To be honest, even on my “degraded performance” setup, the laptop is pretty snappy. Regarding the noise, I would recommend not getting the dGPU. The gpu fans are a lot louder.
Just wondering if I can take advantage of the bigger fans by using the graphics module “without the actual GPU” (i think i you can get the graphics module without the GPU in it)would it work to help dissipate the heat faster?
I had a similar idea, but as far as I’ve been able to test, there is no benefit to the cpu.
Did you already test this scenario?
- were the larger fans louder?
- the graphic module is larger and without the gpu in it i would have imagined heat has a large area to dissipate and the advantage of the bigger fans? No?
Mine had the dGPU in it, but I tested this more or less. I ran the dGPU fans on a cpu intensive load with no GPU load. The fans are significantly louder (“night-and-day” louder) and during the cpu benchmark (cinebench screenshot posted above) I was unable to get more performance or average wattage. My performance had actually decreased by ~400pts compared to my relatively recent benchmark with the shell fans. I’m guessing that degradation is more related to the issue this thread has been dedicated to though than to the change in fans.
pain. i got at least 14000.
re-paste?
Not correct. the cpu contact area is now copper, with some new baffles and even more thermal material around the powere delivery area.
Not correct, The Pictures in the Framework Marketplace have been there like this before i ordered my Spare Heatsinks and my Spare Heatsinks had alot higher Serialnumbers than my Stock Heatsink but had been the same Structure, without the Exposed Copper Core but still the badly soldered Shim with hexagonal Structure. The Pictures are pretty Sure the Pre-Serialproduction/ or prototypes.
Yeah I’m likely going to do something, just waiting for the resolution announcement from Framework so I don’t spend more time on it than I need/want to. I got a score in the 14000 range once, not no more though.
I am rather skeptical, since I thought if they have a fix (new revision) they prob would have annoucned it. Thanks for confirmation.
Fortunately this is not as serious as a issue like “oh my SSD don’t show up” or “oh my computer don’t turn on”.
45W on the integrated SoC is “mm, ok that’s expensive but not outrageous”. 80W is “ooh, spicy”
Yup, basically where my heads at.
additionally, normal thermal paste can squish out.
Thinking about trying this out next year. Where do you source the copper shim from? The other stuff I think I can figure out how to source and do on my own.
If you remove the GPU, the heatsink for the CPU doesn’t increase in size, as doesn’t the opening for the CPU cooling air to go through.
Fw16 uses separate heatsinks for CPU and dGPU cooling. GPU is cooled by rear exhausts, CPU by side exhaust.
You might benefit from bigger fans, or you might not, if you block the rear exhaust holes.
I think @PSierra117 found his on amazon, I found mine on ebay, just look for copper shim laptop and some should pop up in different sizes and thickness, the ones I got were 20*20mm size and 0,3-0,8mm thickness (in multiples).
Nirav confirmed today in a QnA that the FW16 is moving from Liquid Metal to PTM 7958 in production, and they will have a replacement/upgrade kit available.
Clip (not the full clip, as you can only make those up to 1min on YouTube, watching past the clip will give additional info): https://youtube.com/clip/Ugkxjdc7w9Sr-hWhkbuCcWSGYjl9l_qZ69_P
There is also now a knowledgebase article for requesting a replacement kit: Phase Change Thermal Pad Kit Request form
What’s the advantage of PTM7958 over PTM7950?
According to the QnA they’re the same thing, but the 7958 just works better in a manufacturing/machining context
So with the announcement from Framework on switching to PTM, I decided to finally make the change on mine.
Pre-change, my R23 score had dropped to 13,350
After change, its now 15368 (on only the second run).
Will be excited to see what happens as the PTM burns in.