they have focused on the issue at all.

This will be the only purchase I make from Framework, and will actively discourage others from purchasing it. I have not been impressed with the support nor quality.

2 Likes

Over the past 2 days or so, I’ve been testing out s2idle behaviour on my now 4-year old system (11th gen, 64GB RAM, 1TB SSD, 4x USB-c card). Not touching S3 / deep as Framework mentioned it’s not supported by Intel.

Long story short, for a 55Wh battery, it’s still draining at 1.1% an hour. It seems like it really hasn’t been addressed yet.

(Wanted to test it out personally, wanted to be optimistic…but reality is still kicking me in the balls)

3 Likes

I ran the amd-s2idle test script on my brand new Ai340 laptop. 1.2% per hour….. It seems like I can’t get power consumption down on this laptop.

Bummer…so there’s no improvement even with the latest FL13. Is this an upstream / platform / standard / s2idle issue in general? Are other brands dealing with similar level of battery drain during s2idle?

I have no idea. I CAN tell you that I’m going to be purchasing two more usb-c cards. Apparently just having the usb-a cards in the laptop add an extra 1.5W of power draw. Yes, I have them in the recommended ports, but still high power consumption. I can’t get the laptop to a ā€œniceā€ power consumption state either. Lowest I have been able to get is 4.5-5W with the screen all the way down, backlight off, and no apps open except a resource monitor. As soon as i open a web browser, I’m hovering around 7-8W, with spikes up to 10-12W. I was really hoping to get a power efficient laptop, but not sure if I will be able to accomplish that with this machine running Fedora.

Yeah, that’s what I did as well…just to keep power consumption to a minimal.

In my case, 11th gen, 64GB RAM, 1TB SSD, screen at 20%, wifi connected on 5GHz, bluetooth on (nothing paired), Manjaro at 2.45W idle. With NixOS, it was idling at 2.8W. With Fedora 42, it was idling at 2.55W. (Just for comparison)

I just installed TLP, removed tuned-ppd, adjusted all the TLP settings, and ran a test with:

- Display brightness turned all the way down

  • Keyboard backlight off
  • Wifi connected 2.4ghz
  • No apps running except resource monitor and syncthing
  • Bluetooth off
  • 32gb Ram, 2TB ssd

4.8W of power consumption. Not sure what to do. I might swap the drive and try fresh installs of Fedora or Bazzite to see if it changes. I have the 2.8k display, maybe that’s why? I would love to be in the 3W range, but I’ve never seen that with the screen on. With the screen off I’m sitting at 2.5W.

Here’s what my overnight suspend looked like. I don’t have concrete numbers and times, but I put the laptop to sleep around 9pm at 64%. I opened it up this morning at 8am (11 hours later), and I’m at 33%. Whopping 31% in 11 hours, 2.8% an hour. I’m going to see if other distros change, and I just ordered two more USB-C cards so I can install those instead of the USB-A cards. It’s very difficult finding information on the new AI laptops, so I’ve been using recommendations for the AMD 7XXX series of laptops, last gen. I just picked up a thermal camera yesterday, maybe that can give me some insight into where the power is going.

That’s weird. I’m seeing roughly 1% per hour (ca 10% over 10 hours during the night.)

AMD 350, single 32 GB RAM.

I slept the laptop overnight. Stopped using it at 90-95% at 5pm, opened it this morning around 9am and was at 83%, so much better performance. I don’t really understand why it’s so inconsistent.

can we change it so that closing the lid just shuts the laptop or something. losing 10% overnight is crazy