I’m wondering who designed the fans in this thing.
They are strange, centrifugal-oriented design.
Nothing optimal in them. I say that as a professional physicist, doing hydrodynamics.
If you want something nice and quiet, you need to do mixed-flow.
Axial flow, radial flow, mixed flow, mixed flow needs the largest volume. Are you sure that you can put a mixed flow fan into a laptop?
There is amount of how much “mixed” you want, that would be fine. The current fan is IMHO bonkers inefficient. ..trying to say more blades does not make it better.
You mean, the industry standard for laptop fans?
But also, it would seem that if it was worth adding, manufactures would add them.
I’m definitely not a physicist, so my answers are based on some AI prompting.
Here are the results - note, I even asked it for an example where this would be a benefit (the tl;dr, its not ideal):
Mixed-flow fans require a conical or volute housing to manage a 45° discharge vector. In a laptop—where you usually have less than 10mm of vertical clearance—there is no room for that geometry. If you flattened a mixed-flow fan to fit a chassis, the “mixed” angle would hit the top and bottom of the case immediately, causing massive turbulence and flow separation.
The Physics:
- Static Pressure: Laptop heatsinks are incredibly dense. You need high static pressure to “punch” air through those tiny fin gaps. Centrifugal (radial) fans excel at this; mixed-flow fans are designed for high-volume, medium-pressure environments like HVAC ducts.
- The 90° Turn: Laptop cooling relies on pulling air from the bottom and exhausting it out the side. A centrifugal fan handles this 90° turn natively. A mixed-flow fan discharges at ~45°, which would just slam air into the laptop’s “ceiling” and “floor,” killing efficiency.
The Verdict: In a 1-inch thick server or a 6-inch duct, mixed-flow is great. In a 0.5-inch laptop, it’s a geometric and aerodynamic mismatch.
The Physics: P-Q Curves
In any cooling system, as Static Pressure (P) increases, Airflow (Q) drops.
- Laptops are “High Resistance” systems: The heatsink fins are packed so tightly (often <0.2mm apart) that the “System Resistance Curve” is very steep.
- Centrifugal (Blower) Fans: These have a “flat” P-Q curve. They maintain high pressure even as airflow is restricted. They “punch” through the resistance.
- Mixed-Flow Fans: These have a “steep” P-Q curve. They move a lot of air in open space, but as soon as they hit the resistance of a laptop heatsink, their airflow (Q) collapses.
The result: A mixed-flow fan in a laptop would spin at high RPM, make a lot of noise, but move almost zero air through the actual copper fins.
When would mixed-flow actually work?
To make mixed-flow viable, you have to remove the two biggest constraints: Z-height and Static Pressure.
The Example: A “Desktop Replacement” (DTR) Laptop Imagine a “laptop” that is 2 or 3 inches thick (like an old-school Alienware or Clevo monster).
- Chassis Depth: With 50mm+ of vertical space, you finally have enough room for a conical housing that allows air to discharge at a 45° angle without hitting the “ceiling” of the case.
- Low-Density Heatsinks: If the laptop used massive, thick radiators with wide spacing between the fins (reducing air resistance), the mixed-flow fan could take advantage of its higher CFM (Cubic Feet per Minute) capacity.
In short: It works if the laptop stops trying to be thin and stops using dense heatsinks. At that point, you basically have a portable desktop, where the physics of a 120mm case fan (or a mixed-flow equivalent) finally start to make sense.
Nope, the flow is always “mixed”, assuming 45dgr flow is incorrect. The incident degree is always something. In a laptop fan, maybe 10-15 dgr could be utilized.
Always funny when someone tries AI against a professional.
But to my eyes the fans look like they have been designed by children. Could do way better ones. Would cost like crap though.
1 Like
I mean, I was completely up-front that I used AI - and it seems to defend an argument against your ideas more than you are defending your argument for it. As you are a professional it would be interesting to see you debate an LLM and see if you can convince it you are right. As they can be manipulated.
In short though, and this part I did know - with laptops, the intent is static pressure, due to the density of the copper fins. Without that pressure, the air wont pass through the fins.
Apple, Razer, Asus, etc use the exact same design for this reason.
Argument that you need more fins for that is just incorrect.
Fans we see have been produced as cheaply as possible, so you don’t see any proper designs, as you see in some GPU:s - think why NVIDIA’s own cards use just one fan…
Yep, because they are cheap. Plain and simple. But I can still whine about it.
Hi,
If you can design a fan that fits the space and is quieter and still provides enough cooling. There will be a demand for it here. I say, give it a try.
4 Likes
You’re saying the fans you want to design would cost too much?
To me, it sounds childish to think one can just not take into account costs, ease of production, and the other factors at play in consumer products.
But I understand the feeling. The annoyance of it. Often it feels like few things we touch are designed as well as they could be, even without increasing costs.
I guess an alternative cooling would be if one was somehow able to use the entire top cover (where the display is) as the heat sink / cooling surface, and somehow use heat pipes to get the heat there.
That hot surface would then be away from the base of the unit.
Getting heat pipes to work across a hinge would be an interesting challenge.
1 Like
It sounds like you’ve got a hypothesis here that the fan design could be improved. I’d love to see that tested. Is there a way you could build a prototype of how you think the fans should be made and then test the two side by side?
2 Likes
They mostly don’t do that anymore anymore
Sounds like an untapped market if true. I am fairly sure a bunch of people here would pay quite a bit for even marginally quieter or higher performance fans and they are fairly modular.
2 Likes
Agreed. While my 5070+Framework control center+noise cancelling headphones make for a more than adequate setup for me personally, I wouldn’t mind spending upwards of 75€ to have a “better” cooling solution.
If you’re in the position of designing and producing prototypes, go for it! I’d happily test them.
AFAIK (Gamer’s Nexus videos with thermals specialists) designing fans is hard. Tradeoffs in size, noise, CFM, cost, power and pressure must be met. The laptop space is even worse, due to stricter size constraints. My knowledge ends here, I can’t wait to see prototypes!
I am not sure one would even need a prototype.
If one can use a computer model simulation of the current fans, showing their air flow, sound level etc. And then your suggested improvements. Showing the improvement as output from the computer model simulation. That would be a low cost start to convince people there is a better way.
That’s true, that definitely lowers the barrier of entry for prototyping. Naturally, you’ll lose some information around manufacturing feasibility and the subjective “annoyance” factor of the sound produced, but it’s a solid starting point for sure.
Water cooling ? Where the lid is the radiator with flex pipe to the heat sink 
2 Likes
I recall LTT water cooling a laptop at one point or reviewed one that could be water cooled (I can’t remember which). I’ve thought about how cool (pun intended) it would be to water cool my FW16. I don’t have the capability or means to do so (or the time!), but certainly an entertaining thought. I’d think the FW16 would be perfect laptop to try and design a water cooling solution in given the modularity of the laptop.
I do think there’s plenty of room to make better cooling systems for the FW16. I feel like those who get a dGPU, like myself, aren’t as concerned with size of the laptop as others and would consider the tradeoff of a thicker laptop worth it for better performance/temperatures/accostics/ect.
I’ve noticed that the part of the bottom cover above which the left heat pipe runs heats up substantially and there’s a lot of flat space there to attach some kind of heat sink. Could be really a flat-ish pipe of the same shape as the one used inside, but with active circulation.
I have no complaints about the Framework’s thermals, but now that you mentioned it I’m tempted to check how my poor man’s phase change material - candle wax - fares.
A 10x10x0.3cm sheet would only absorb around 2 minutes’ worth of heat from the CPU running at full throttle before melting, but obviously that’s not where most of it goes when the fans are running, so my guess is that with such a setup I would have closer to 15-20min of weak, but measurable additional cooling.
I also haven’t had any issues with thermals but the idea of using water cooling is to pull the heat away from the laptop. As you said the bottom cover does heat up and I frequently game with my laptop on my lap while I’m laying in bed. Water cooling my not be the best for my use case but even a better fan or better dissipation from that area would be great.
On a note about heat dissipation I think I recall a small gap between the exhaust in the chassis and the heat sink fins. This leads me to believe they could have put a slightly bigger heat sink in the FW16.