Yes, that’s what I’m telling.
I don’t really understand what your point is then. You’re pointing at the MBA as an example of a 30W silent machine, when in reality it’s benching as a 9W machine, that also runs it’s chip at 108* when doing so.
Others on this forum have already shown the existing FM (and possibly many other machines) can run at ~20W sustained load which is higher than the MBA’s results. At this point, the FM portion of the design (heatsink, chassis, power limit) is already proving to be capable of running silent, the constraint is the chipsets available.
Apple is not going to sell anyone else their M series chips, and they wouldn’t be able to run Windows anyway (they’re definitely not allowing FM to use MacOS). Unless Microsoft makes ARM windows a real thing finally, and Qualcomm CPU boards become common, I doubt you’ll find much demand for non x84 boards in the near future.
I think more likely is you eventually see someone mod a FM cooling system to use one of those AirJet coolers, giving it enough cooling for most tasks, while remaining silent compared to a fan.
9 Likes
My point is: I never said MBA cools 30W CPU passively, I said that MBA cools 30W system passively, and actually I was just rounding numbers; according to the review the actual figure it 28.5W.
I don’t see the point of isolating the CPU either. You should always think of it as a system, at least I find it more useful.
I agree with a previous comment,:
I have disabled turbo boost and the fan almost never kick in with intel 12th gen. I have NO noise during my day, even in hot summer. Only when I do a video render eventually, or gaming.
I think with AMD 5nm process it will be even better (its basically 4X more power efficient). That s the node Apple is using on their fanless laptops.
You might want to disable 2 big cores aswell , to mimic the 15W intel CPUs (2+8).
Interesting I thought there was something sketchy going on there.
Don’t particularly care about the 108 (who cares about silicon degradation if the soldered storage dies and turns the whole thing into a brick after a couple years), but the 45C skin temperature for 9w is not great.
What’s the obsession with disabling cores here, if they help using more cores at a lower frequency is more perf/w and if they don’t they sleep anyway.
Only thing it does is limit the max performance which you could also do with a power limit (and still get the better perf/w).
This is a bit tone death, making a fanless laptop and any hardware in the first place is extremely hard, I also doubt Framework has the finances to pay for ARM licensing, if they went with RISC it would be the RISC-V I imagine.
Tone deaf*
And yes, making computers is hard. Ya got me. My point is that if you can make a laptop, and Framework seems to be able to, then it’s not that hard to make it run cool if you’re willing to compromise performance.
The second part about ARM licensing…you’re sort of implying ARM is the only way to build a fanless laptop which is just not true, but then continuing on to assume what, they would need to build their own custom ARM chips? Why would they do that? What can’t they afford? How do you know this? I’m sorry, but there are just too many assumptions and logical missteps in that one short post.
You could probably setup a script/automation to reset at every reboot
Alternatively you could also manually wire the fan to the 5V rail or 7V rail (negative on 5V, positive on 12V), I think it’s the best of both worlds, its virtually inaudible but it still moves some air to remove some heat
While this was a fun hack on desktop I’d advise you to not do this kind of thing with laptop fans (assuming you can find a 12v rail somewhere XD), putting gnd on anything that isn’t gnd is at best going to throw off rpm readouts and fan control and at worst blow up most of the main-board.
I would like a fan-off mode for one specific use case: setting the laptop on a bed and watching a movie.
In that case I want to know for sure that the fan won’t come on, because I don’t want it sucking all the lint off my bed into the grill.
The rest of the time I’m happy to have the fan come on as normal.
I have a GPD Win Max 2, which has a default TDP of 25 watts. But it has a keyboard shortcut that will turn on “silent fan mode” that reduces the fan speed and drops the TDP to 15w. The future is awesome, and I use it fairly frequently.
I’d like something similar to this, but with the ability to go one notch further and turn off the fan entirety and cut the TDP down to whatever it it can cool passively, probably 3~7w.
(I don’t have a framework laptop yet, but I’m looking at getting a 13-inch one soon. I’ll probably dual-boot Windows and linux, so a software solution that works and at least one of those would be nice. A hardware solution that works in both would be even better.)
Under linux that is quite easily configurable. Did that on my laptop. There are quite some tools available for that.
Just make sure you lower the CPU frequencies etc.
I see a perfect fit between the Framework 16 expansion bay and the AirJet technology from Frore Systems since the fans are in the expansion bay and can be swapped out. That would be an awesome collaboration.
Update: Just to be clear, this only works when your aim is to improve the performance of a silent machine. If performance is the name of your game, you need something else.
I see this AirJret Tech commentary all over the place, and in almost no instances is it a suitable replacement for a fan. The AirJet is designed to replace a giant finned heatsink on low TDP systems i.e. 7w or less. When you have most laptop chipsets sitting at 15w at the lowest and 28w at the preferred operatinig range the amount of power going to the cooling system i.e. the AirJet would need its own cooling solution. The tech is simply not there yet when it comes to cooling these kinds of chipsets at a reasonable energy cost.
1 Like
You’re looking at it from the wrong angle. It’s a great way to make passively cooled devices faster. So the idea of a passively cooled Framework 16 becomes more feasable this way. Obviously a performance oriented machine has no business using AirJet stuff.
Buy an AMD, and shutdwon 1st turbo boost, test , then CPUs from the BIOS and bam you have a fanless computer 
I switched of turboboost on my intel 12th gen and never heard the fan , except if I do games or video encoding. On intel if you want the U from intel i5 just deactivate two BIG cores out of 4
From what I have seen the current gen airjets are quiet but nowhere near silent, the noise profile may actually be more annoying than a normal fan to some users.
But anyway those things aren’t really ready for most use-cases jet but if they have generational improvements anywhere near what they claim they are going to be quite interesting in a couple generations.
2 Likes
The way they’re intended to be used in laptops is putting multiple of them on either a vapor chamber or heat pipes which then contact the cold plate and soc.
You put 3x minis to get to 15w or 4x to get 20w. OTOH at this point I think there’s only one shipping product using them at this point so it makes sense to view the marketing with skepticism.
However my understanding was it was costing 1w per 5w of heat dissipation. So my 12th gen would need 7 units at 7w. That is more than my entire laptop uses currently on battery. They simply are not there yet in efficiency to make any sense on a laptop.
3 Likes
The power to run the unit does not take from what it can dissipate. It does not run that power through the soc/CPU. It would, however impact battery life.
For clarity: if you run 4x minis you get an effective 20w of dissipation for CPU/soc - you don’t subtract the power to drive the unit. Similarly, 3x of the pro is 30w of effective dissipation for the CPU/soc. Again though the power has to come from somewhere, and if you aren’t plugged in, that’s the battery.
I see I screwed up the math too many things at once today. My point was that the expected power draw for the units would in short drastically reduce battery life.