Laptop Fan Ramping Up in Sleep

This behavior has been reported elsewhere, but I’m not sure there’s a canonical thread for it. That said, it’s immensely frustrating that Framework hasn’t made any progress on fixing this behavior. It’s impossible to put the laptop into sleep with the charger plugged in (either before or after the attempted sleep), without it immediately getting hot and the fan running at full tilt.

I don’t know if this behavior is similar to what Linux folks are experiencing, but I expect better from a Windows-compatible laptop. I feel like we’ve given Framework enough time to address this issue and I’m rapidly losing my patience.

cc @nrp

What’s the result when you do this?

I was going to say that the 3.06 bios actually turns on the fan if modern standby is doing something while it is supposed to be sleeping. (I mean whoever thought up modern standby just needs to be shot. In what world, when I want my computer to sleep, does that mean, please keep working, draining my battery and heating up the system?)

If your system is using this, then it is because Windows is ignoring your desire to sleep and is instead trying to get background work done. This happens while plugged in because on AC power is a condition of this work being done.

It would be great if Framework would moan to MS as they have more cloat than us. TL;DR: Modern standby is stupid. When a computer is put into standby it should be doing nothing other than conserving power. Full stop.

Fun fact this happens when you leave some programs open doing absolutely nothing. Happened to me when I left DaVinci Resolve open

For anyone still having this problem it seems to happen to me when I switch from better performance to best performance in the battery settings on windows when plugged in.

1 Like

This happens to me too. I’m using “best power efficiency” on battery and “best performance” when plugged in (windows 11).

For me, it only happens when plugged in. The fan will really ramp up as soon as I close the lid, and stay that way for several minutes. Interestingly, if I open the lid again the fan ramps down back to normal.

1 Like

I have this issue too. Sometimes the computer won’t sleep properly when the lid is closed. The bigger annoyance is when the laptop is sleeping, lid closed, and I then plug it in to power which wakes it up kicking the fans on. Occasionally I’ll go back to the computer a couple hours later and it warm because it never went to sleep.

1 Like

Revisiting this thread as I missed it the first time. With AC attached, and putting the laptop to sleep, windows will sometimes exit sleep randomly (sending the modern standby exit event) and start running background tasks like windows update/search indexing which can be pretty processor intensive. There may be some registry tweaks to adjust these settings.

3 Likes

Popping in here too. Something-something “modern standby” (I call it “newspeak standby” because it seems to not be sleeping at all - just a really, deeply flawed low power mode).

I don’t know what it’s doing and I HATE that there seems to be no way to hold whatever-it’s-doing accountable. Task Manager doesn’t update while it’s in Newspeak Standby, so when it wakes, it’s got that stupid look on its face like
(✿◠‿◠) What? I wasn’t doing anything, I swear!

Every time I have the computer asleep and plugged-in, the computer heats up significantly, and the fans start up. I charge my computers solely from solar power using a power bank (just a symbolic gesture really - the solar setup it charges from dumps 80-90% of its daily stored battery energy into the wall because I only need 10-20% of it every day for charging everything I use). Thus, I have granular visibility into how many watts it’s drawing at any given time. It also makes the issue significantly more painful as I can’t rely on charging the laptop while it’s asleep and plugged-in.

Just a moment ago, “asleep”, fan roaring: 49 watts.
Now, typing this with display at 30% brightness: 37 watts.

This is way more than just battery charging, and “every day, every time” is way more often than doing menial occasional tasks like Windows Update or indexing (indexing is complete, Windows is up to date).

When it’s unplugged, it’s better behaved and sleep behavior is satisfactory. When it’s in hibernate, of course it consumes near nothing (though 1.2 watts while fully charged is still… not great - my power bank auto-shuts-off at <1 watt and it never reaches that).

Mine is also a brand new Framework laptop (batch unknown/not shown anywhere, but has the new Tempo audio chip) with a new install of Windows 10 Pro, and it’s exhibited… unusual firmware-like issues, such as not exposing the “network in standby” options, and having a weird hard-defined 20-minute hibernate timer set by default (in another thread, this was pointed out as unusual in my build).

1 Like

I kept the laptop running with Task Manager open and just… watched it for a few minutes. Guess what popped up? This little f*cker:

SurSvc.exe - part of Intel’s driver updater stuff. Isn’t it nice when you go to download some “recommended” stuff from a legitimate company, and it installs a bunch of crap that drags down your system performance as well?

Not sure if that’s what eats up CPU in sleep, but it does seem to pop itself up high (chomping a full core for multiple seconds - definitely not innocent activity) only when you’re “not looking”. Software that runs when you’re “not looking” and doesn’t make its presence known is pretty crappy IMHO.

“Improves system performance” my entire :peach:.

Next culprit:


System itself (drivers) starts chewing up time after a couple idle minutes, then hides as soon as any key/mouse movement is detected. Man, how deep does this mess go?

(BTW, neither of these are updates or indexing :joy::joy::joy:)

1 Like

There is indeed so much running in the background on our systems today, PC’s, Phones, etc.

How much of it is really necessary?

And many times, the name is something like service, with no clue what it is.

1 Like

Most of it benign, idle, taking up memory but not consuming resources. The impact of its existence goes as far as turning some idle 0’s into 1’s and 0’s in memory, which takes no more energy. It can simply passively exist and consume nothing. That’s fine.

Perhaps a better question is rather, how many individual Chromium stacks can be found on any given computer? :joy::joy::joy:

This is true, but the covers aren’t hard to pull back. “Go to service” in Task Manager will reveal if it’s a service, and right-click->“properties” will reveal where it lives and often, who made it / why it’s there.

The tools are there for the most part, but some things hide really well. Having seen that “System” starts spooling up and going crazy during idle, makes me suspect a driver is being poorly-behaved in sleep – as drivers continue to run while the CPU should be sleeping. They’re basically unrestricted. It’s much harder to hold drivers accountable as you have to start digging into threads at that point…

1 Like

@Matt_Falcon , thank you for good hints. Appreciate them, I just wish that average users who just want to get things done didn’t have to worry about those things.

2 Likes

I’ve been struggling with this issue for months now and had resorted to unplugging the laptop before putting it to sleep. By sheer coincidence, I noticed that Windows used 20% of one core when I didn’t move my mouse for a few minutes, and stopped as soon as the mouse moved. Turns out Windows runs its own version of memtest when idle! This is started by 2 scheduled tasks under Microsoft → Windows → MemoryDiagnostic.

Lo and behold, it’s this same task that gets triggered by putting the laptop to sleep. “Modern” standby indeed. Of course it’s a great idea to peg a cpu checking memory when the computer is sleeping, of all things /s. So, try disabling the memtest scheduled tasks.

3 Likes

The solution @Mr_Darcy provided has worked perfectly for me. The two scheduled tasks
are ProcessMemoryDiagnosticEvents and RunFullMemoryDiagnostic. I have also attached a screenshot of the Task Scheduler for reference.

I have no idea how I haven’t seen this in hours of googling about the problem. Trying @Mr_Darcy ’s solution now.

Edit: I’m kind of in shock, this modern standby issue is so pervasive that LTT made a video addressing it and had enough signal boosting to get in touch with Microsoft about the problem, yet disabling two tasks has had my gripes completely go away. Thank you so much, and we need to signal boost this solution as much as possible to see if it works for other laptop models!

1 Like

I know I reported in another thread that @.Mr_Darcy’s solution fixed it, but well, it appears to be back. Not as bad as before, more of a gentle breeze than a jet engine, but still there.

I did briefly catch some other cpu test in the task manger upon waking it. I’ll see if I can pin it down.

Intel system usage report is probably Intel gathering, well, intel of your device and send the report to the CIA

Probably crypto miner

The solution I have used long sense forgetting about this thread has been to disable modern standby (S0) and just use S3. You can do this via registry, powershell or command line (elevated). This stops the system turning on when in standby.

HOWEVER, on Windows 11 by default the wifi card is allowed to be woken by network traffic. This will wake your machine for nothing. You need to go into device manager, find your wifi card under network, and then under properties and the power management tab ensure that this is not enabled.

Doing these two things will give you predictable and usable standby behavior.

If you insist on using SO then you can make it network disconnected mode, which will prevent as many wakes as well. Why you would want to use SO though is beyond me.

3 Likes

We Penguins get the fan-in-supposed-sleep too. It’s really vexing me. EDIT: And there’s no bloomin’ indicator on the outside of the case, so one cannot tell when the thing is asleep (except in that, if the fans come on, you know that it isn’t). Further EDIT: There may well be a solution (of sorts) for, at least, Linux Mint. See here. I am afraid that that will not help you, @obfuscurity (and that’s a great username, by way!).