“refined” 
I’ll order a new heatsink in a couple of weeks.
Do you remember if you had to reduce the size from 20mm x 20mm?
I just had a few delivered today (20mm x 20mm x 0.8mm ) and I want to get it ready for when I buy the new heatsink 
I just sanded the edges down of the shim, as I think they are press cut so you have upstanding edges on the cuts. There is no need to reduce the size though, if placed correctly it doesn’t interfere with anything.
2 Likes
That’s perfect, thanks!
Last time when I checked it was a bit too close to the capacitor.
I’ll make sure to sand the edges.
@Alex_Uta nope didn’t resize the Shim, only lapped it and deburred the edges.
I swapped from Liquid Metal back to PTM7950 with my soldered Shim and i am back to the Same Numbers as with the PTM Sandwich.
Thanks!
I guess the benefit of the soldered shim is that it won’t move at all.
Do you have some numbers you can share?
How many PTS do you get with everything closed and, what’s the sustained power?
Thanks folks! I’ve soldered the shim and put some PTM yesterday.
I’m getting 55W with 8-10C delta between cores; power spikes at 65W initially and stabilizes over half a minute as heatsink heats up. Room temperature is around 20C. On LM I was getting 32W and 30C delta, with almost no initial spike.
Sum of frequency over all CPUs increased from 42Ghz to 52Ghz, that’s 24% improvement, not that much. More importantly if power is restricted to 30W, CPU stays around 65C with much, much less fan noise.
2 take aways:
- vapor chamber copper is very thin and easy to bend, I’ve got some visible dents from removing the shim with a knife, and from quite light pressure applied when removed solder with sandpaper (now I’d use grain 800-1000 and a really light touch),
- when I was installing the heatsink, I got power button cable under the pipe and got it pressed to the chasis. Laptop powered on, and there was a tiny bit of magic smoke in the area of power cord connector. The button still works, though, not yet sure what was lost.
3 Likes
That doesn’t sound good.
If you need to do this again, please have a look at Enable Battery Disconnect.
1 Like
Well 24% IMprovement over something that should have been from factory is quiete impressive 
Good Job!
I am now also pretty pleased and as i reverted from Liquid Metal to PTM i lost about 5W sustained TDP from 63W to 58W. But thats okay for playing save on the possible Shortage Front.
2 Likes
I put the dGPU upgrade in yesterday and reran cinebench, thinking bigger/faster fans might make an improvment to my score. Instead, I lost 400 points and dropped to 13020 
I don’t know for sure that the drop is a result of the fan change.
What’s your CPU temperature before and after the dGPU change?
Well 400 Points can be one extra app running in the background. As everyone with below 15k you are definitly affected by the lm run off 
I have only the dgpu not the Expansion Bay, so i have no reference. Is the System louder with the dgpu?
100c constant, same as the shell fans.
This was the first thing I noticed, the noise level different is drastic. I went from flipping my device around to see if the fans were spinning to being able to hear each individual step increase in the fan speed. The dGPU fans are very loud in comparison to the shell fans.
I swapped the LM with some noctua paste and now running 51W in R23.
1 Like
what noctua product?
NT-H2 Put on probably a bigger drop that I was supposed to but it lowered my temps considerably when gaming. I think, at least in WoW it used to hit 80 degrees, now it sits around 65.
1 Like
Hi! My friend recommended FW after I mentioned I needed a new laptop and I was instantly sold due to its modularity and upgradability.
I’ve been trying to do as much research as possible and stumbled across this heating issue. I’ve read through this thread and just wanted to get some advice before finally placing my order.
(I plan to get the FW16 7840 with Windows and without the dGPU.)
Please forgive my ignorance wth these questions!
The website says current production uses Honeywell PTM7958, but has this fixed the performance degradation? Is anybody still experiencing issues? What kind of power readings and benchmark scores are people getting and what readings/scores would be acceptable?
Would it still be a good idea to “repair” the heatsink and replace the shim for the uneven thermals? And by soldering or by PTM sandwich? I’ve got limited experience with soldering (replacing blown capacitors in TVs at home). Are either of these methods something I could do without majorly screwing up?
I’d like to have this laptop for a very long time and given the price I want it to keep performing well throughout and hopefully without too much fan noise (my current laptop is very noisy!)
Any advice is welcome and much appreciated!
Cool, I have some of that, where I used it with my AMD 5950X, it’s good stuff
With the AMD 5950X, the difference between NT-H1 and NT-H2 were noticeable for me, where the maximum temperature were eventually the same but significantly longer for NT-H2 to get to the maximum and when the load dropped, the temperature decrease rate was noticeably higher than NT-H1
======== offtopic
nearly 10 months now since I’ve had my FW16, never used my AMD 5950X
FW have stated they are working on a fix if the described problem in this thread occurs. I have no idea if they are implementing the fix already in the production, but that could be where they are testing it right now.
So either it will ship without this problem (in the future). There is always the chance you won’t suffer this issue with your FW16 (individual production variance) and if it does, FW got a fix in the works. If you don’t mind tinkering you can always fix it yourself as many including myself have done in this thread.
If this is a concern for you, you know what to do.
1 Like
Same here. I used to do everything on my desktop that had a 5950X, and since getting my Framework 16 I’ve turned my desktop from my daily use machine to a NAS.