4% accuracy - but faster! 
3 Likes
This is my personal review and if anyone is on the fence about getting this laptop…please read.
My spec:
Ryzen™ 9 7940HS - Radeon™ RX 7700S
32GB or ddr5 ram 3600mhz
2tb WD sn8000x
After receiving my laptop, I loved it and was so excited to use it! I love the modular aspect of the laptop and what the company stands for, BUT it ends there.
The light task you can do without any real problems, but as you start to do something a little more work intensive the fans would blast sooooooo high/loud.
I didn’t mind it as I would put on airpods to drown the noise. Then a month later the motherboard died on me. No matter what drive from any brand I put in, it would just not boot. Took me about a whole month to get this fixed and replaced via support.
Now that I got it replaced, every time I boot up my laptop the fans would ramp up and my CPU would get up to 90-100 degrees Celsius while booting/idling. I temporarily fixed it after reinstalling Windows for the 5th time. Again, I am faced with the same issue when I run a game (not an intensive game. I tried bluestacks with 1gb for both memory and CPU, league, and even super auto pets…) or my virtual machine. I am reaching temps up to 90-103 degrees Celcius.
If I had known this would be the case, I would’ve held back on getting the Framework 16 and just invested in a gaming laptop at this point. I would hold off on getting this Framework until its much more refined. I get that this is the first product they released with a GPU, but I feel like its not there yet. On a side note, I hate AMD software…almost 70% of the time their Radeon software to tune and change the settings for your laptop does not work. I have to uninstall and reinstall whenever it decides not to work. It’s just such a bad experience and I wish I could just return this and get a different laptop, but that’s just what you have to deal with for being the guinea pigs…
Hope this honest review helps someone. Again, this is just my personal experience.
2 Likes
I got a Mainboard Replacement because Hitting 100C at an instant and being more than1500 Points off the Median 7840HS in CBR23 and having a Tempdelta between Cores of more than 10C under full Load is not right for the 7840HS.
Well certainly there is some Problems with either the Cooling System or the “Revolutionary Liquid Metal” as there are more and more Threads and RMA’s for the Temperatureproblem
With my new Board there was still a Thermaldelta and it didn’t feel quite right, but it was better than with the First Board.
I decided to remove the Heatsink and take a look. The Heatsink looks, ehm funny, it has hexagonal indents. I Suspect them to be the Reason of the High Tempdelta on the FW16. As those Bumbs has to be filled by Thermal Interface Material. Possibly those are there to help with the surface Tension to keep the LM in Place, But they create a very big Gap for Heat Transfer.
Yes i know, the LM Pad is still not orderable. I Decided to remove all the PTM LM and install PTM 7950. Well, The Thermaldelta got reduced and i got even more CB R23 Points after the initial Burnin. And then i removed the Heatpipe again and took a look how the PTM 7950 was spread.
Same but more Uniform.
I have decided to “Lap” the Hexagonal Surface to reduce the Dents and reapply PTM 7950 (just a little to see if it helps (600grit/1000grit/2000grit). Now if i ever want to get Liquid Metal under there again i have to get a new Heatsink, as it was coated Copper. And bare Copper without Coating hates Liquid Metal.BUT now i get nearly 15900 Points in CB R23 and the Thermals are more even across all Cores. (the highest Delta is still 10C) With the Original Assembly it got 15400 Points. The CPU Power goes up to 61w and reduces to 51W during an Single Run. But still hits the 100C Thermal Limit instantly.
I Just did a Cinebench24 Run and you can see my old Mainboards Score and the new Score with my new Mainboard and modified Cooling Solution over 70Points more is massive in cb24
Even my Single Core Score was massivly affected before. 102 vs 80 Points. Thats Wild!
I did further Testing by Limiting the Maximum Temperature with Universal x86 Tuning Utility after i modified my Thermal Solution
MaxTemp / SOCWattage / Clock (ghz) / CBR23 MC Score (Single Run)
60 / 21 / 2,95 / 11130
70 / 28 / 3,50 / 13014
80 / 36 / 3,90 / 14735
90 / 42 / 4,10 / 15190
95 / 47 / 4,15 / 15539
100 / 50 / 4,30 / 15918
My Conclusion there is a Quality Problem with either the Thermal Interface Material or the Cooling Solution. Why are there those Dents which in my Investigation clearly induce the high Thermal Delta between Cores one Core sits flat agains the Heatsink and the Other Core has a Dent above it. Is there enough TIM or has it Airpockets induced by the dents?
As a next Step, i will take the Cooling apart again. I will try to further lap the Contact Area with a file and again fine grit + polish grit to reduce the Dents if possible to zero. Major Concern is, if the Clamping force is still enough and will the surounding area hit any component on the Board or the cpu median.
I will see how far i can drive this Test without damaging it.
The Cooling Solution of the FW16 should be able to get more than 16k near 17k Points out of the 7840HS as even the Razer Blade 14 and Asus ROG G14 are able to hit those numbers, and they share their Cooling System with the GPU and they do not hit 100C at an instant.
Maybe someone from the Team could give a little Feedback if there is a spread Problem @Destroya
Greeting Psierra117
7 Likes
A shared cooling solutin will do better if ther is no GPU Load im not sure if that comparisons works.
As it wont be the same if you are running CPU+GPU Stress test.
But its really ineresting that the PTM7950 works better then the Liqud Metal.
Could you show how you laped the CPU Cooling solution?
Sounds really interesting but i hope you don´t brake you Cooler because i don´t think they will send out a new one as Waranty replacement.
1 Like
Thank you, I’ll share this with the team.
5 Likes
Interesting…and brave! So to clarify, you are showing that you remedied a faulty application of the liquid metal by using the PTM 7950, not that the liquid metal is overall inferior, correct?
1 Like
It sounds to me like there may not be quite enough LM to fill the space between the CPU and the heat pipe plate, resulting in reduced heat transfer.
1 Like
I threw the Conclusion in the Room, that it is possible, that there is not enough Liquid Metal to fill the Holes/Dents fully. It is quite possible, that the PTM7950 is just “more” mass filling the cravices.
The Main Question is why are there those Dents? I can suggest, that they are there to help keeping the LM in Place to not short anything out, but the Barrier Sponge and Glue all around it should be enough to keep it there. Are those really necessary? They are quite fine machined hexagons (Cost?). Its a Plate glued/welded/soldered to the Vapor Chamber. I suggest its a Vapor Chamber between the Heatpipes and the Die. Is the Plate possible to be removed and replaced with a thin Copper Element?
And i would think the PTM7950 vs LM should perform within 2-5% Margin, They are not far off in Comparison. I did alot of Testing and Comparing of PTM7950 vs Thermal Paste on the Steamdeck and Legion Go and even one Test LM vs PTM on my Ryzen 7 3700x and they sat in between margin of error in favor for the LM.
*edit
And here is a Die Picture, it was on both Mainboards for me, that Core 4 has hit the Thermal Limit with a Gap to the other Cores. I don’t know how its counted upon the Picture, but you can lay it over my Spread Pictures and see, where there is more or less Thermal Material. Or where the most Dots are seen 
The CPU Cores are in the Area closest to the Revision Number on the Mainboard.
*edit2
There is a Great Comparison Video on the 7840hs vs the 13700h from Jarrods Tech Testing on a Legion Notebook. And there the 7840hs is hitting around 63C on a 50w TDP, which is nearly 40C less on the same TDP the Framework hits Thermal Throttle on a similiar/smaller sized Laptop with a shared Heatsink.
*edit3
Jarrod Tech did even Review the Framework 16 The Benchmarks he showed on his 7840hs are not matchable with my Modification and far far off with my First Mainboard.
There he states, that the Cooling Manufacturing between the Review and Customer Units will get somedifferent but quote “with better Results”
Here another Full Load Test, with added GPU Load.
3 Likes
TL;DR version? Thermal paste better than liquid metal?
1 Like
No.
TL;DR: Maybe not enough Liquid Metal, maybe wrong designed Mating Area between Die and Heatsink.
3 Likes
little Follow-Up.
further Lapping worked and gave some extra Points (more than 200) But its just a after some Single Run heating and cooling cycles, not the 10min Loop. All Tests in the attached Pictures before also had been the single Run.
Before further lapping
during the Process
nearly finished before polishing
And a 10minute CB24 Loop for my own Verification
3 Likes
Those “dents” are probably the “etched pattern” that frameworks presents in its deep dive on liquid metal, “etched pattern” aimed at holding the liquid metal through surface tension.
Well than my assumptions in the other Threads had been right. I thought it had to do smth with the surface tensuon to hold the lm. But its “shit” as filing those away and using worse TIM inproves performance 
1 Like
I could imagine the liquid metal not filling those small recesses properly, resulting in lots of tiny air gaps across the mating surface. Just a guess, but how do you evacuate these recesses if both the flat area and and the liquid metal sheet are perfectly level? If this were the problem, adding a grid of small channels between the recesses might help, like so:
I hadn’t read the Deep Dive properly, but my conclusion upon my findings was the same. Either not enough material under the plate or probably airpockets in the edged dents. I Hope those Reviews will help to produce a better Cooling System V2.
3 Likes
I understand the one piece trackpad idea. But how many? One left of center, one centered, and one right?
That might get to be a lot of unused inventory…
Hey @Destroya - I just had to create a ticket for an RMA for this same issue.
R23 score of only 14k in linux and even lower… 12300 in Windows!!!
Constantly ramping up to 100c the second the task starts… noticing more and more thermal throttling in games.
I have a feeling level 1 support is going to put me through a ringer of questions that wont have anything to do with the issue.
Before I go the crazy route of putting PTM 7950 on the CPU and risk breaking things removing the current LM - can you help make sure the ticket gets adequate attention?
I love my F16 - but I want to get the performance I paid $2200 for 
2 Likes
I just added this post Framework 16 - Uneven CPU thermals - Steps to test / Questions from support
It should help you skip about 20 emails and cut down the debug process by 80-90%.
2 Likes
In the Batch 20 Thread and the uneven Thermal Thread. There are more and more People complaining and testing after i started modifying my Heattransfer with PTM and lapping the Heatsink. I just started this after i received a Replacement Mainboard with also bad Performance and Thermals.
2 Likes
Yeah, i dont think im willing to file down the heatsink… yet…
but if the RMA doesnt perform well… ill be pulling off the LM and getting some 7950