Uneven CPU thermals!

I bought it online it’ll arrive tomorrow

Thermal conductivity numbers given by paste manufacturers are almost useless, at least across different manufacturers.

PTM definitely outperforms paste and approaches if not matches liquid metal, didn’t really belie it before I tried either.






results are in and it looks good! The CPU temps have also never been this low before, they’d usually go straight to 100c but it’s staying under 95c on the hottest core!

2 Likes

Great to see! Really Great Result with alot less manual Labor as my Workaround!
If my Spill Guard wasn’t completly torn i would have considered another LM too.

Okay your Heatsink looks different than mine. Mine has its QR Code etched into the Metal.

So my Point all around is proven. Framework has to get a Fix for the Cooling in the FW16. It’s sadly lacking alot of Performance.

1 Like

yeah I had to agree with you on that

I’ve done some more benchmarking with 3dmark and the CPU clocks seem really unstable


The funny thing is that the best score 3dmark shows was done by me on my current motherboard.

My CPU score lines up with the other scores, so it’s my GPU that has the low scores.

Fixed it! turns out it was an issue with my power plan.

1 Like

Thanks everyone for all the help and support! In the end, my laptop has a lot better CPU thermals so I’m kinda happy this happened lol.

3 Likes

I am glad we could help. And congrats for trying something risky and succeeding :wink:
But be advised, that the liquid metal you used can be problematic when carried vertical, it is easily able to creep out. Be careful please. Yes there is a spill guard, but better save than sorry.

Yeah I made sure I didn’t add too much Liquid Metal. I saw a couple youtube videos on how to apply it on laptops and they didn’t add much of a spill guard. I’ll make sure not to drop it or anything but I think that with surface tension and the spill guard I’ll be fine.

I just did a Cinebench R23 10 minute run on my machine (R7 7840HS w/ GPU) and got a score of 13595 with a temp delta of around 12C with one core being at a constant 100C and another between 88C and 89C. Seeing as my score is way, way lower than what you got on your old board I’m considering trying to RMA my board as well.

2 Likes

Just created a ticket for this same thing as well.

If the second board is bad… maybe Ill go the PTM7950 route… but for now. Id rather have Framework fix it and give me the product I paid for.

Love the company, love the mission, love this latop - but… my CPU is being more of a bottleneck than it should be right now.

1 Like

Can you remove the heatsink without removing the motherboard?

I looked at the picture of the heatsink and I think you can - each of the screw holes are visible while in the frame.

1 Like

Yes you can, but you have to remove the GPU/Expansion Bay, to be able to remove the black top Bracket with the Power Button. Its 5 Captive Screws and the top left and right Mainboard Screw, that hold the Heatsink.

1 Like

That’s perfect, thanks!
I’m currently making a plan in case the new board is not better.

1 Like

Reminder that the mainboard screws are not captive!

Cross-post from the batch 20 thread - pretty much as expected.

Big shout out to the support team, however. They were straight forward in the request of what they wanted and didn’t run me through a script. @Destroya - big shout out to the support team - I am not sure how much exposure they have to the forums - but it definitely feels like the Company is aware of these issues.

Just some info for everyone else. FW support reached out with the request to run R23 and provide thermals.

I did so - and, the results were pretty much as everyone else said:

As seen, a temp delta of about 15 degrees, a really low score, and Core 4 was at 100c pretty much the whole time.

Biggest issue is only getting about 36 watts of package power.

4 Likes