Add to this - here are the pics of the LM after removal of the heat sink - of note was the piece that was completely outside the “pad” area
4 Likes
Brand new board installed, presumably with the existing LM solution, will be closely monitoring for any changes on a weekly basis, intial scores started around 13000.
Monitoring thermals with HWINFO during testing showed significantly reduced core temp delta going from 20+ degree differential down to under 7+.
3 Likes
Has anyone here requested a kit with the request form?
I can’t input my order number, it says R+ 9 digits but the input field only allows for 8?
I requested it and no issues. I put in this format-> R123456789 and no issues.
Redid it and now it worked 
2 Likes
I still haven’t received my keyboard deflection kit that I submitted for, so I’m not sure if/when these kits will arrive. I did submit my request for the LM replacement though.
1 Like
People have already received the keyboard deflection kit. If you have any questions about it, you should reach out to support.
1 Like
I received, two, one for my original board, one for my replacement. I did ask for the 2nd. I got it no questions asked though, just gave them the ticket id for the replacement.
1 Like
I have a feeling that the shim is more important than the thermal interface material. The thermal interface material can help getting to 45W, but not 80W. I’m rather hesistant.
1 Like
Hello Nirav,
With the move away from LM towards the PTM solution, will Coolermaster make modifications for vapor chamber? I assumed the solder on shim was to create a barrier to be protection against the different types of reactions that could happen with copper and LM. The shim wouldn’t be needed and thus that process could be skipped?
1 Like
Switching from LM to PTM on mine, with no shim change, resulted in a sustained 57w when running R23.
4 Likes
hm.
Well, I have some improved paste compared to my industrial one (seem to tend to boil off), so let’s see.
Edit: 42W doing Cinebench.
Not great.
1 Like
Does anyone know if this issue has been resolved for current buyers? Or where we would be able to find out in future? I’m interested in buying but would like to hold back until its been sorted. Thanks 
1 Like
What are your CPU temps like with the thermal pad out of curiosity? (idle & under load)
1 Like
idle after a Coldboot its right now Sitting around 34C, I hit 98C max after a Coldboot with 75W TDP and falling to 54w (92C) after 15s as i modified PPT Fast Time to 15s and the System falls back to 42C running the Browser(Idle).
After the System heated up once i still manage 70w for for the Peak TDP within the 98C
Note: I reduced the TJMax to 98C from 100C with UMAF
1 Like
yippee!!!
3 Likes
2 days in with the PTM from Thermal Grizzly:
Things to note, I now do not hit 100c at all during the process. So I would argue I am no longer thermally limited.
I get a sustained 57w - the 7840HS is spec’d at 54w
With LM, my initial runs on R23 were 14700 and after a few months (new board), it was down to the mid 13s.
7 Likes
If the newest FW16 coming off the production line are PTM instead of LM.
Is there anyone here who has a recent FW16 who could do either a cinebench if on windows, or something like a Linux kernel compile and amdgpu_top and see if the 8 core temps are even or not.
On linux, compiling the kernel I get core temps in C:
79,76,85,83,98,89,99,90
If that is core 0 to 7 from left to right.
I have a 76 and a 99, making the core temp range being 23C, and I get the power reported as hovering about 41W.
As you can see core 0,1 are the coolest, with core 4 and 6 being the hottest.
I would be interested to see the 8 core temps from someone who has a more recent FW16.
Mine was Batch 16.
1 Like
I can confirm that all of my testing was conducted under the “High Performance” power profile in Windows 11 with 180-watt USB-C power brick connected to the Framework 16. The difference between 13391 and 16400+ is quite a large delta. Also, my core temps are all over the place with one core only hitting 79.8C while another is at 100.1C which seems a bit extreme since on my desktop CPU all the cores stay within a few degrees of each other at full load. But, I wasn’t sure if AMD was different since all my desktop PC’s are Intel based.
I can also confirm that the fans weren’t obstructed on the back of the unit, and it was sitting on a hard wood surface. So, this should be the best-case scenario for thermals. I also noticed some people in this thread are getting much higher scores than me with the lesser 7840HS processor which I found to be a bit odd, but it makes sense if my processor is having to heavily thermal throttle because it’s hitting an upper bound. I’m also a little surprised that the cores that are running cooler like Core 0 aren’t boosting up more since they have more thermal headroom, or do they purposefully not do that if one core hits 100C?
This Framework 16 laptop has performed so well that I had never considered that there was more performance on the table. But, after looking at all my results and others in the thread it’s pretty obvious that I’m suffering from the same issue. Most of what I do is GPU bound and I don’t have experience with AMD CPU’s before this so I really didn’t have a point of reference until I saw this thread to show me what true performance I should expect. However, I will note that the benchmark’s I ran did show that I was underperforming the average for this CPU but I just attributed that to it being a laptop CPU versus desktop CPU but apparently that isn’t the case.
Does Framework have an official solution for this yet? I suspect the sanding option is more of a community solution rather than an official one? I would prefer to wait for an official solution from Framework 16 since I want to echo the true end user experience if they purchase a Framework 16 in the future after this is resolved in any of my reviews.
2 Likes