Honestly this isn't more than I cared to know about PC fans. I've worked extensively with industrial appliances and fans are bloody important and underlooked: fan failure is often the step that precedes component failure. The one part of the spec I wasn't aware of was just how insanely high the maximum current draw is; I'm glad that we've made more efficient, longer-lasting, effectively silent air-movers that can fit easily within that power budget.
The problem with that is when the fan would block the natural venting if it stops working. These devices are engineered to have the heat move the air through and out of the device, and though a fan will usually do a better job of that, it will ultimately fail. When it does, if the fan body is in the way then the intended convection is blocked and the entire unit shuts down or is damaged.
Cable modems are often placed in poorly ventilated spaces that overheat. So you can get the best of both worlds by including generally unneeded fans which enable the device to function in sub optimal conditions, but can also extend lifespan in better ventilated spaces.
The alternative of passive cooling in such spaces is unreasonably expense because only a subset of devices are subjected to such conditions.
I have thought about adding a bathroom style vent fan to the closet that houses the network gear for the house. Exhaust near the top and have a cool air intake near the floor.
Bathroom fans work too if you’re able to vent through the ceiling. I can recommend the Panasonic whisper fans (available under a variety of models with “whisper” in the name). I’m using one to exhaust the warm air from my home theater emitted by the projector. I think it’s this one:
yeah, if you look at the equations at the bottom, power goes up with the third power in relation to rpm. whippin' and drivin' that much air takes some oomph. connect that fan to any normal motherboard, and you're going to burn out your fan header pretty much immediately (especially since inrush current is probably even higher than the steady state).
Hooking them together would provide 24VDC. On a 350W PSU, I see -12V rated to supply 0.3A, which would allow for 24V fans drawing up to 7.2W of power, if desirable.
I don't see any reason to prefer 24V fans, however. The current losses over such short ranges are minimal, and I would think the larger economies of scale of 12V devices will keep 12V fans less expensive.
Do not use 24V fans for anything, if you have any control over it.
24V fans in general are poorly stocked. 24V 4-wire PWM fans are, these days, more or less special order items. There are thousands of 12V 4-wire PWM fans in stock at Mouser and Digi-Key. There are less than a dozen 24V 4W fans in stock, and most of those are sizes you probably don't want.
I'm a consulting EE these days. One of the jobs I get to do is clean up designs that aren't working. We always flag these fans when we see them... then nothing gets done because they're "available enough"... then when build time comes we can't source the fans because the "stock item" is completely out of stock and all possible substitutes are out of stock. But 12V versions are available.
So, yeah, stick to 12V fans whenever possible (or 5V for the little guys). Also genuinely never use AC fans, but that's a different story.
This led me to wonder why we still need the -12V supply - I know the -5V supply was removed a while ago because it was only needed for ISA slots. According to Wikipedia, -12V is for RS-232 ports and PCI slots, both of which are hardly seen on any current motherboards. It should be safe to use it for fans, but 7.2 watts is not very much power. Looking on Mouser, I see 24VDC fans that consume 1 watt or 18 watts depending on speed, so one could easily overload the -12V bus if they are trying to get a lot of CFM. I don’t see a compelling reason to use 24V fans in a PC.
From what I've read, ATX12VO really is 12V only; peripheral power connectors (sata/molex) will be connected to the motherboard, where the 5v and 3.3v (if needed) will be generated. 3.3V on sata power plugs was never reliably available and was rarely if ever used, and one of the pins was redfined to be a disable pin, so it's usually best for everyone if 3.3V isn't provided at all.
-12V for RS-232 is easily generated from +5V or similar sources; max232 chips have been available since the 80s
I'd suspect they'd go up to like 48V. It's the max voltage for USB-PD 3.1, which corresponds with being the effective highest voltage you can have before you're no longer "low voltage". This is also why PoE is approximately the same voltage.
People can easily overload 12V motherboard fan headers, too.
> if they are trying to get a lot of CFM.
If so, they probably have a PSU larger than the entry-level 350W unit I specified. Higher wattage PSUs will provide more power on the -12V rail as well.
> I don’t see a compelling reason to use 24V
Neither does anyone else. Still, correcting a bit of misinformation in the article.
huh, i didn't know they could be connected like that! i only knew of the -12V as a source for RS-232. thank you for the tip; i'll investigate this (i assume you're correct, just want to understand it for myself) and update the page. thank you very much!
> huh, i didn't know they could be connected like that!
'Ground' is not necessarily 0V, it is just convenient to use it like that. If 'ground' is 0V then 12V - 0V = 12V, but if you make 'ground' -12V then that is 12V - (-12V) = 24V.
I'm surprised it doesn't mention hysteresis some fans have at low speed.
Cheap fans often don't have actual proper speed controller on them which means they might start spinning at say 30% PWM but only stop spinning at 15% PWM. That was probably the thing that surprised me most when I upgraded to nice Noctua, at certain PWM it just started, always, and below that it stopped, always.
I wish noctua would sell aftermarket replacement CPU fans for laptops. I thought there was something wrong with the loudness of my Dell laptop's fan (with low confidence), but I couldn't gauge that any replacement product wound actually be quieter or cared about noise levels.
Most laptop parts are specifically designed for a particular laptop model. This is very different from the highly standardized desktop parts. Noctua would be sure to lose money selling laptop fans.
This was a great read, I have always had a love of PC fans, and the technology behind them. It'd quite a complicated engineering challenge to produce a high static pressure, high rate of airflow, and minimize noise while maintaining high reliability. I'm one of those folks that has no problem dropping significant money on fans every time I do a PC build, and so far I have very very rarely ever had a fan fail on me, even after decades of running at high duty cycles. The thing I always marvel at is how high the draw is for some of these fans. At one point I even retrofitted a system with an external radiator for water cooling that used 120V 120mm Panaflos and was separately powered from the computer.
I wish folks took cooling more seriously. There are so many devices that benefit from active cooling but rarely get it, or even being strategic around placement, positioning, and mounting to make convective cooling more effective. Thermal limiting is such a common occurrence on modern electronics if you don't give due consideration to cooling, as we've moved towards more passively cooled devices, including PCs (laptop especially).
Also, one thing I've always wondered: why do people want to use hwmon to set fan RPMs? (Or really, why do this from userspace at all?) It seems inherently dangerous to me, as you're asking a process who might not receive CPU time for whatever reasons from the very much not realtime OS to control a fan; if the current RPM is too low, and the system starts generating heat, but the fan controlling process doesn't get CPU time … then what happens?
It seems to me you want fans controlled with something dedicated to it.
The other thing I don't get is all the plethora of options my motherboard gives me to set fans only to fixed RPMs. Am I crazy in that I want the fan to be controlled by heat? (More heat => more RPMs. Keep the system cool, but if there isn't much thermal load, spin the fans down and reduce the noise?)
But by fixing an RPM, it seems the only valid input is "100%"; anything else could be too low under stressful conditions.
I could also have a cheap motherboard. (I definitely won't be purchasing from this manufacturer again, and the motherboard does have other severe quality issues…)
> why do people want to use hwmon to set fan RPMs? (Or really, why do this from userspace at all?)
Some people want to have a mode switch; normal use should be silent/quiet, but when you know you're going to do something big (game, big compile, etc), fix the fans at full so the noise is consistent and cooling is best. (the cooler the chip, the more the boost)
Some people have no good options from the system firmware, and getting _something_ configurable is better. I've run on systems where I couldn't tell the system to actually run the fan, so things would get hot and throttle. Userspace configurability is better than nothing. This tends to be a bigger issue on things that are sold as a whole computer, like laptops, and small formfactor things (which are often pretty much laptops without a battery and built in user interface devices) but also some name brand desktops.
My recent motherboards all seem to have a pretty nice fan configuration tool. Presets for quiet/performance/full speed, and a simple graph based UI to set % by temperature. Most of the fan headers can be set to follow the cpu temperature or the system temperature. When you buy the nice Noctua fans, they also ship 'low noise adapters' that I assume drop the voltage and limit the maximum RPMs and limit noise. Depending on your overall cooling design, that can be reasonable or asking for trouble.
> Some people want to have a mode switch; normal use should be silent/quiet, but when you know you're going to do something big (game, big compile, etc), fix the fans at full so the noise is consistent and cooling is best. (the cooler the chip, the more the boost)
Yeah, I don't doubt someone is like that … I'd just rather it be automatic.
> My recent motherboards all seem to have a pretty nice fan configuration tool.
Mine has a "flashy" tool, I would say. Certainly looks pretty, but again, it's all constant RPM options.
As I lament in the other thread, this is something that would differentiate boards at time of purchase, but no mobo manufacturers marketing dept. seems to have it's shit together enough to get such a differentiation across to the consumer. Instead the focus seems to be completely on the aesthetics of how the board looks.
And again, I've chalked this up to having chosen poorly. But there-in is the problem: assuming I chose poorly, assuming some mobos do support sane defaults/fetures … how do I end up finding and purchasing one of those? Any knowledge I acquire during a purchase is useless the next time around, given the constant product churn HW manufacturers nonsensically do.
The two boards I've gotten recently advertise the features:
https://www.gigabyte.com/Motherboard/A520I-AC-rev-1 look for "Smart Fan 5", there's a tab you can click and see what the customization UI looks like (it's in the firmware settings usable with keyboard or mouse). ITX does mean this isn't a 'value' board, but when I got it the premium above mATX wasn't that much (and probably mostly went to the wireless I don't really need and barely ever use)
My other board is a bit more upmarket https://www.asrock.com/mb/AMD/B550M%20Pro4 it doesn't show anything on the marketting page, but in specifications it mentions "Smart Fan Speed Control" and the UI to configure it is pretty similar.
You get to set about 5 temp -> fan % settings and I can keep things cool without being noisy until I've got sustained load and then it's noisy and warm anyway. The ITX systems can only do so much with a cooler + heatsink height of 36mm (at 37mm the fan housing touches the mesh side panel), and the b550 currently has an anemic Wraith Stealth. Even with 65w target chips, that's not enough to keep them below 90C at high load.
I think my older boards have basic quiet/loud/full speed settings but not detailed ramp settings; but it's been a while and they're either hidden in the basement/garage or not used often so I didn't care about noise. :D
In the old days, it kinda was - at least to your hardware.
Then people realized that blowing up components because a fan failed, or became unplugged, or a filter clogged with dust, maybe wasn't a great user experience and/or caused more in-warranty returns that required replacing hardware ($$expensive$$!), and implemented thermal throttling and thermal cutoffs. Nearly two decades ago at this point, I helped a friend diagnose his computer randomly turning off. It turned out to be a CPU fan unplugged itself, causing overheating to trigger a thermal cutoff. No other apparent harm done.
Fans aren't the only means of limiting heat: slowing stuff down and turning stuff off also works. And it turns out users sometimes would rather stuff run slow than run loud, and maybe your crappy motherboard vendor shouldn't be writing a ton of code running in kernel space - with all the potential stability and security issues that might entail - for whatever network-connected bloatware syncs your RGB lighting and fan settings to the cloud. And they will do exactly that, if that's what's required to give their customers what they want.
Just exposing fan RPMs to userspace might be far less dangerous.
> and maybe your crappy motherboard vendor shouldn't be writing a ton of code running in kernel space - with all the potential stability and security issues that might entail
This wasn't the suggestion I was making. I was suggesting that the motherboard, itself, should be controlling the fan RPMs (or should at least provide such a mode). I don't feel like taking a temperature input, and mapping that to an RPM output should take much circuitry at all, but it that (somehow) required a full-blown CPU, I was thinking a (very small) auxiliary chip, dedicated to the task.
But yes, if you're going to do it on the main CPU, then in userspace. But now you incur all the problems I mentioned in the original comment, some of which can exhibit death spirals: CPU has to throttle due to heat, meaning less CPU time, meaning it will take longer to get to the code responsible for alleviating the problem of heat by notching the RPM up!
In the worse case, you hit the CPU's critical trip point before the problem can be brought under control.
On typical desktop motherboards, the Super IO chip handles all the temperature monitoring and fan control. Those chips usually have a few modes to configure some very simple control system for mapping temperature inputs to fan speed outputs (never anything as advanced as a PID controller).
The main problem is that the Super IO only has access to the temperature sensors on the motherboard itself, and on the CPU (these days, through PECI). There's no standard way for the Super IO to do out of band monitoring of temperatures on your GPU or storage drives, so if you want those to affect fan speeds you need to implement it in software.
Servers typically have BMCs controlling fans, and even Apple's x86 machines have their SMC; in both cases you typically see a more thorough monitoring of component temperatures, configured out of the box with a proper awareness of which fans are blowing across which components. But that stuff doesn't trickle down to the build your own desktop market.
Hmm. I guess I have no idea how this chip interfaces with the kernel then? The only knobs that seem to be documented to exist ever are direct RPM controls for manual control.
I know it's possible (despite the insistence to the contrary on the other thread), as basically every laptop does it. Only do desktops seem to struggle with this concept.
How fan control is configured depends on the SuperIO chip, so it's different for eg. Fintek SuperIOs than for Winbond/Nuvoton SuperIOs.
On Linux, a supported SuperIO will be exposed as a directory under /sys/class/hwmon. On one of my systems, the SuperIO is a Nuvoton NCT6791, so the relevant driver documentation is https://www.kernel.org/doc/Documentation/hwmon/nct6775
Relevant sysfs files to note are pwm[1-7]_mode to toggle between DC voltage and PWM control, and pwm[1-7]_enable to switch between full speed, pure software speed control, and several Nuvoton-specific automatic speed control modes.
> I was thinking a (very small) auxiliary chip, dedicated to the task.
Extra cost (both in design and manufacture), new potential point of failure (both for manufacture and in the field), when the CPU failing was already a single point of failure for the machine. It's not that a full-blown CPU is required, it's that you already have a full-blown CPU that can do the job, simplifying the design. Well, you might see dedicated fan control hardware for server motherboards and other more industrial focused applications, but they often need to coordinate pumps/fans for an entire building - reliability in this context is more about redundancy, and alerting maintainence to the need for repairs, or perhaps switching to a new primary datacenter if the failure is big enough - not in attempting the impossible of ensuring 100% reliability for any individual component.
> CPU death spirals
Have you actually seen one of these that noticably bogs down the fan-driving firmware/drivers and causing issues? I haven't. I've had fan failures. I've had plenty of hardware controlled fans go full apeshit 100% power to the point of being not merely a nuisance, but a problem (audio, vibration, wear+tear, ...). I've heard of building cooling failures. But I don't think I've seen so much as a blog post about the CPU getting so starved that it can't spin up the CPU fans.
And I've had fans not working hard enough - but I'd rather flip a setting in software than open up a case and go hunting for the right jumper, typically. Less disruptive - and the machine is typically usable enough I can still download/install missing software, and google appropriate documentation, which is frequently a lot more difficult to do with the case open.
---
I guess my main point here is that reliable hardware must already assume the potential for cooling failures, and that extra hardware or engineering for a minor improvement to a "purely theoretical" failure mode doesn't sound like it'd pay for itself.
100% have hardware temperature throttles and cutoffs though. Those cut in for a lot of very real failure modes, that I've only heard of actually happening, but personally experienced. Those will pay for themselves.
> Extra cost (both in design and manufacture), new potential point of failure (both for manufacture and in the field)
"But the BOM" is a tired trope in any discussion about "why can't it do X?". My money would be where my mouth is. But HW manufacturers routinely make design decisions that are simple negatives.
Motherboard manufacturers, in particular. Mobo marketing is focused on colors and flashy effects instead of on quality of production and actual function, beyond a bare minimum of tech specification. AFAICT the products are indistinguishable as no attempt is made to stand out.
> when the CPU failing was already a single point of failure for the machine
… and so is the RAM, the disk, the GPU, the northbridge, the southbridge, the eastbridge…
That's not a valid rationale for "let's a adopt a fundamentally flawed design".
> Have you actually seen one of these that noticably bogs down the fan-driving firmware/drivers and causing issues?
"Fan-driving firmware" is essentially the suggestion. It's what I haven't got.
> I've had fan failures. I've had plenty of hardware controlled fans go full apeshit 100% power to the point of being not merely a nuisance, but a problem (audio, vibration, wear+tear, ...).
No; in my current rigs 10 year lifespan, the fans have been run continuously at 100%, and no fan failures.
The noise, of course, is a nuisance. That's why I'm looking for something that can control the fans in response to temperature, such that the fans can be driven at a temperature-appropriate RPM.
But without doing that in userspace, where the controller might just "die" or effectively die, for any number of reasons, outlined earlier.
> But I don't think I've seen so much as a blog post about the CPU getting so starved that it can't spin up the CPU fans.
I couldn't find any posts reasoning about anything. Either it's not a problem, or just nobody is thinking.
> and that extra hardware or engineering for a minor improvement to a "purely theoretical" failure mode doesn't sound like it'd pay for itself.
Yeah, I was just trying to learn before attempting, literally, "IDK, try it and see if you trip the critical temp cutoff."
Yeah, no doubt the various hardware safeties will save you (although it might put some undue stress on stuff to be at >100℃…) but it might also leave you with a rather hard-to-debug situation if you start experiencing stalls or shutdowns.
… and it's only in desktops that this problem seems to exist. On every laptop I've owned, fan control is in response to temp. (And not handled in userspace, although I presume I could download any of the various fan control programs and have it be that way, but the point is that the HW is, out of the box, doing the sane thing.)
you might want to use different algorithms for automatic control than those exposed through the motherboard firmware, which are pretty much limited to either constant, manual setpoints (usually with extrapolation), and a linear temp-to-rpm curve (ideally with different possible sources).
if you look at my linked Counterforce project, i believe there to be value beyond these simplest management schemes. we'll see.
The pinout is wrong, is it not? As it is drawn, if you mate the two connectors, the wires won't match up. (You'll need to mentally rotate it: the hook on the back of the male connector slots into the rails on the front of the female connector to make them impossible to separate without destroying a nail.)
I think the female side is right, it's the male connector that's reversed.
Not if both connectors are drawn looking top-down.
Note the ribbing and clip on the female connector, the bottom of the female connector is at the bottom of the drawing. You'd mate these by lifting the female straight up, rotating about 90 degrees along the vertical axis, and pushing down onto the male connector. And when mated that way, the labeled pin-outs line up perfectly.
Oh, you're right; it's drawn such that the business end's holes for the pins are hidden on the bottom; the wires should be coming out the top, but they're not rendered. (So, the pinout is right.)
(Edit: in my defense, the CAD model's square holes on that end are pin holes, not wire holes, which would almost certainly be wider, not square, and take up more of the space available on that side, and do on the connector I have. Hard to find a photo on Google, as nobody photographs the non-business-end…)
One sad thing with fans in PC is how bad is fan controller / SuperIO driver support in Linux. AFAIK, there is no autodetection for SuperIO chips (like there is for PCI or USB devices) and for many chips there are no drivers anyways.
Seems like using USB connected microcontroller with open-source firmware instead of embedded SuperIO seems like easiest way to have reliable fan control in PC.
Nick has been a prolific writer as long as I have known him. This is some great, esoteric stuff, about fans. I have used case fans in unusual applications not involving a computer but never went deeper than provide sufficient power that it spins full speed. For my own builds, I usually just plugged the fan in without a second thought.
What a coincidence, 4x of the 120mm Noctua Redux fans just arrived this morning and are sitting on my living room table waiting to go into my rig. I love in-depth analysis of stuff like this.
How do y'all like to manage (chassis) fan speeds? I have been out of the PC world for so long that I forget all pieces of hardware need to be managed.
FWIW I used to use software (Speedfan, and later something else I don't remember), but after upgrading to Noctua fans, I found whatever is built-in to the motherboard works just fine and gave up on the complication of software. I'm on my second system like this, and I've found they've both been a good temperature and nearly silent most of the time. Even maxed out, it still seems to be quieter and cooler than it would be with most (cheap) fans.
Dig the fact that this thing comes with little temperature probes - my current case is probably a little too small for the setup I have right now and I am debugging the temperature zones / airflow. It's a Fractal Define C, i7-12700K, Radeon RX 6750XT OC. Initially my CPU cooler fans (2x) were pointed in opposite directions, lol. Would love a thermal camera, but these probes would probably be just as helpful.
Ah reminds me of the time when I was a kid and my PC would randomly crash and would not start for a while. Turns out the processor fan was janky and would not start occasionally requiring a push with a pen or something to get it going. So the POST process in my case had an extra step: "have human check CPU fan"!
Does anyone know if there are high-heat-resistant PC fans (and at what temperatures normal PC fans will melt)? The only fans I can find that claim to be heat-resistant are the Noctua Industrial PPC series which, aside from not being cheap at $33/piece are "only" heat-resistant up to 140C and the airflow isn't crazy high - subjectively the ones I've got feel like they move substantially less air than a standard diesel heater blower fan.
Ideally I'd be interested in something that's >180C resistant but if there are any other options around say 140C max then I'd also be interested.
Unfortunately many plastics, and the most common types of solder, start softening before you get to 180C. You can get smoke extraction fans and chimney flue fans that will tolerate higher temperatures, but they're generally an all-metal affair costing more than $33.
Assuming the 180C is at your system's ventilation output, could you move your fan to the air inlet so it only has do deal with ambient temperature air? That's what's done for blacksmith's forges etc
Yes, chimney flue fans were next on my list but surprisingly enough the airflow for most of the shelf models is quite low. Unsurprisingly with all metal parts and less competition in the sector the performance is also far worse in terms of watts/CFM. Pushing ambient air sadly doesn’t work for my application. That said those would have all been great suggestions though so thanks!
Using furnace vernacular, you have forced draft fans which blow air into the furnace or induced draft fans that sucks out the hot flue gas. These fans are typically centrifugal as they develop higher pressure than an axial fan. You need the pressure to overcome the system flow impedance.
If $33 is "not cheap" then you're not going to find anything lower. Perhaps an oven fan, but that's a different form-factor. The problem is that fans with integrated motors are limited by the temperature of the motor, and motors designed for high temperatures are also not cheap since special materials need to be used.
...and what's the reason behind needing such a fan?
$33 is cheap enough for a handful for myself of course but as you may have guessed I’m thinking of releasing a product that would need them. Not a serious thing, more so a handful of bulk bought parts sold together to scratch a long standing itch in a very very specific field. Sadly $33/ea would really eat into the BOM, even on an expensive mass produced item like an iPhone that would be one of the most expensive parts.
If $33 is not cheap then maybe not. But I buy my delta fans at mouser or digikey, and they have a wider range of specs than typical consumer websites. They come without connectors though so there's some crimping/soldering required.
For our little rendering cluster I bought a bunch of 12.000 rpm fans to cool the rack mounted threadrippers. Worked fine, no watercooling in the rack needed and 4ghz on all cores under load.
For some reason I got all opinionated about fans and settled on delta as my preferred brand. I think because delta actually had data sheets. The cool thing is that digikey has the exact fans I want, the downside is that the fans digikey sells have no plugs.
I tried once to terminate it myself but I was too cheap to buy the correct tools, so that did not go well. and am now a bit bummed about the whole thing, I found a vendor I can trust to sell me real delta fans.. and they have no plugs. sigh will probably stick with nocturna as a good second best.
One interesting thing to note is the P/Q fan curve. It plots how much air a fan, at a specific PWM frequency, moves against a variable restriction (measured in pressure drop). Here's a sample - https://noctua.at/en/nf-a12x25-performance-comparison-to-nf-...
Hypothetically, knowing the heatsink airflow curve, one can find the amount of air a fan will move across.
There are PQ curves (either reported or measured) for a bunch of fans and some radiators out there.
This is great. I determined some time back the most important thing I need to know is "buy Noctua". They're expensive but have consistently outlived all other PC fans I've ever tried.
Those kinds of fans typically use a bronze bushing rather than a bearing. These bushings are long-lived, but eventually wear. When the bearing hole enlarges as part of the wear process, the fan rotor shaft now has additional room for lateral movement in the bearing and can oscillate in the the bearing rapidly, causing the "brrrrr" noise you hear.
You can often extend the life of such bearings by adding a drop of light machine oil to the bearing (usually you have to pull off a sticker on the back of the fan to access the bearing).
I'd imagine the fluid bearing gets a bit worn over time such that it's a bit "looser", i.e. doesn't hold the fan rotor as firmly in a straight/central position.
Not enough movement to sense by hand, and no roughness, but enough looseness / free-play to allow a kind of resonant vibration or "wobble" when the fan is spinning faster.
Edit: or maybe it's just very slightly worn, has debris in the bearing or something, such that it's out-of-round. I think it's the former though (resonance) as that kind of noise can be intermittent.
I haven't assembled PC in long time, by now I'd expect you can build passively cooled PC if you use PC just for some regular office job and surfing without gaming and resource heavy apps or we are still not there? Back in the days undervolting was for me for interesting than overclocking to make PC as quiet as possible.
Some passively cooled PCs exist, but they're niche and expensive.
It's a lot easier to make one that's quiet, but not silent. Big fans that spin slow and not going crazy with bleeding-edge hardware do wonders. Yeah, if you really want, you can try to underclock/undervolt some things here and there.
I think it's useful to draw a distinction between a silent computer and one that is merely inaudible. For a typical PC, all you really want is for it to not noticeably raise the noise floor of the office it occupies. Going down the route of truly silent PCs forces severe compromises, and in almost every use case it would be better to have fans, properly configured to always spin, with a minimum speed low enough to be inaudible.
It doesn't. They're almost always open collector/open drain signals, with the pull-up on the host (PC) side. No power needed, if you've got mechanical switches doing the commutating!
(Of course those aren't so reliable over the long term, so transistors are usually used. But swinging a base or gate around is all that's needed, and that is easy to do with almost any available power source. Like the main power for the fan!)
Generally the fans are powered with a steady state DC source and the PWM is just logic signaling. If the fan is powered by a PWM source the tach output will not be a clean square wave (but could still be reconstructed, if the PWM frequency is sufficiently higher frequency than the fan's rotational speed).
I can't remember the last time I encountered a motherboard that didn't offer you the choice between DC voltage and PWM for fan speed control. Both are still mainstream options; CPU heatsink fans are almost always 4-pin PWM fans but case fans are very commonly 3-pin DC fans. AIO Water cooling kits also commonly use 3-pin fans, especially when integrating their own fan controller or multiplexer so that they only need to occupy one or two fan headers on the motherboard.
(Edit: if you mean that in practice, the motherboard approximates DC voltage control using a PWM-based method, that may be true, but it's beside the point; a 3-pin fan is missing the control circuitry that a 4-pin PWM fan uses, and the 3-pin fan's behaviors like turn-on point will be specified in terms of DC voltage not PWM duty cycle, and you can't drive a 3-pin fan with the PWM signal intended for a 4-pin fan because the voltage and current supplied are both far too low.)
Just when we thought Linus Tech Tips had exhausted every normie, low effort way to "investigate" which fan is best to cool a consumer CPU and GPU for gaming...
