My DIY edition running Win10 is getting stuck clocked at .39Ghz. This happens even when plugged in. After a reboot it runs at full speed for a minute or two, then reverts to .39Ghz.
I’m following the instructions on this thread and this did not fix it.
powercfg /a output (note: after disabling modern standby):
C:\Windows\system32>powercfg /a
The following sleep states are available on this system:
Standby (S3)
Hibernate
Fast Startup
The following sleep states are not available on this system:
Standby (S1)
The system firmware does not support this standby state.
Standby (S2)
The system firmware does not support this standby state.
Standby (S0 Low Power Idle)
The system firmware does not support this standby state.
Hybrid Sleep
The hypervisor does not support this standby state.
The SleepStudy (what a great tool): https://drive.google.com/file/d/1D2KxcyyVQk0KHwCbqESpTkp1qOmhZPkG/view?usp=sharing
Just to cover the bases, do you have the driver bundle from here installed? Windows 10 Installation on the Framework Laptop DIY Edition - Framework Guides
Yes, I have the driver bundle installed.
I don’t know what frequency the laptop idles at (CPUs can idle far belong the base clock speed), but I’ll assume that’s not what’s happening here. (Edit: irrelevant)
I wonder if something is very wrong with the cooling system. Could you check out what range of CPU core temps you’re getting with a program like Open Hardware Monitor?
It’s also been mentioned that the CPU can throttle to as low as 200 MHz if too many high-drain devices are plugged in to the laptop. You don’t have anything like that, right?
400 MHz means it is potentially hitting a PROCHOT, which means either the CPU hit 105C (which shouldn’t normally occur), or something went wrong in the firmware state. Do you have any devices plugged into laptop when it is entering this state?
@nrp i figured it out. When I first setup I noticed it couldn’t charge via USB-C on the back-left port, and plugging my USB hub in there raised some “this device isn’t working plug it in elsewhere” warning. I ended up moving an HDMI expansion card to that back-left port and have not used it.
If I remove that HDMI expansion card the clock returns to normal, and when I put it back the clock eventually drops to .39Ghz.
Putting a USB-C card in that spot does not produce the clock speed problem, maybe those are pass-through and the HDMI card isn’t?
With a USB-C expansion card and a hub plugged in to it, the back left port does not cause this problem, but it also doesn’t work.
That is very interesting and problematic. USB-C is indeed passthrough while HDMI uses USB-PD and DP Alt Mode. It seems likely this will need a mainboard swap, but I will let @Kieran_Levin provide thoughts on this as well.
I am also encountering this problem where it drops to ~400 MHz. I am running dual 4Ks connected with the DP expansion cards on the left side of the laptop. On the right side I have the USB-C for charging + USB-A for a hub which has my keyboard, mouse, camera and microphone.
I have the machine loaded up with 32 GB RAM and am using VSCode + MS Edge for web app development. All of my heaving lifting is done on a remote docker container so I don’t think I am over-taxing the machine. I am basically using VSCode as a editor frontend and then of course the browser. No compilation is done on the machine.
This normally runs okay but when I do a Zoom video + screen share on one of the 4Ks, the machine regularly goes to 100C and then throttles down to 400 MHz. These Zoom screenshare sessions are typically 8 hours a day as pair programming is the normal way of business for the place I work.
Am I driving the machine too hard or is there a legitimate thermal issue? I did see talk about a different thermal paste. Would that remedy this issue? Is there some other thing that could improve this?
I really love the laptop and the concept here, but it’s real difficult when the CPU clocks way down and the whole system drags for a min or two.
Thanks!
Hi @Jordan yes, that one sounds like it could be the thermal paste issue. We’re getting in a batch of thermal paste syringes with the new paste that we can send out. Send an email over to the support team with your system serial and shipping address, and we’ll mail you one.
@Adam_Hupp since you are only experiencing this on one port this does sound like a potential hardware issue. Two additional things to try are.
@nrp I did contact support and they put in a order for me for new thermal paste. However, it has been a week or so and it hasn’t shipped out, so I ended up buying some GELID GC-EXTREME thermal compound and made sure to coat the CPU/die completely. It did definitely lower the temperatures, maybe on average 15-20°C but it is still overheating and now even throttling to 199 MHz. This happens very regularly throughout the day when I am using it.
What can I do? It makes my entire workflow screeches to a halt while it cools off.
Forgive me if I am repeating something that was already suggested.
Most of the threads I have seen referring to running at these reduced clock speeds are caused by an issue where the foil cover for one of the expansion bays is contacting something it shouldn’t and causing a short that shouldn’t exist, and the throttling down of the system is a self-protection method.
If it is indeed that issue, please see this forum post in the “An Expansion Card bay isn’t working” section: Known issues on early Framework Laptops
The fix is described here: https://knowledgebase.frame.work/en_us/one-port-on-my-laptop-does-not-function-correctly-r1mqMnTet
Hi @Jordan sorry about the delay. We are just getting the thermal paste through our inventory process for shipping. With the updated thermal paste you applied, are you seeing temperatures hit 100C? If not, it’s possible the throttling is being caused by something intermittent that is separate from the original thermal paste issue or thermals generally.
@nrp It is reaching 95°C. I looked at the An Expansion Card bay on my laptop does not function correctly (frame.work) guide and opened up the machine. 3 of the 4 seem to have this problem:
The one that seems okay:
The incorrectly installed ones:
The big problem now is that when I tried to move/adjust the last sticker, it ripped in half:
What can I do with that? Should I try to fix the other ones that are wrong/not yet ripped?
Can it function without these?
Thanks
@Jordan We don’t suggest adjusting the EMI stickers unless the port has a functional problem such as expansion cards not recognized or not charging. You can continue to use your system without the sticker and there should not be any difference in performance. If you want to, you can just place some tape over the top of the sticker to hold it together and reapply it.
@Adam_Hupp
Since you have a consistent repro case could you try downloading and installing the intel power gadget
And then running the following command:
“C:\Program Files\Intel\Power Gadget 3.6\PowerLog3.0.exe” -resolution 100 -duration 600 -file cpulog.csv
Which will log your cpu speed, temperatures, and throttling for about 10 minutes.
If you can start this as soon as the system boots, and capture a log of the system performing normally and then transitioning to the low frequency state it would be helpful for us.
If you can share the csv with us we can take a look and see if we can find the source of your throttling.
@Kieran_Levin it was reproducible for a day or two, the stopped. Possibly because I’m not using the pass-through power on my janky USB hub anymore. I installed that tool and will grab a log if it starts happening again.
So, I’ve noticed a slight cpu clock issue with mine. Occasionally, it’ll decide to run at .39 GHz while under load for 15-20 seconds, before slowly getting back up to normal speeds. I’m fairly sure it isn’t temp related as the cpu has run for extensive periods at 100% usage around 3.5 GHz and not had any issues, but, (as my most recent example) I was running around 25-30% utilization when it decided to clock down to .39 GHz for the aforementioned 15-20 seconds before speeding up again.
Is there anyway to find out what’s happening (like running some sort of terminal while I’m using the laptop to see if it throws an error code?). Additionally, and this could be the cause, I was running it with an egpu at the time. (should I follow the advice from this previous post?) ↓
