8 posts were split to a new topic: Fan Control
Can you just boot up the laptop without expansion bay installed?
Would need a massive heatsink to spread the heat out passively.
Depends on how much heat, the 13 can already do quite a lot with the fan off, depending on how high you want to let the skin temp get. I’d hope the much bigger heat-sink on the 16 would do better than that.
Hi.
I have to agree, that the work that the framework team has done, is outstanding. In regard to cooling solution and build quality.
But, it’s been almost half a decade since I’ve seen an quality “rugged” laptop/barebone that can survive different “harsher” environments. Such as heat and particles. And it would be great if an manufacturer like Framework would come up with a solution for said problem.
And of course the easiest solution to said problem is making a laptop that does not need fans.
Other manufacturers seem to think that, all users will ever need is: “bigger numbers” on their benchmarks (which is not always the case… I would prefer my laptop to actually work more than a few years). Which is why I still have my old 2018 Omen 17 (7700hq: 1070 laptop not mobile, which is just full 1070 with power limit lowered). And it’s a great machine, it even played cyberpunk without lag, but it’s really old by now. And even the weakest integrated cpu’s beat the 7700hq.
I live near a beach, and so I live next to an active road and the sea (which translates to heated salt water in the air and lots of dust particles). And ideally I would like to be able to use my next laptop: in bed before sleep (watching movies or playing visual novel type games): on public transportation to-and-from work (I currently spend more than 3 hours, on a work day, on public transportation): in the beach house cafe (to enjoy the outdoors even a little bit, and possibly get some sun).
Now some of you might tell me to get an tablet for those, but… As i have bad blood supply to my ligaments, then that is not an lasting solution. Meaning that, my arms go numb if I hold them up for for than 10 minutes and my hands cramp up when I clamp something with them- or extend an finger out (out of the normal position, for touch controls for example) for an extended period of time.
One thing that manufacturers seem to be dead set on is making the smallest/thinnest/lightest computer possible. While it is a good thing, I personally don’t think that it should be one of the main sales pitches. I mean honestly, I couldn’t care if the laptop is 1/4 of an inch or 1/2 of an inch thick. And same goes for the weight: I don’t care if the laptop weighs half a pound or 3 pounds (still way lighter than my Alienware M17xR4, which is 6kg or 13 pounds). And I don’t need an laptop that I can break on my knee.
While the fact that Framework let’s us change the components in our laptops to reduce e-waste is amazing… It just leaves me with the question of: “Why do we have to change the whole cooling solution, while we maybe only want to change the fan?”. The current solution not only pulls dust into the fan (which is normal and fully acceptable), but also on the motherboard itself, which can cause serious problems for it’s longevity/life expectancy. Now I understand that it’s still a work in progress, and it’s bound to change. So I can’t really fault Framework for that, but i’ld like to read some of their thoughts on that.
Now, as I don’t want for this post to be all criticism and me harping on their product (their so-called baby), then I’ld like to offer up some possible solutions next.
- Use a bigger/thicker chassis: this allows for more room for air to move and possibly apply better cooling solution and/or cooling surface.
- You could use a motherboard with an lower tdp cpu on said bigger/thicker chassis to reduce the cooling required and/or remove the active cooling solution entirely for an passive cooling solution.
- Seal off the active cooling solution part from the rest of the cooling solution, so it wouldn’t push dust into the rest of the system internals.
- Make the fan an separate item. So people could just buy the fan itself to replace their old one.
- Switch from bottom air intake, to side air intake. As the only solution that might take in more dust than bottom intake is top intake, which only leaves side intake. And as most laptops are used on a flat table or on lap or on a blanket, then there is almost always an airflow restriction on the bottom intake active cooling solution.
- Design sealed off “air ways”/“wind tunnels”. As to prevent hot air circulation inside the system.
- Use higher tdp cooling solution (even on lower tdp cpu’s), if you can. There has never been a problem with having too good a cooling solution. This applies to both: fans and possible future ‘Solid State Coolers’. As just enough does not last very long.
Now i know this post is probably too long, with me yammering on, so my apologies for that. But those are my thoughts on that matter. And I am kind of happy that I got to give my two cents on that somewhere, where it might actually be useful.
So, thank you.
The CPU thermal throttles anyway, so one could turn off the fans with ectool if needed.
The air actually get pulled in top and bottom and immidialty exits through the side. I’m pretty sure one could seal it dust proof with some cardboard and tape.
Even the 15W cpu? That would be an big design problem, especially if it already is in a bigger chassis, with better cooling solution.
As for the fan, i meant the mainboard solution fan.
Okay, I did not really know that (thanks for the correction). But cardboard and tape? Well first of all, I am pretty sure, unless they solve the heat problem first, that cpu/apu will be toasted in a very short time. And secondly that point wasn’t completely about dust proofing, but about the dust and heat collecting on the internal electronics (a.k.a the motherboard). As some might know, battery and hard drives do not like heat (of which you already corrected me on).
As I’ve only used an framework laptop a few times for work. But I have looked into buying one at some point.
I mean that the CPU would throttle if you turn off the fan, so it’s not necessary to limit the TDP.
There are no other fans.
That is probably the worst solution ever. To let it thermal throttle, instead of lowering tdp. That amounts to intentionally baking the cpu.
Hmm, then what is this? -
The CPU is designed for 100°C, but sure limiting TDP is definitly also an option. Not sure if ectool can do that yet.
There is definitly no way for air to go anywhere but out.
It’s for the FW13, this is the FW16 subforum.
Yeah, I am so new here, I didn’t even know that. So wait, all the cooling is in the expansion slot? Also, is there a 13 fanless subforum?
Well if they were to put 13 in bigger chassis, it would become 16, but i get your point.
Umm yeah… I don’t see any way to lose 45W+ with such a form factor just from passive cooling, sorry. Maybe peltier elements or something similar in nature.
Sorry for the mix up.
The fans are in the expansion bay. The heatpipes etc. are part of the mainboard. There is a slot in the plastic between them to let the air through.
Peltier elements would certainly be a) kill the battery life b) exceed chassis temperature limits. You just can’t have everything, but I think limiting TDP - either directly or indirectly thourgh thermal limits - gives still enough usable performance for most tasks.
My fans are definitly off right now, even though I’m writing you this message and download something in the background. Current CPU temperature is 50°C.
So I think the FW16 is already like 90% of the laptop you want.
Yeah. A year ago I saw they made an 11th gen intel one (15 inch) that was almost there with passive cooling (there were actual laptops that had no fans), so I thought they might have improved upon it.
And you are correct actual peltier would be: first and complete overkill as cooler (it goes farther than just freezing). And secondly, the other end would be too hot (would probably burn the table and users lap, not to mention the other components).
And yes, it will drain battery.
The Framework 16 also seems to draw the same amount of power as my 2018 Omen, which means 2h 30min of battery life with the graphics card add-on for 90Wh battery.
I almost want to try to use the expansion shell bay and disconnect the fans to see what happens.
For non-CPU heavy work, the fans doesn’t seem like they kick on. If they are, they are super slow and quiet. So far I haven’t found a way to monitor the RPM of the fans in Linux (lm-sensors nor nvtop). I wonder if they show up in Windows (hwinfo64 maybe?)
They don’t show up in HWInfo 8.00
Huh… I wonder if disconnecting the fans would cause the laptop to throw errors or warnings then.
I am also kind of curious, if the new usb3’s are safer than the old ones:
I used to have an synthesizer, that had an unknown adapter (1.6V higher than the requirement). And one day when I turned it on before hooking it up to my Omen, it fried completely (puffed the smoke, melted some of the plastic, etc). Now why that happened was cause my usb3, upon connecting, pushed the voltage back with 18V (not the regular 5V). And the adapter was only an 16.4V one.
I found that my machine works quite well without using the fans for non CPU heavy loads, when I reduce the CPU temperature to about 60-70°C (with ryzenadj) and switch to a fan profile that won’t switch on the fans until they go over that temperature.
Even short bursts of CPU load are easily soaked up by the heatsink, and even fanless compilation is possible, although at much lower frequencies and the whole machine gets quite warm then, but I still see all cores at around 2GHz without any active cooling, even at full load for a couple of minutes (compiling a smaller program for example, I use Gentoo).
that’s very interesting and encouraging to hear. i have a fanless machine that couldn’t manage to hold up to 1 ghz when under all core load. (which really made the i7 questionable. lessons learned)
I only tested full load for a few minutes at most, the case gets uncomfortably warm without any fans. I now have it set so they run at about 1000rpm when the CPU runs hot for more than a couple of seconds. At that speed they are nearly silent, and the machine can run at about 2.6GHz for hours, case still gets warm, but not uncomfortably so. My old machine only hit 2.6GHz on all eight cores with the fans screaming, and this one does so with the fans whispering.
I’m amazed by the cooling system though, today I tried to reduce fan speed while compiling, to see how low I could go before losing speed, and I was still hitting 4.2GHz on all cores at about 3500rpm (running at 95°C, I like keeping my CPU at least 5°C below the design temperature), which is definitely noticable, but considering the speed, it’s quite impressive. My compile was done before I could see if I could go lower without the CPU throttling, so in time I may be able to get even slower. I’m curious if I still get 4.2GHz at 3000rpm, or even 2500rpm without significant losses in speed.
It seems higher fan speeds are really only needed when using the dGPU, dissipating up to 100W is a different ballpark than the 35W the CPU generates at full load. But I’ll wait for a 240W power supply to fully utilize the GPU, I dislike it when the battery discharges when I’m gaming. So for now, I try and keep the CPU below 60-70W, to ensure the battery doesn’t discharge.