I Turned MIT Award Winning Toroidal Propeller Into A PC Fan | Fan Showdown S5 Special
Вставка
- Опубліковано 5 лют 2025
- MIT has just been awarded an R&D 100 award for the work they did on toroidal propellers. MIT set out to create a drone propeller with similar performance to standard propellers but while being much quieter. Well spoiler alert they nailed it, demonstrating a Toroidal propeller on a consumer drone that seemed to check all the boxes.
Well..... as soon as MIT published there findings I had tons of people reaching out to see if I had seen the study and if I could turn it into a PC fan.
Well if you know me you know I can turn anything into a PC fan, so today I take MIT award winning toroidal propeller and turn it into a PC fan.
My Twitch
/ majorhardware
Material Used Formfutura Galaxy PLA "Jupitar Brown"
amzn.to/3DuW6cd
Fan Showdown Instructions
• EVERYTHING you need to...
Fan Dims
www.thingivers...
Submit your design (at least .STL file)
fanshowdown@gmail.com
Cameras used.
BlackMagic Pocket cinema 6k pro
amzn.to/3m2RLVw
Reddit
/ fanshowdown
The fan blade has more blades, which also cuts down noise compared to a 2 blade prop. It sounds like you get a decent amount of vibration noise too, which is either because your toroidal isn't weight balanced, or you could have some structural issues. Could figure that out pretty easy in Ansys.
ayyy sup
Yes, definetly some vibration sounds occuring due to inbalance
Could some of it also be due to the fact that these are ducted? I was under the impression that ducted fans were more efficient mostly due to the reduction of those tip vortices
let the man cook
Came to the comments to say this, noticed real engineering already chimed in. 👌
Also noticed he acknowledges the same things at the end of the video.
Yes a Sharrow Prop is worth it, if you're on the water professionally, or you're a very avid fisherman. You will save enough in fuel in one year to pay for the prop twice over.
wow. that's a hell of an efficient prop then. I was wondering if the prop helped reduce fuel consumption as well.
So buying a 5k prop will save you 10k in fuel if you go out on the boat enough times in one year?
That is quite good, and if everyone did the same I wonder how fuel prices and supply would be effected?
@@Koronosavilon unfortunately it will not be affected. Boats make up a minority of fuel consumption... now, if we are talking cargo ships, we could perhaps improve cargo ship fuel economy. That would mean lower cost of imports.
@@NoobNoobNews Surely cargo ships already run the most fuel efficient props known to humanity, the engineering into making those things run cheaper is incredible!
@@NoobNoobNews I hope that you x thought that those cargo ship propeller is just as simple cheap propeller 😂…it’s more complex and far more advance with multifeatures
One very big difference is the fans you are testing are ducted. In this case the "duct" is the body of the fan. When Noctua designed their fan they designed it to work optimally within that body. The Noctua fan also has many more blades, with much less open space between them. Therefore it can push against higher static pressure at lower RPM than a typical propeller can.
Remember, the noise of the high speed propellers on a drone mostly comes from the tips of the propellers. Such propellers normally have relatively few blades, so the toroidal propeller is a closer match to that.
Further, the toroidal propeller is intended for use outside of a duct, where the airflow around and through the propeller behaves differently from how it behaves inside of a duct.
In the end, while this test was interesting, it was more like comparing apples and oranges.
If you want a more accurate comparison, consider making a fan with blades just like the Noctua, but with only 3 of them. Then test that against the toroidal propeller inside and outside of the duct.
I'm blown away nobody else saw this. The toroidal design is basically half-way between an open and a ducted fan. You're spinning 3 short ducts. So putting the three inside a larger duct is not only redundant, I suspect it detracts from the efficiency.
@@AlexKarasev
You might be right about that. That's why I suggested testing both inside and outside of a duct. The differences could be very interesting.
Exactly the point I just made with my post before reading this, LOL
LOL was about to type about the ducted configuration causes performance difference and the build purposes, glad I looked at the comment section before I went brrrrrrr.
Yep. I came to this video because of having looked at some vids about toroidal props. I like this guys style, he seems very straightforward and good testing. And super talented making it from a picture. But all that is sort of moot because the point of a toroid is to create something like what a duct does. The duct is already breaking up that turbulence at the end point of the blade. As Laughing one points out the comparison to do would be a ductless version. But in that case it's no longer a computer fan.
Have you ever tried printing the A12X25 yourself, to truly see if the differences you are measuring are due to the design, or if the print quality also plays into it? I see a lot of ripples in the surface of the prints, which to me seems like it might contribute to turbulence and noise levels. Though i don't know much about aerodynamics so might also be way off.
Yes, two years ago. ua-cam.com/video/vsqi6SP0NhE/v-deo.html
This 🔝
He did a pretty long time ago I think.
There's an episode where he did that.
As David Cheek already stated, he's done this before (I've already replied to another comment in another video actually) - he did it before the fan showdown became a thing: ua-cam.com/video/vsqi6SP0NhE/v-deo.html
A major difference is that MIT seems to be focusing on wing tip vortices, but your "tip" boundary is different. Instead of being in free air, its in a static ring. So the torodial was optimized for a different set of conditions.
Came here to say this ^ , thank you.
This is the key part. Inside the fan shroud this is really only a 3 blade fan with cool toroidal shape blades.
And possibly for a very different RPM-Range at comparable diameter.
@@miagi1337 everything here is what is stated on the video, did any of you guys watch it in full before commenting? xdxd
exactly, probably lots of shear from the outside edge of the toroidal to the wall, probably over double on the exterior since its roughly a 2-sided surface ring rotating on 1 surface of a ring.
Honestly I think the fact that you were able to match the noise level of one of the best fans in the industry with just a new shape and nothing of the polish work that a company like noctua puts in their product to make sure they are that quiet, I find very impressive!
With refinement this design could well double the operating speed and cfm of fans while keeping the noise at a more acceptable level.
Yep can't wait for a company to implement this design
Matching the noise level with lower hydrostatic pressure and airflow is a phyrric victory though... because to match the air flow you would need to double the speed, meaning a huge increase in volume. You could match the volume with any blades if you turn down the rpm enough.
It's definitely a cool result and first step, though.
5dB of difference is quite a lot, and not “matched noise levels” lol
@@DashzRight Yes, but keep in mind this is a first attempt home made fan vs the result of hundreads of tested desingns. Now it would be insteresting to see noctua pick this design up to see a real comparison.
The toroidal is designed for open air, not a duct. Also, Sharrow used to machine their props out of one huge block of stainless steel, about the size of a 5-gallon bucket. They are now working with cast stainless, to get the cost down.
Yeah lol I imagine that would make a difference
Opportunity for 3d metal printing then
I mean you could use a converging and diverging duct setup to minimize the static pressure the fan has to deal with, but that'd going to some great lengths just to use a fan not designed for the application.
I have thought of doing a small version using 3 120 fans on a 3x140 radiator but again that's going to kind of a silly length for probably minimal gain at best.
@@andyparker8631 Pretty sure that casting is wayyyy faster.
@@andyparker8631 Cast and cleaning is much cheaper/faster for production and they'll be able to drop that machine time down significantly while retaining a lot of quality of the billet part at a fraction of the cost. 3D printing has it's place but is still way out there when it comes to mass production.
I think that the Toroidal Prop isnt Quieter than the Noctua also has to do with the fact that in the PC fan mount the Fan is encased with the tips right up to the mount of the fan. That probably combats those tip vorteces wich are much more prevalent in a free Propeler like on a boat or a drone
Yeah, acts more like a ducted fan rather than an open air prop
it would be interesting to make it spin at 10k rpm.
I was about to leave this exact comment, also why modern jet airliners are much quieter than older prop aircraft!!
I was thinking this as well. Also the toroidal fan isn't designed with the housing "legs" in mind so the blade-pass-frequency could be quite bad.
I'd also say the MIT fan is probably also squishing air against the sides of the , shroud?
1:23 That amount of immediate reverse power is ungodly impressive
If you read their paper, the intent of this was to avoid adding the shroud around the blade tips (ie case-fan)... you might find that their design pushes more air than the noctua fan, if you remove the case-fan frame from both. :D
Not through a radiator or filter
@@VndNvwYvvSvv as in static pressure vs airflow?
That's what I thought. Maybe a ducted fan already limits blade tip vortices.
Good argument, but is a caseless fan still a PC fan at that point?
@@EdwardChan.999 I think that's the point, that PC fans aren't the best application for toroidal propellers, nor were they designed to be. Every prop has an application.
I think it is really tough to evaluate the MIT design unless they release models. Aerofoils and aero structures are impossible to eyeball and small changes can make significant differences.
^this^
The math behind fluid flow mechanics is fiendishly complex, and TINY adjustments have massive downstream impacts on efficiency and performance. I’d be curious to see a design that was fully optimized for the task and see how it performs.
@@munchkingod6 Yea, or at the very least smooth out the surface that's still jagged from the printing method...
yeah this is what I was thinking the entire time, the only thing this test showed was whether or not the design was viable for a fan at all not whether it can actually compete with standard bladed fans in PC cooling. a quieter drone propeller is useless if it can't keep the drone flying. granted that is the final conclusion he came to as well, so it'll be interesting to see a similar design with more stages of optimization and QC
@@munchkingod6 To be fair, he did mention that the results from MIT were based around much higher RPM.
That´s why I am witing ro a A.I. / chatGPT model / design!
I get the feeling that if I hang around this channel for a while I will turn into a fan. One way or another.
I see whatcha did there. Nice. Don't be a quiet fan, though. ;-)
Nailed it, its out of the park.. Def fan
Well, I'm a HUGE fan after watching this, and it's my first time!!!
fr
me with a fan:
I read an article about the MIT thing over the weekend and immediately thought, "Man, it'd be cool if someone made it a fan for this show." Day and a half later, you delivered.
It might have been because he mad a community post here? I saw a post on reddit about it in a sub frequented by the kind of folks who watch his channel, someone else probably saw the post and posted about the tech on reddit xD
Same! Immediately thought of MH when I saw that :D
But why? It's fun, but as he said a PC fan never spins that fast. And most importantly, the PC fan is shrouded, while a drone propeller is not.
Before I even read the article that I saw, I skimmed and saw pictures of boat props and my head immediately went to fan showdown lol then I read it and watched a clip of both a boat and a drone using variations of it... Pretty cool stuff. I always wished I could fly a drone near someone without them hearing
The noiseblocker eloop did this years ago. They also quoted the reduction of tip vortices as the main advantage.
Loop rotors where initially patented by Dr. Rudolf Bannasch in 1999, see EP1196696B1 or WO2001002742A2 (Rotor with a split rotor blade).
The patent recently expired so I'm not surprised that a lot of people are "inventing" this now.
Ahahahah MIT inventing old german technology as per usuall XD Good one for spotting that out
Underrated comment.
Idk about other videos but the one I watched they did say that’s what inspired the build and designing of these blades so he never claimed to have invented the idea just did testing to improve it.
The fan "shroud" functions to break tip vortices like the Toroidal Propeller, so it makes some sense that the Toroidal Propeller doesn't make additional improvements in a shroud. Toroidal Propeller is really a design that improves shroudless fan/propellers. As others have mentioned, more blades for the A12x25 is also likely to be better over the 3 blade design in the case of a shrouded fan for both air movement and noise.
Yeah, if the shroud was removed, and the blade angles were optimised, the Toroidal propeller (while possibly still not beating the Noctua) would still be a decent, high performing and quiet option.
I wonder how it'd go on a graphics card (or CPU) air cooler, where the fan wasn't covered by a shroud vs a dedicated aircooler with the standard fan style.
Yeah toroidal fan has no advantage inside a duct.
I think that the frame of the A12x25 helps the standard fan a lot. Spinning in free air, it probably wouldn't do nearly as well.
On the noise aspect, definitely. Was thinking this too, Noctua isn't comparable to a drone propeller if it's in it's frame. Well, not otherwise either (blade count and shape to start with), buy anyway; the frame will help with those tip vortexes.
@@Garbox80 ducted fans for drones and ducted propellers for boats exist too, so i wonder if the toroidal prop design has advantages against those beyond weighing lighter.
@@sepg5084 I mean I feel like weighing lighter is the name of the game lol.
The rough surface on the 3D printed blades (against the much smoother one on the commercial fan) are a factor that impacts negatively on the dB levels.
Maybe you could sand and polish it before measuring again. It WILL make a difference.
I think it impacts on the airflow as well but not so sure.
2000 rpm might also be too low. Smaller fans use higher RPMs. Maybe try with a 40mm high flow fan? Because from my experience, small fans can be very noisy in the higher frequencies.
agree, I'd suggest printing the propeller on a resin/SLA printer instead
but even if it dropped the noise levels by 10%, it would put it right in line with the other fan. While it is cool to think that it is as quiet as a noctua which are notoriously quiet for pc air cooling, 2k rpm just isn't fast enough to see a huge change in db.
I think a better option, would just be to print a design similar to what's found on the noctua. So they both have the same disadvantages.
@@internetcancer1672 Good idea
As others have pointed out, the vortices concept is reduced in significance when the fan is ducted and not open like a drone propeller. The blade area ratio between the two tested is also massively different, you'd see better airflow from the toroidal if the BAR were higher. Overall it's actually really promising, especially given it shifts the sound profile to lower frequencies, which are far less annoying.
The F-35 comes to mind here. It's apparently 20% louder than any of our other aircraft. It has a much deeper sound to it though, especially compared to the Harrier which it replaced. The harrier still sounds way louder and far more annoying to be around lol.
What I find fascinating is the suboptimal build quality and speed of design was able to nearly match the noise performance of the noctua fan, which is about as good as it gets. This might actually have potential with more R&D
**Edit**
Yall, I'm not a fluid dynamics expert. Idk why this comment keeps getting replies. It's an interesting topic for sure, but Idk wtf I'm talking about.
I was thinking the same thing. You just eyeball it, slap it in, without as much as treating it, much less, like, sanding it glossy, turn it on, and get a result with less high frequencies and almost the same level of noise. While Noctua are like the cutting edge engineers in the field. These blades probably went through hundreds of hours on the design boards and dozen of physical edits.
Some 600-grit sandpaper, maybe a coat of something just to stick in crevices, and a whirl at 'wheel-balancing', for instance.
No it wasn't close. Noise should be compared at same airflow and/or pressure - not RPM. Otherwise I can make a disk "fan" that's just a flat spinning round piece of plastic, and it'll be practically silent on its own but move ZERO air.
@@AlexKarasev great point
if the fans could go up about 1k rpms youd notice the benefits of the toroidal prop. it would out preform the standard fan.
This is a ducted fan (the fans are in a duct, natch) which probably affects the tip vortices quite a bit. Compare them in open flow (without the duct) instead, as well as at higher speeds as you suggest.
Dude congrats on the 2.9 million views!!! I've been following you for years and am glad to see you get a good view count!
You deserve it man!
The longer this Fan SHowdown goes on, the more convinced I am that Noctua really do know their sh*t.
anyone who sells a brown fan, clearly has their s#&t together.
I am a Noctua fan, and I can tell ya we ain't full of hot air! but quite the opposite we push cool air at components quietly and efficiently.
@@JustKanji haha noctua fan
@@JustKanji punny
@@JustKanji
Im black and homosexual transgender muslim. It's very important.
This is pretty badass as a first stab. As RE said in his comment, with additional engineering effort to tune the airfoil shape, improve fit/balance, and some configuration studies (what if you overlapped four or five toroidal rings?) you'd probably see it close the gap significantly
That's my first thought as well. Increase the fan depth to allow for over lapping rings at higher density.
Essentially, the “peak” shape just becomes sharper but there is significant loss of air pressure due to the “openness” of the shape itself. High pressure usually = loud.
You should sand the toroidal fan smooth and re-do these tests. The roughness of the print layers likely contributes a bit to airflow turbulence and a louder sound
Great video! I think one of the factor may be the number of blades. Noctua has 10 blades but Toroidal has 6 blades equivalent both spinning around the same RPM. If you can get 10 blades for Toroidal spinning as the same RP, then it will be more interesting. Additoinally, I think one design advantage of Toroidal vs the traditional drone fan is the shroud or the lack of it. Because Toroidal creates a virtual shroud around the blades by the design. But, Noctua already has a shroud. The decrease in noise will be more prevalent when there is no shroud. A well designed shroud decrease the noise for the same air flow.
Also, with this Toroidal design, the self shroud of the propeller is interfering with the shroud of the assembly, causing noise and power losses.
You can hear the toroid rattling in the shroud also wich surely accounts for some of that volume
I noticed the number of blades also. One more thing to check is how much power is being used, check both with that toroid and one with more blades, with an AC current they will tend to spin at about the same speed.
I would suspect total surface area towards the air would be the dominant factor.
The least Cd w/ the greatest surface area would win.
using that idea, increase the blade count by making each shape more oval instead of more circular. sort of like daisy petals instead of round ones. that way there's more blades actually doing the pushing. because in this layout, there arent any FULL blades, just 3 sorta full ones and 3 less than that.
I love the way you did your research. Fully scientific, with real tests, collected data and conclusions. Nice sharp video, good sound and efficiently long to keep our attention focussed.
I sent you a Möbius strip fan with three turns or something quite some time ago. Would have been cool to see if it worked. The designer name was Archer. Hope I am not annoying with all the fans I sent. Have not done many lately.
Maybe you can do a demo of it and see if it would compete.
Might not have a printer
@@di_spencer7372 I have not been to a public library in the last 2 years that doesnt have one now, but then again I have only gone to 3 different ones & they were all within 1 hour of either Milwaukee or Chicago aka not extremely rural (one was legit in a small farming town though).
@@von... lmao this city to poor to let us have nice shit
Hey there, the transition on the toroidal fan front blade to blade was quite aggressive. Smoothing that out would help with the tests quite a lot. Another thing to take into consideration is the weight, flexibility, and quality of the 3d print. A thicker print with groves from layer lines will most definitely cause some performance issues.
@MajorHardware Also, the 45° Angle of Attack is WAY too aggressive! It should be around 15° at the root and slowly decreasing towards the tip. Angling the propeller so much will cause it to stall, which will drastically reduce performance and increase drag (due to the massive amount of turbulence generated)
This is the smartest comment section I've ever seen. Cool stuff!
I'd love to see this on a drone with a sound test, see if it even works lol.
Yes yes drone test this!
The point is to reduce high frequencies that travel far. To test this the sound meter should be 10+ meters away from the fans in a stadium or other very large room.
do it!
You will see these types of propellers come out for purchase from high quality manufacturers in the next couple months.
and a set of a12x25 blades on the drone for the sound comparison lol
You are running them both in a ducted configuration. The duct on the standard fan is going to provide a lot of the same benefit for reduction of tip vortices as the toroidal fan, and is probably quite a bit more efficient (especially since you have more blades).
This is the kind of thing I’m hoping the UA-cam algorithm will recommend me because I don’t know what to search for to find it
I would love to see you scale this down to a 40mm fan and see how the noise levels compare at high rpms. I think you are probably right about the speed being a big factor in the apparent performance differences here.
I have a few 1U servers with a bunch of 40mm high speed fans, and, let me tell you, they're loud as F (especially during reboot or CPU intensive operations)
@@davestorm6718 22000 RPM goes BRRRRRRRRRRR *crackling headphone noices* RRRRRRRR
I was thinking he needed to reprint this for a high speed delta fan and retest it
@@davestorm6718 I had a system in 1U, watercooled, and use radial fans in it.
Now I got another one, and need to get some more big radial ones. Got a new gamign system in 2U and it is a lot quieter, I think if I had an easy way to mount the gpu and not needing a Protocase custom order or heavily modded case, I';d have done it in 1U again.
You can help the noise if you really want to.
The MIT design you copied only had 3 blades. The toroidal version likely performs better than a conventional 3 blade fan. In order to perform as well as the Noctua design, you would need to make a toroidal shape with an equivalent number of blades.
Definitely needed to be doubled up, the back pressure is going to punish anything that is very open
I agree. An additional lever to pull: adding depth like the Sharrow water props would also increase the surface area and probably help channel the air and keep it moving towards the output.
Quadmovr put toroidal propellers on a stunt/racing drone, and they ended up being substantially louder than his more conventional 8 blade propellers. The toroidal had more of a howl to it.
There is a drone delivery company called Zipline. On some of their drones they are running a 2 bladed propeller. The blades sit at maybe 25 or 30 degrees to each other, and on different planes on the hub. Opposite the blades is a single coin shaped counter balance on a stalk that is a bit shorter than the blades. It makes sort of a shushing noise with little, or no buzz.
Amazing experiment as always. Try to print one more fan using SLA so that you won't have grooves generated by layers of 3d printing. And measure the difference. Also, there can be slight variation in weight distribution causing nonuniform in moment of inertia, resulting in vibration. SLA method reduces this too.
It's really cool that you jumped over to the marine application. I actually came here interested in both PC fans and boat props but did not expect to get both in this video lol
In a computer case, since the path around the fan is usually quite long (from one side of the case panel to the other), the pressure gradient at the edge can't easily get around the shroud to dissipate energy in the form of a really long tip vortex. However, there is still the abrupt pressure gradient at the tip, where the flow turns very 3D, and a toroidal could soften this, but fans that need to overcome static pressure gradients inherently need all flow to be sufficiently energized by the fan to avoid local flow reversal at the shroud. Fans with lots of blades are good at evenly applying energy to the flow so that very little of it will reverse away from the increasing pressure gradient.
I can definitely see more blades improving the airflow and static pressure. The angle of the blades could also be better optimized.
One piece of the Noctua's design is the ridges on the blades that are supposed to help reduce the noise they make, so I wonder if the same kind of ridges could be added to the toroidal blades.
The fact that this quick unoptimized design appears to compete so well at all with the top-dog Noctua A12x25 I think it pretty promising.
Exactly. I would love to see like overall stats from the various ones tested. So you could theoretically then combine the advantages of the various optimal fans. Though the official one seems to be highly optimized already, while all the others are barely test of concept stage. So any doing vaguely decent against it is a huge sign towards that design being easily better, in certain ways at least, after a few refinements. Then material testing and and refinement to fix any changes due to that and you have the best min/max'd fan ever.
might also be interesting to try it out with doubling the thoroidal blades, fusing one set into the other
Except we're missing the scale. If the noctua is top dog, how much worse is average dog and bottom dog, and how does the toroidoal compare to those respectively?
Not sure how well it competes, really. It's moving half the air and making a little more noise. Not to say it couldn't get better. But it's not like it was just barely worse than the Noctua.
@@BrettBreeden Keep in mind that airflow and static pressure are not the same thing. At half the static pressure, this unoptimized design isn't something you'd want to put against a radiator, but in the smoke test, it doesn't appear that its airflow is that far from the Noctua.
Maybe Toroidal would work bit better if it had more blades also starting and ending at the middle points of the other arcs, 4 arcs would basically count as 8 blades and 8 blades would be 16 if you followed this extremely quick explanation of an idea, I'm really glad and you deserve kudos for that.
Worth noting that these props when scaled EVERYTHING changes. There's 3 semi-chaotic phenomena happening with props that if scaled, their effect on the result changes by different amounts for each phenomenon. Also the shape plays a HUGE role so 1 degree on the attack angle for example could make something 100% better or 100% worse. Prop design is hard and especially this one.
Considering then that it even came close the the noctua's fan shows how much improvement there is to be made (unless of course you hit the perfect design which is less likely than hitting the lottery but not impossible haha)
and the profile of the wings does as well change effect. I think that the lower frequencies might have come from vibrations due to unequal material thickness and/or excentricity
I really appreciate the community writing comments under this video. They all are contributing some ideas and everyone (it's most certain not everyone, but a Hughe fraction at least in the top comments) is friendly. That's what I like the most about videos like this and I think that's one reason I appreciate taking part in nerdy communities
It always brings a smile to my face :)
And I wish you a nice day
I think refining the angles, the shape of the torrid, and the trailing edge, will yield insane results. Something perfectly tuned to 2000rpm,
Yes, since you know the speed it should work you can make it work perfectly for just that speed
Yeah, I think a lower angle of attack on the blades would be optimum...say 30°or less possibly
The fact that it got relatively close to even being compared to a Noctua fan as just a first prototype, which was actually designed for that purpose and has had several years of rnd behind it, says alot. With some development, more testing, more refinement, and proper optimization I think that it could be leagues better
Static pressure has specific use cases. If you're pushing through a rad or heatsink they're good. For just moving air through the case the toroidal would be better because the decreased pressure doesn't really matter and you shift the tonal pitch to a range more comfortable for human hearing.
I would use traditional props on my water setup and then use an array of the biggest toroidal fans I could at the lowest speed that's pleasing to the ear. Lower decibels combined with lower tonality is a good formula for silent airflow. Every additional fan you add (same model/specs) will only add 3db to your overall noise output meaning 3×30db fans are better than one 50db ripper. 3×30db fans sum to 36db but will probably perform similarly or better.
I want this in 120mm+ and offered in low rpm/dB models for keeping ssd's, nvm's, ram, and board components cool. Even if they had to run at higher rpm they'd be preferable in this use case.
That's interesting - my dad did some work for GE making fans more quiet back in the late 70s/early 80s. That was before there were PCs. But he did manage to make them more quiet. I wasn't privy to the details, but I think it has to do with the angle of the blades that break up the incoming air and how it shapes the air once it's gripped it, as well as possibly creating an echo off of other blades or parts of the fan.
For an eye-balled 3D printed fan not designed to be a computer fan, it is doing really well.
I use a dust collector in my small shop. They are crazy noisy, I can't help but wonder how much quieter it would be with a better blade design and if it's significantly quieter there is probably a good market for them.
For our shop, I build a closet for our Clear View cyclone dust collector. It's built with two layers of 5/8 drywall and uses an exterior steel door with more 5/8 drywall on the back side of it. Then I built a set of baffles floor to ceiling and about 18 inches wide that just have a 3-inch wide opening that is alternated from left to right with each layer of baffling, which are 3.5 inches apart. It's quiet enough to have a conversation in the room while it runs. We have really high-end (I want to say 1 micron, but I would have to look to be sure) filters on the outlet and recirculate the air in the shop so we can easily maintain temperature and humidity. A better fan design is a great idea, but this was easily done at a low cost.
Dust collectors would use turbine/impeller type blades. The difference there would definately be interesting
@@ericapelz260 yes there are ways to work around the noise if people have the space. But there is a reason so many high end vacuums sell compared to shop vacs with a noise barrier, some people will pay for the convenience. That's one thing that bothers me about the festool groups. When a problem comes up their solution is always to buy more festool.
Clear View is an expensive brand. If there was an option to buy a quieter unit with a 20% markup I believe it would be a popular choice.
Some nice outta the box thinking!
@@Otto-W I agree a quieter unit could draw a premium. We already had the Clearview, so this was a retrofit. It could also work with a shop vac if you had enough hose to reach things. There always is more than one solution to any challenge, and each will have it's place. (FWIW, we went with Clearview because we wanted to re-circulate the air but wanted to minimize the health hazards )
I think a Thrust test like we do for drone props would be a great new test to add to a season
Lol the legit boat advice thrown in was awesome
Add six more blades to the toroidal, to match the blade number of the other fan, and you'll probably see a significant improvement in air flow by the tiroidal fan.
You might even get much better airflow by just adding three more blades to the toroidal; you could then test the six blade toroidal against the nine blade toroidal fan.
Super interesting! I think that design probably works best for applications without a shroud or duct. The blades almost form their own duct-ish thing. Just my guess. Sure looks cool, though!
Problem is unducted PC fans don't work all that great to begin with!
In PC space there have been some unducted fans, particularly the Arctic F Pro series, but they make up reduced performance by just being THICC, and also only partially. 40mm instead of the usual 25. They failed to make an impact and largely got themselves killed off. They do sound pretty nice and unobtrusive but they also didn't shift a lot of air.
Though i think with this sort of blade construction it may be worth another shot?
My thoughts exactly. The whole idea is centered around silencing non-ducted propellers. So having them in a ducted PC-fan kind of defeats the point.
The comments under the first showdown that the Cheater competed in made me aware that fan ducts aren't as common in PC cases as they used to be, despite being so obviously useful in improving airflow. I wonder if it's because they get in the way of gamers' oh-so-important lightshow. 🙄
@@stevethepocket You're thinking of the wrong kind of duct; the fan frame fully envelops the fan blades and thus makes it a "ducted propeller"; as opposed to an "open propeller" like what the original paper is aiming for for its usecase.
Ducted cooling assemblies, where you have a shaped pipe connecting some outside opening to a fan, have only ever really been a thing in mass-produced prebuilts, not in standard-parts custom-built PCs, because it turns out this is kinda impossible to standardise across so many possible part combinations. And with prebuilts leaning towards overpriced throttling garbage today, there really isn't much incentive to put thought into airflow engineering, just slap the cheapest thing on it that barely works, people who buy these things don't know any better anyway.
I would actually welcome a step back from tempered glass sidewalls and back to polycarbonate ones. For one they're pretty robust and cheaper to replace; for other you can do things with the sides - drill vent holes and such. A large downdraft cooler, side vent for it, and some DIY ducting to help it draw the advantageous outside air could be a thing. You can have aesthetics (or "aesthetics") and functional design both. This is actually about what i've been doing but i've got that old Scythe Grand Kama Cross and just steel sides. Apropos functional design vs. aesthetics, what's up with enclosures that have so much glass even covering the front, they have no chance of supplying the requisite air?
@@SianaGearz If you're saying that cases with built-in ducts aren't feasible because of the wide array of places the CPU could be positioned on the board, that may be true, but that's when you just go with what my last off-the-shelf PC did: having half the panel covered in holes and making the duct attach the heatsink instead. You'll want all those extra holes anyway so the graphics card (and any other components on the board that get hot; apparently even SSDs and RAM have heat sinks now) can get fresh air too.
Actually, depending on your computing needs, that graphics card might have _more_ use for a ducted fan than the CPU would, and for that... yeah, I've never seen anyone do that and I have no idea how it would be done.
... God I feel called out. When my dad and I pulled the boat out of the water at the end of the season last year, I forgot to lift the outboard.
Holy shit, this episode is DOPE. Keep digging around in the scientific papers and bring more of this stuff to my eyeballs.
I think the one you decided to print is the initial proto since the intern started using FDM printer for the first prototypes. The one I think you should copy is the blue prop. The one printed in SLA with the form lab.
What i like about this fan is that it looks so much easier to clean.
Fascinating; thank you so much!
*paused at 3.32*
That's a really interesting blade geometry; I can see the compromises against a traditional blade, along W several advantages...
They lose a fair bit of the end-blade area W the highest diameter, and therefore speed+swept volume; that's going to cost something...
It's going to be heavier; that will go either way depending on specific applications & part of the curve...
But the structure is inherently stronger, more rigid. The straight edges integrate some of the ducting/perimeter of forced/contained flow. And Because there are no hard blade ends & a more consistent mid-circumference optimal blade area, there's fewer vortices & potentially a more uniform performance curve...
There's also potential for a mixed curve set, and an optimised internal diameter/ external diameter to blade ratio; that mid area join zone is likely a net loss, but also potentially a gain for static pressure..
I Do think a toroidal fan optimised for this use case would be different, but W such an emerging geometry type there's no way to know how....
I have some ideas, but not enough skill to really do it; I'll try to see what I can put together & submit...
This blade lends itself to so many things.... Integrated ducting, conada effects? Soap bubble geometry... 3d blade array... Omfg...
Testing your drone with the Noctura fans would be really interesting. If a more bladed fan could reduce this noise a bit that would be fantastic
If we are talking about drone props, here is how bladenumbers affect things:
Fewer blades = more speed (due to less drag), lower efficiency, more noise.
Dual blades for highspeed builds, very loud, need beefy motors and high-C batteries.
Quad blades have less endspeed, but more bite in corners, motors can be smaller and lighter.
Tri-blades are a nice medium, both for racing and freestyle drones.
I cant figure out why all the camera drones have mostly dual blades...
my guy just hustled us to do the work for him and make a toroidal pc fan, challenge accepted!
You might not be noticing as much difference in sound due to the fact the propeller blades on the A12X25 are shrouded/ducted which reduces the tip vortices. So no real advantage to the toroidal propellers in this particular setup. If the fan blades for both were out in the open you may see a difference where the toroidal propellers are quieter.
Interesting idea. I do think a significant difference from the drone configuration is that these fans are ducted. That is going to have a massive impact on the tip vortex behavior and therefore the sound. I think it's plausible that the duct is already mostly mitigating the tip vortex induced sound anyways, thereby erasing any advantage the toroidal would have in open air. The RPM difference you pointed out does make some sense too, though to me the Reynolds number is already plenty high enough even in the pc fan situation that I don't think there would be a significant difference in terms of turbulent vortex induced sound.
I agree, that's the first thing that struck me - ducting changes the circumstances of a relatively straight blade. Essentially the toroidal blade is a halfway house between an open-ended blade and a ducted blade
"Have you seen what MITs been up to?"
Me at 5:30 am : "No. No I haven't. What are they up to?" *Clicks video*
I hope your channel gets big enough you can afford to buy your dad that prop without batting an eye mate. Love your content!
I bet adding a similar to noctua fan curve to the blade will increase performance drastically. Also you probably should try to add number of blades while making them thiner
Here's something tp think about:
The noise as you say mostly is a result of tip vortices. Notice how in high performance cars, they always have plates at the end of their wings? They help dissipate vortices faster and in some cases, reduce their generation altogether.
Now, in pc fans, the fan casing essentially acts like those end plates, hence tip vortices don't have that big an impact on noise and performance anymore
Love the Fan Showdown. I would like to know how the Noctua System works with e shape like the Noiseblock NB eLoop. These are toroidal for years
Very interesting. Considering how close some of these numbers got with just the initial design using a 3D print, I'd say it probably has quite a bit of promise. If I were doing it, I'd definitely try more blades first; probably would start by taking the existing design, duplicating and rotating it to fit between the existing blades.
Back couple years back when I was in jail, the noise on the drone would a;ways give away that a parcel was being dropped, the man dem gonna love this one.
I wonder if the tips being close to the walls of the fan are reducing tip vortices in the stock fan? I am currently designing a wind tunnel and we're using a drone propellor for the blower fan and we think we can reduce turbulence+swirl+noise by keeping the tip clearance close to the wall of a duct. That's basically what's going on with the stock fan. Drone propellers usually run in free space though, not ducted, so that might be the reason that stock+duct is quieter than toroidal+duct. The toroidal isn't getting a lot out of being in a duct from a noise reduction standpoint I imagine.
Yes, the tips have to be very close to the walls in wind tunnel fans. You can also bleed air out just near the tips, afaik, so you get cleaner air down the stream.
Yeah normally PC fans can't get their fantips close to the cage, because everything warps a little and then SCRAPE. This problem hit many PC fan manufacturers trying to optimise fan designs, indeed there are a few reject scrapey scrapey ones known, particularly as you run them up, and then the blades unfurl a little, and scrape. So they gradually upgraded their materials and casting methods and one of those accomplishments for the Noctua is that they can get in pretty close.
I think this would be much more effective with smaller 40mm fans for 1U servers where they run at much higher speeds and create much more noise.
In fact I’d love it if you could test this. I’ve been searching for years for a small 40 mm quiet fan configuration.
Noctua actually makes some of these smaller fans. They are of course not designed specifically for servers that have custom mounting options. But the NF-A4x20 is very much able to be this option.
That's such a funny thing to immediately work on.
You should 3D-print a toroidal for the drone, and a copy of the noctua for the drone, and see which one flies the best.
I would love to see a copy of the noctua fan on a drone! Especially given the pretty high volume of air it shifts, plus noctua's legendary low noise profile.
And use the drone blades for cooling the computer.
These props ARE as good as they say. BoatTest did a pretty thorough side-by-side (same exact engines, same boat, different prop), and it ended up being both more efficient (GPH) and higher speed at the same RPM. They're also more resistant to damage, because the leading edge isn't as exposed to logs and other debris, and the circular design has a tendency to bounce off debris as opposed to sinking into it.
Notice every use case scenario for it is always in the 2500-5000 RPM range? Also the wind noise can affect the decibel meter, so measure it blowing away from the mic, not towards it.
If I've understood correctly, you can simulate the turbulence of air by making the same experiment in water with slower speeds. As a result, it should follow that whatever design works in water (e.g. commerical toroidal propeller for boats) should work well in air as long as you drive it faster. If the commerical boat propeller is designed to run around 2000-4000 rpm, you might need to spin the similar design around 20000-40000 rpm in air for optimal performance.
And running anything with such rpm will require near perfect balancing and strong materials.
Reynolds number (ratio of intertial to viscous force) is length x velocity x density / viscosity. You won't get that to be the same.
Specific speed (rotational velocity x flow**0.5 / (g x head height)**0.75) might be the same.
I sense a spin off series coming called the Drone Prop Showdown.
ha, spin off
That would be great!
Get. Out!
The thickness and smoothness of the blade, and material itself, probably all detract
You should print out A12X25 blades AND toroidal ones for you drone, and see how each works in flight.
Great idea.
Personally, I'd love to find quiet drone props. Noise is the main reason I don't fly my drone very often.
It would be interesting to see if the noctua fan design would translate to drone props. Also, do you think that a more aggressive, progressive pitch on the blades would make a difference.
drone props tend to twist like actual propellers and to do the flips and acrobatics, they need to produce sufficient opposite thrust spinning backwards.
i don't think the noctua fans would work very well for that and other reasons (weight, etc) but i'd like to see somebody try, I could be wrong. ;)
the pitches on the drone props are designed for their speeds/flight envelopes, and yes, you can buy them in all sorts of sizes, pitches and blade counts. The blade pitch on a prop is analogous to a gear ratio in an auto transmission, you don't want too much or too little, depending on what/where you drive.
lolz I was just about to post about this to your channel somewhere-- ofc you're on top of it ;D
Imagine how many everyday designs are the way they are because that is how they always have been, and how much more they could be improved. We need to think outside of the box to get even more performance
This is a very good point, and if someone bothers to go to the effort, it's actually one of the best uses for current non-quite-AI: iterative improvement by taking the original, creating a good-enough computer model with a set of measurable goals, let the AI run wild for a bit, then printing the result. It's not perfect, but quite often an AI will get a crude low-hanging-fruit result that's measurably better than the current standard even before refinement just because it's not constrained by human thought boxes in the same way.
@@blumoogle2901 Genetic algorithms (a method of doing what you talk about) was one of my favourite topics in college. Took those concepts to almost all of my jobs and managed to improve something by letting the machine just go ham. Sometimes only marginally, sometimes drastically.
I would like to see you make these for your phantom drone I would def watch that
i think its more like comparing a 120mm fan to an 80mm fan, on a drone the props can be a bit longer to fit the toroidal bits in but in a PC you need to shrink the "blade" portion to fit the torroidal bits
This looks sick! Results were to be expected since there's so much space around the blades I guess. Now the next question, which of the designs can lift better?
correctomundo - different profiles for different purposes
5:34 - incorrect test. He puts usual fan to the tube and increased its performance.
Great video, thanks for your efforts! Mechanical engineer here, agree with the 'need more blades' consensus. Suggest printing a wafer-thin snap-on ring for the A12X25 to essentially tie all the blade tips together -- that *theoretically* should reduce noise with *presumably* minimal impact on performance -- but need to functionally verify. Also *much* faster to design and print a simple ring than a complete fan blade. Downside being that it might be tricky to balance, and it doesn't look like there's much clearance between the blade tips and the fan housing. You got a new subscriber!
Yeah that 5000 is tough!! It kind of goes to say what it's going to save you as far as other aspects of owning a boat though. I guess the question is, does it??
That is pretty cool and interesting! These toroidal blades have been all the rage lately :) I saw your subscriber fan blade show down as well. I did a few videos on quick ways to use fusion 360 forms to build this toroidal blade as well as some others. If there is something crazy you want to try I am happy to model it up for you to give it a shot. I think the idea of the toroidal blade inside of a ducted body will never quite match the traditional blade. They are just designed for different purposes. I would bet(not really cause i don't bet) that the closed end blades on the toroidal are hurting its performance in this case.
I think the noise problem might be cus there is so big variance in the outer diameter of the fan.
Wow, this one is very interesting. Thank you for leading me to this subject.
It seems that there is more space to improve with this subject.
My immediate concern was how few fan blades there are in the toroidal design vs the finned out noctua.
I just saw a video on the toroid prop design and love that you already had a video out using it for fans. I think the problem is blade density for lower rpm pc fan use. Try a design more like the propeller where the depth is much greater like a 3x fan depth allowing for more blades to be added closer together. Then you can try it with and without a surrounding shroud and different shroud depths.
Interesting I guess they didn't want to talk about the militarizes prop with the angled cut tips that send out counter vibration waves to cancel the initial noise wave.
First of all: imagine how much time, effort and money Noctua spent on developing those fans. Considering your design was pretty much a shot in the dark, the results are pretty amazing.
And second: The toroidal design itself is a rough principal. You can't just take a drone propellers geometry and transfer it to a PC fan. Comparing the toroidal drone and boat props you can already see significant differences. PC fans are an entirely different application again. I'm sure with a little (or a lot 😀) tuning of all the parameters (toroid geometry, number of blades, pitch angle, shroud geometry, print quality, etc.) a toroidal fan could close the gap to or even surpass the Noctua fan.
you should also consider adding another measurement factor, the V/A draw of each fan. on boats this design has a gas savings of about 20% over traditional designs. and maybe the RPM of each design as well. this might offer some interesting insights to designers.
the other thing to consider is fan mass. the "winner" looks like it might have much thinner blades, or made on resin printer vs fdm printer. the lower mass per blade and overall mass could have a huge change in results.
finally i was looking foward for this
Did you consider emailing the MIT team for some of their dimensions and ratios? Researchers can sometimes be rather free with what they learn especially if your interest is to further experiment and not commercial.
Would make for a really weird-looking helicopter wouldn't it?
Probably work in a sci-fi setting
I think that'd be pretty much impossible to have the blade angles tip to work like a heli unfortunately 🫤
@@androiduberalles, yeah helicopters need collective pitch for attitude control
@@riongronberg1435 What if the whole thing could be flexed?
I don't see how it would be possible but it's interesting to think about.
@@ddegn Maybe some pivot mechanism like this ua-cam.com/video/55G-mwT9uCE/v-deo.html
I would be super keen to see you try different blade designs on the drone actually. 3D printed drone upgrades seems awesome.
I just learned of all that’s to come and I’m looking forward to see where this leads , stronger , quieter. I’m just thinking of all the implications and applications this will improve on