PC Fan of the Future?

baptiste writes "While we marvel at the latest CPU release or new motherboard chipset, we still put the same old fans in the neon lit cases and then complain that it is too loud. Well, maybe someone has finally come up with the next generation PC fan. Y.S. Tech has announced a new fan which is driven at the blade tips by a magnetic motor in the housing. Without the motor in the middle they claim a 30% airflow improvement meaning, potenitally, you could get the same airflow of today's fans at a lower RPM meaning less noise. They also claim the fan tips result in the bulk of fan noise. In this design they are sort of enclosed, reducing noise further. There is also a PDF datasheet available."

Comments

All I can say is

[SmittyTheBold]

It's about time! I have seen some fans that apple curved tips and other methods to reduce noise, but this sounds like something that could be even more useful.

It seems to me that these fans sould also be more reliable. You don't need the axis to be anywhere near as complicated, since all it's doing now is providing a center of rotation. The motor is less likely to burn out, and also runs at a lower speed, which is always a good thing when it comes to longevity of hardware.

Re:All I can say is

[MtViewGuy]

I am somewhat surprised that cooling fan designers haven't studied how airplane and submarine propeller blades work and recent design improvements that allow for slower turn speeds without sacrificing the amount of movement of air (or water in the case of submarine propellers).

By putting in a 10-12 blade fan with a relatively narrow center with advanced fan blade designs, a CPU or system cooling fan could easily run a lower speeds while maintaining the same CFM as airflow as regular cooling fans. Lower speeds means not only less noisy fans, but also lower power consumption, too.

interesting, I thought submersables used this

[Romancer]

interesting, I thought submersables used this already to get internal waterflow. because they couldn't have long lasting waterproof motors within the pipes, they had a fan that was driven by the outer ring so they didn't have the actual motor and electrical perts exposed to the water, I remember that they said something about the pressure and salt water being corrosive and the normal pumping systems not working well.

I'm still trying to fing an article, but it might have been in an issue of popular mechanics.

Re:Price?

[Anonymous Coward]

Since samples aren't even going to be available until April, I'm guessing they don't know what all of the manufacturing costs are going to be yet.

longer lifespan?

[NightHwk1]

Since the motor is located on the outer edges of the fan, would this have any effect on the lifespan of the fan? How about dirt accumulation?

The main problems I have with my cooling fans are balance, and dirt. When a fan has been running for a long enough time, somehow its balance shifts, and makes horrible loud noises. My server box is located in my closet, on the floor, with plenty of dust getting sucked into the case. It eventually got so bad that it caused the power supply fan to stop moving...
If this new fan design can overcome or lessen these problems, then I'll buy them as soon as they are available!

Re:longer lifespan?

[bani]

It's likely to have a much longer lifespan than your usual center mount motor fan. The bearings will be a LOT simpler, and there should be far less wear. Also, mounting the magnets on the edges means you should get a LOT more torque than a center mount motor. So even when the bearings go wonky, the magnets should still be able to spin the fan.

Why not make the fans bigger?

[phr2]

Think of the huge, slow-turning ceiling fans in "Casablanca"--well, ok, not that big.

The point is that airflow increases with rpm times the SQUARE of diameter, which means if you make the fan 2x as big, you can make it 4x slower. The original NeXT boxes used a big, slow-turning fan to keep noise down, and it was quite effective. PC's usually use 3 inch fans but with some case mods could probably be set up with 5 inch fans, which for the same airflow could quiet things down a lot.

Add to that a hard disk with a fluid bearing spindle motor (pdf) and you've got a rather civilized box on your desk. I'm using one of these drives now (Travelstar 30GN) and the difference between it and an old drive is wonderful. The quietness is like getting rid of a toothache.

Almost all Apples SILENT (Apple 2,Mac+,IIfx,iMac !

[Anonymous Coward]

Almost all Apples SILENT (Apple 2, Apple 3, Mac+, IIfx, iMac, Cube, iBook, Newton, etc.

That is almost the main engineering case design criteria and why so many fussy demanding people like artists, musicians, scientists (chemists, biologists) and even programmers continue to support apple products.

PeeCees are like load aircraft carriers or at least Vaccuum cleaners comapred to the majority of the most popular selling Apple Computers.

Did you know the Apple 2 had NO FAN at all?

Same with the Mac Plus... no FAN at all.

and most all the products since the IIfx have a variable speed fan that increases rotation only if truly needed... like the Wallstreet G3 laptops.

When Steve jobs created the NeXT workstation he designed every cable to be one long 10 foot single cable so that the machine (almost quiet) could be hidden insode a closet or put very far from your monitor! He hates noise of all kinds.

Hurray for quiet or semi-quiet hardware.

Pure PR

[Anonymous Coward]

There may be good economic reasons to put the motor parts on the OD of the fan and that is good. But, all statements about the air flow and cooling capabilities are pure PR.

First, the velocity of the blades near the hub are too slow to put work into the air and therefore the air tends to leak backwards near the hub. This is why putting the motor in the hub area does not degrade fan performance.

Second, while the fastest parts of the fan (the blade tips) make most of the fan noise, changing their shape does the most to reduce the noise. The noise has nothing to do with the "motor parts" that reside in the blade tips in this design.

Third, the pressure generated by the fan is solely dependant on the fan design and has nothing to do with how the fan is powered.

Fourth, of course the torque is more effective at a larger radius, but this has nothing to do with efficiency.

Fifth, any increase in air flow has to do with the shape of the blades and the speed of the fan and can probably be improved a somewhat by blocking the hub a little.

Sixth, I do not see how the fan improves the efficiency of cooling anything. Increased air flow will cool anything better, according to how much horsepower do you want to put into your fan. And, it matters a whole lot where the parts to be cooled are in relation to the exit air stream, etc.etc.etc. Nothing but PR here.

Re:Pure PR

[brad3378]

Whoa man!
looks like both of you guys should be yanking your old Heat Transfer, Accoustics, & Fluids books off the shelf and plan on doing a little review.

I can debunk almost all of (both of) your claims.
Firstly, you both argue about the significance of Airflow around the hub. BFD. Airflow is not everything!! Heat Transfer is everything.

There are 3 types of Heat transfer.

Radiative heat transfer: this is what makes your black colored car much hotter than the surrounding summer breeze. In this case it is insignificant compared to Conduction & Convection. (therefore it wouldn't matter if your fan was white or black)

Conduction heat transfer This is the movement of heat from within ONE Material. In other words, put just an edge of your frying pan on the stove and the rest of the pan will eventually get hot. The main application of Conduction heat transfer occurs within the heat sink.

Convection heat transfer This is the movement of heat between 2 materials. Convection occurs between the processor & Heatsink, Conduction then moves the heat to the other side of the heatsink, and then Convection transfers heat from the fins of the heatsink to the air. Sound simple? It's not.

Since There is a heatsink that acts as a buffer to the heat transfer, I argue that the 2nd A.C.'s claim is misleading. Yes, if we mounted a fan directly to the processor it would likely help to aim airflow at the center. In that case it might be nice to have a hub-less fan. Since only fools would run without a heatsink, it doesn't really matter much.

Remember our lesson on Conduction and Convection?
Well in our case, conduction moves heat away from the core (the center) and does it much more efficiently than convection. One reason for this is related to Fluids.

In fluids you learned that as a fluid (in our case Air is the fluid) moves around an object, a boundary layer occurs. For Maximum Heat transfer, This is a Bad Thing. We want to introduce a "micro turbulence" that will break up the boundary layer. The great engineering problem to solve is to find the best comprimise between airflow, turbulence, and noise.

So what's the best solution?
I doubt that there is a single solution for all computer users, but here is what I like:

http://www.procooling.com/articles/html/quiet_fa ns _-_power_control_met.shtml

It's a circuit that will Automatically
adjust the speed of your fan to maintain an appropirate temperature. Unlike adding a Rheostat, it doesn't create a larger current draw, and you don't have to continually adjust it.

For those of you that would like to learn more about heatsink/fan benchmarks check out this site:
http://www.dansdata.com/coolercomp.htm

Re:Price?

[cybermage]

I might spend as much as $1 more for this new fan design, but I wouldn't spend twice as much.

I'd spend twice as much if it'll last twice as long. Maybe three times as much if it'll cut the ambient noise level down by 10 decibels or more.

Re:Almost all Apples SILENT (Apple 2,Mac+,IIfx,iMa

[TaliesinWI]

Don't I recall lots if not all of those early computers having an external AC/DC converter, either as a wall wart or as a box midway up the power cable? IOW, your primary source of heat was moved AWAY from the computer case, and since 6502s really aren't known for their heat problems, of course you wouldn't need a fan. The PC has a switching power supply because it was far more expandible than anything else out there at the time, and the power requirements needed to be more flexible. This was not limited to the IBM - my TI-994/A's external expansion box (circa 1979) had a little power supply in it (don't recall if it had a fan, I think it did) because again, I could put a variety of peripherals in there, and it needed to be able to run them all.
Yet, my base TI-994/A ran off of a wall wart.

Recall that until around the 486 DX/2 days, there (in general application) wasn't anything hot enough inside the PC case to warrant its own fan. The fan behind the power supply is/was always meant to cool the *power supply* anyway.

Flame exchange.

[Erris]

Well well, looks like I've been trolled by two bogus AC's. Well the first made more sense. Let's start with the second's rather ugly start: Glad you're not an engineer.

He made more sense then you do. You: I thought of this concept a while back, even started modelling it in Simulink (part of Matlab).

Nice work, drawing pictures lends credibility. Nope. I'm a Mechanical Engineer. I've done some rudimentary fan designs (you know vector diagrams, work balance, that kind of thing), but I won't play a fan designer on TV or stomp reasonable people. Him: "First, the velocity of the blades near the hub are too slow to put work into the air and therefore the air tends to leak backwards near the hub. This is why putting the motor in the hub area does not degrade fan performance."

You: First of all, the hub is smaller, meaning the there is more air intake near the hub...anything which calls for a greater air debit at the top of the fan is good. The air will only 'leak back' if the air pressure under th fan is greater than on top, which doesn't happen until you hit high RPM's. What you will get is vortices under the fan near the hub, which is good...it means a lot of airflow over the die of the chip, which leads to cooling of the chip. Note that this happens under the fan, which is where you do want vortices.

He's right, again. If your fan does not give you any pressure increase, you have a poor difuser or a poor fan or both. There is zero motion at the axis. The best thing to do with that air is to redirect it outward towards your blades. If your hub does not do this, you might get some recirculation there that will do what your hub should have. Him again: "Second, while the fastest parts of the fan (the blade tips) make most of the fan noise, changing their shape does the most to reduce the noise. The noise has nothing to do with the "motor parts" that reside in the blade tips in this design."

You: The thing is, the tips of the blades are in the ring surrounding the fan...now the air first gets led through a bounding area before it gets spun by the fan...it goes from vertical to horizontal directly, like in any good pump, without having that airflow distorted by the boundary conditions created by the tips of the fan being near but not touching the side of the fan case. Vortices on top of the fan only make for a decrease in airintake which is bad for cooling.

Huh? What does that have to do with what he said about decreasing noise by changing blade shape? Any axial fan will create an axial vortex above it unless you put stator blades in the way. If you want to get rid of the blade tip vorticies you change their shape, or you could put a ring around them but that makes other problems. Him: "Third, the pressure generated by the fan is solely dependant on the fan design and has nothing to do with how the fan is powered."

You, again: But here the method of powering the fan has led to an improvement in fan design (more air intake due to higher area of the fanblades).

You are starting to repeat yourself, which would be OK if you were correct or even consistent. You know, he's right again. While you fail to actually contradict him, you do contradict yourself by not bothering to say anything about pressure differences that you don't understand very well. Him: "Fourth, of course the torque is more effective at a larger radius, but this has nothing to do with efficiency."

You: Thiuss hasd no bearing on the situation other than stating that the larger a fan is, the better it works...duh, we already knew that.

Ummm, that's not what he said, but it's nice to see that you have the spirit to be condesending. Can you explain why a larger fan works better for us? Is it because the ends may be turning faster? Isn't this why axial turbines are generally rows of blades mounted further away from the axis than they are long? Would they be more efficient with a large hole in the center? Him again: "Fifth, any increase in air flow has to do with the shape of the blades and the speed of the fan and can probably be improved a somewhat by blocking the hub a little."

You, sigh: No, that would only limit the area of air-intake, which is bad...you want a big area of intake, to get as much air down the fan as possible. Blocking the hub is a Bad Idea(tm)

Do you know any other tune than, "this thing rocks because it has a smaller hub"? I think I've alrady stated the purpose of the hub. Go figure. Him: "Sixth, I do not see how the fan improves the efficiency of cooling anything. Increased air flow will cool anything better, according to how much horsepower do you want to put into your fan. And, it matters a whole lot where the parts to be cooled are in relation to the exit air stream, etc.etc.etc."

You: Yeah, you got the idea...this new design improves airflow, which is good!

Same tune. You know what is good but not how to achieve it. He does. But wait, you have a conclusion: Not PR, just airflow...they increase the airintake (by having a smaller hub, which also means better cooling near the centre of the fan) while lowering vortices near the airintake (by having a seamless fancase-fan interface), thus creating better cooling

OK. Does anyone have any numbers around here? Like flow vrs current for various radiuses? Yes that would be useful, a repeatable experiment comparing various available designs. The data sheet is slashdotted. Perhaps you can CAD up some drawings of a test bench for us?