Uneven CPU thermals!

The way yours is acting with a new shim with ptm compared to how mine acted with the original shim with ptm really makes me think that it’s an issue with how the two pieces are joined together from factory. I am surprised that my laptop being a more recent build didn’t have a new heatsink design with this issue resolved. Surely they’ve figured this out already? Maybe it’s still just a heatsink lottery?

That makes sense. I haven’t run linux desktop for at least a few years now and I’m just trying to dabble in it again so I’m a bit rusty.

I can say that sensors also do not report correctly in windows when running a benchmark with “realtime” priority, though the benchmark numbers seemed to be better in the end once the system started responding again.

Mine was a Batch 20 shipped in June 24 and i got an Board RMA with a more Recent Board with the added Keyboard Deflection Kit in July. We suspect it is:

  1. the Shim/Heatspreader bonding to the Vaporchamber is Crap and
  2. the Liquid Metal running off.

Those are the Main Culprits that cause the Degradation.

The Shim bonding is an Issue Framework knows of since the Preproduction and Press Units as they Said they had about 20% of the Units Underperforming.

So the Liquid Metal has to be revised, and Coolermaster as the Heatsink Manufacturer has to redo their Heatsink and the QM on the Unit Assembly has to be revised to find Units with higher Heat Resistance.

Oh wow.

12210 in CB R3.

Core 4 locked at 100c, lowest was core 1 at 73c, and peak power 34W.

image

image

image

Thats RMA time, right?

(edit - raised a support ticket)

Thanks for creating the ticket!
The more we report this, the more they’ll need to consider fixing it.

I hope you’ll be lucky and get a good mainboard and heatsink, but I think the issue is still there.

2 Likes

New Record !!! (but not in a good way)


Seeing this thread made me test my PC (I had some doubt because mangohud nearly always reported thermal throttling in game but I didn’t had the courage to test it)

Temps while running cinebench r23 multicore
Framework 16 Batch 1 with GPU module and r7 7840HS
Peak voltage 36 watt

i got an RMA reply from support. new board is coming soon. mine was sitting core 5 100c and core 1 50c. 34w…

For everyones info, the first contact with support is as follows:

1 Like

Its always different Questions, depending which Support Operator is working on your Ticket. I had a 7 Question Form with similiar and other Questions.

This is the list of questions I had when I opened my ticket:

Follow up questions:

1 Like

Round 2

  1. Kindly remove all the expansion cards and try to test again to see if there will be a different result.
  2. Do you feel the air blowing out of the vents?
  3. Could you also please tell us if you are using the dGPU module for this test?
  4. Kindly test on both the dGPU module and expansion bay shell to see if there will be any difference in the result.
1 Like

You have been send to the Person who doesn’t have any Clue. Its the Script they sent out when the Issue is in the wrong Department. But yeah you will get sent to the Right one after some Rounds.

1 Like

What is the best method to find out if your board is from a defective batch? Are there any tests that can be run to identify if its defective? I am not experiencing any issues with thermals with my machine (shipped in September) it is 16 with 7940HS. No fan noise or any indication of overheating. I don’t have the eGPU is that why? I am now paranoid I have no way to determine if my board is defective or not!

I shared a list with some predefined questions from the support.

Try to follow all the steps and write a detailed email when contacting the support team.

2 Likes

You could use Cinebench R23 while keeping an eye on HwInfo64.

Based on the results from Cinebench R23 and the CPU Package power you can determine if there’s an issue or not.

In theory, you should have above 15k points. If that’s under 15k or if the CPU Package power is stuck under 45W then… you have a small problem.

3 Likes

Well as i have proven the Cooling Solution should be able to cool around 60w sustained TDP. The 45W Mentioned by Framework on the Marketplace is their Solution to get away from legal Actions :wink:
Anything below 54w sustained is affected, because thats the Stock Sustained TDP programmed in the Bios… anything below 45w is a big Problem in my Eyes.

2 Likes

So does this mean ah “Huston we have a problem”?

Yup. You’re 100% right on this :slight_smile:

I can sustain with the current solution ( once the laptop is warm ) 53-57W with a start from 64-68W on a good day.

But, even with this, I’d be happy to buy a new heatsink once they’ll fix the problem.

1 Like

I can see that the cold one is at 69.2C while the hot one is at 99.3C.

Just from that, we can say that you have a problem.
There’s also the fact that the CPU Package Power is under 40W.

Can you please confirm that you’re using the framework charger with the laptop settings set to “Best Performance”?

Yes, it is set to Best Performance and I am NOT using the Framework charger. Here’s the picture of my charger. Baseus 240W Smart USB C Charger, 5 Ports PD 3.1 GAN. As you can see from the display my machine is the only device plugged into it the other 4 ports are at zero power draw.