"Actual" Voltage and fan speed--YMMV ???

Tad G · Post by **Tad G** » Tue Sep 07, 2004 12:49 pm

Reading a previous post from Russ made me think of this:

For this example (only--NOT REAL #'s):

You have a 120mm Papst rear case fan that you set at 1,200 rpms or 9v in "free air" with the side panel removed from your case. When you reattach the side panel to your case the rpms drop to 1,150.

So is the Papst fan really operating at 9v? Or something else? What is SPCR protocol on this fact of fans operating at higher rpms in "free air" than in actual case application?

Which rpm #'s are Mike C, Ralf, Russ, Ed, etc quoting to the readership?

Many thanks.

Post by **Rusty075** » Tue Sep 07, 2004 1:07 pm

If the voltage is quoted in a review, that's the number that you should go by, since it is usually verified by a measurement tool. (like a multimeter)

RPM's are always context specific, and shouldn't be relied upon as a way of measuring your voltage. There's too many variables to do that: back pressure (as you've noted), but also mounting, and just normal variations in the fans themselves.

Besides, you shouldn't worry too much about the RPM's, since they're only a secondary factor, and don't really directly affect anything.

BTW, what post got you thinking anyway? It doesn't ring any bells. (but I am getting old...according to some on this forum.

)

peteamer · Post by **peteamer** » Tue Sep 07, 2004 2:15 pm

Tad G, If you 'supply' the fan with #V in free air it will be supplied with #V at all times,... even if you stop the blades with a pencil (Fingers not a good idea

).

R.P.M. may/will change... but the #V wil not.

The R.P.M.'s of any fans will depend on the ability of the case to allow airflow and many other things.

R.P.M.'s are not hugely, if at all important, 1R.P.M. is fine if your temps are to your satisfaction. R.P.M. as a value has no 'real' meaning in terms of fans/cooling. (Some trucks/lorries have 800HP+ but my car with many less is faster/quicker, if you see what I mean. )

Russ, anyone who suggests your getting/are old

should have their posts deleted
and if they persist they should be banned.

Pete

Edited for danm tipos. !!

Tad G · Post by **Tad G** » Tue Sep 07, 2004 2:17 pm

Thanks.

Ok, if with a given fan, set at 8v (with a multimeter as I always do) in "open air" the rpms are 1,200 per Intel Active Monitor, and you close up the case....and you now see 1,150 rpms...is the fan actually operating at the preset 8v, or less?

I am sorry to reply I do not specifically recall which post was yours----but I read 100 or more SPCR posts over the Labor Day Holiday...

I liked your reply as to how fans respond in quiet air v actual performance , and that is where you were telling someone how a panaflo "in clean air" performs vs. another brand...as I (think ?) I recall. Does this help you remember?

Here's my specifics, if i set a Papst 120mm at lets say(for purposes of discussion only) 8v with my meter. On the Intel monitor, I read 1,200 rpms........
When I close up the case, I "see" its actual performance as less than what I thought I'd set.

So, is the Papst 120mm fan actually running at 8v, but "laboring" rpm-wise due to case pressures????

Thanks again.

Tad G · Post by **Tad G** » Tue Sep 07, 2004 2:20 pm

Thanks peteamer too!

silvervarg · Post by **silvervarg** » Tue Sep 07, 2004 10:44 pm

So, is the Papst 120mm fan actually running at 8v, but "laboring" rpm-wise due to case pressures????

Your fan still gets 8V, but since it has to fight some pressure in the case its RPM will drop slightly. Lets assume the 8V fan is an exhaust case fan, and you put in a high speed intake fan, then the 8V fan is likely to speed up the 50 RPM that it lost due to case pressure.

On the other hand it is hardly worth the extra noise from the intake fan to get that extra 50 RPM...

sthayashi · Post by **sthayashi** » Wed Sep 08, 2004 8:15 am

Keep in mind that RPM reading units aren't all that sophisticated. Although I don't know for certain, I'd bet that most motherboards use a low-pass filter and an ADC to determine fan speed.

Plus also keep in mind the resolution size of your sample rate. Say you're measuring the fan speed once a second. That means for 1200RPM, you're only getting 20 Revolutions per second. What if you get 19 revolutions in that second? By my math, it's 1140 RPM.

So somewhere between the above two answers is where the truth lies

Gooserider · Post by **Gooserider** » Wed Sep 08, 2004 1:20 pm

Tad G:
Ok, if with a given fan, set at 8v (with a multimeter as I always do) in "open air" the rpms are 1,200 per Intel Active Monitor, and you close up the case....and you now see 1,150 rpms...is the fan actually operating at the preset 8v, or less?

I agre with the previous posters, the voltage will stay the same, regardless of what the fan speed does.

OTOH, I somewhat disagree with the notion that RPM isn't important. Airflow is the key item of interest! I could use the same logic that says RPM's don't matter to say that voltage doesn't matter. Isn't what we really care about is that our fans move enough air to provide adequate cooling, not what voltage they are at? IMHO since airflow is a direct function of fan RPM, knowing what speed the fan is turning will give a possibly better indication of airflow than voltage. In addition, as fans get old and cruddy, they tend to slow down, (Just like Rusty has??

) so a fan that moved plenty of air when new might not when it gets older and its RPM reading drops, even though it's voltage hasn't changed. Monitoring the RPM change could alert one to performance drops before they get to be a problem.

sthayashi:
Keep in mind that RPM reading units aren't all that sophisticated. Although I don't know for certain, I'd bet that most motherboards use a low-pass filter and an ADC to determine fan speed.

Plus also keep in mind the resolution size of your sample rate. Say you're measuring the fan speed once a second. That means for 1200RPM, you're only getting 20 Revolutions per second. What if you get 19 revolutions in that second? By my math, it's 1140 RPM.

I like your logic on the fan speed display changes that makes sense.

OTOH, I don't think you have the right idea on how the sensor circuitry would work. The fan is putting out a sort of digital pulse already, so I would expect that the reader would simply clean up the pulse to whatever degree it was necessary and then feed the result to a counter - no need to do an ADC step since we already have the desired digital signal. Perhaps FanControl can offer his expertise on this.

Gooserider

Gooserider · Post by **Gooserider** » Wed Sep 08, 2004 1:28 pm

Tad G:
Ok, if with a given fan, set at 8v (with a multimeter as I always do) in "open air" the rpms are 1,200 per Intel Active Monitor, and you close up the case....and you now see 1,150 rpms...is the fan actually operating at the preset 8v, or less?

I agre with the previous posters, the voltage will stay the same, regardless of what the fan speed does.

OTOH, I somewhat disagree with the notion that RPM isn't important. Airflow is the key item of interest! I could use the same logic that says RPM's don't matter to say that voltage doesn't matter. Isn't what we really care about is that our fans move enough air to provide adequate cooling, not what voltage they are at? IMHO since airflow is a direct function of fan RPM, knowing what speed the fan is turning will give a possibly better indication of airflow than voltage. In addition, as fans get old and cruddy, they tend to slow down, (Just like Rusty has??

) so a fan that moved plenty of air when new might not when it gets older and its RPM reading drops, even though it's voltage hasn't changed. Monitoring the RPM change could alert one to performance drops before they get to be a problem.

sthayashi:
Keep in mind that RPM reading units aren't all that sophisticated. Although I don't know for certain, I'd bet that most motherboards use a low-pass filter and an ADC to determine fan speed.

Plus also keep in mind the resolution size of your sample rate. Say you're measuring the fan speed once a second. That means for 1200RPM, you're only getting 20 Revolutions per second. What if you get 19 revolutions in that second? By my math, it's 1140 RPM.

I like your logic on the fan speed display changes that makes sense.

OTOH, I don't think you have the right idea on how the sensor circuitry would work. The fan is putting out a sort of digital pulse already, so I would expect that the reader would simply clean up the pulse to whatever degree it was necessary and then feed the result to a counter - no need to do an ADC step since we already have the desired digital signal. Perhaps FanControl can offer his expertise on this.

Gooserider

peteamer · Post by **peteamer** » Wed Sep 08, 2004 3:21 pm

Isn't what we really care about is that our fans move enough air to provide adequate cooling,

is all that really matters, ... with the exception of noise.

R.P.M. & Voltage are actually immaterial. They really don't matter and you might as well say "My fan moves X C.F.M at 3 cheese and 2 onions.

The voltage and R.P.M. don't actualy account for one iota. It is possible to have a fan designed with blades that will make little noise at very high R.P.M. and still deliver movement of the medium required.

Boats/ships/submarines don't suffer from the antiquated pathetic design of blade shapes that the comp. cooling industry/market does.

is an example of the blades in a cheap air con unit. (Note the shape of the lower most blade ) The blades move anti clockwise and are Very Very quiet especially considering how much air it moves, we see no such 'innovation' in blade design.
Not even innovation.... ancient technology actually

.

In fact with such a blade design you can fly in the face of tradition and use a four blade design that is still very quiet and smooth sounding...

This particular unit uses the water it condenses by spraying it in an atomised form over the cooling fins to further aid the efficiency of the radiator... much like rally (Subaru road~) cars... (can you guess my next (watercooling) project ?

)

It doesn't matter how quiet an XYZ computer fan is, ... the blades are still very 'square' to the rotation direction and will create more noise than necassary. Stealth submarines especially would laugh at the noise signatures of the blade designs we use.

Talk to anybody with a knowledge of propeller design and they wil laugh at what we have to suffer.

R.P.M. and Voltage are not our main enemy... very poor design of fan blades is.....

R.P.M. & voltage are just numbers... as is noise... but noise is the only one that actually matters to us....

Perhaps slightly off topic... but maybe not... Discuss............

Pete

Post by **Ralf Hutter** » Thu Sep 09, 2004 6:54 am

peteamer wrote:Boats/ships/submarines don't suffer from the antiquated pathetic design of blade shapes that the comp. cooling industry/market does.

That's why I've always been intrigued by the design of the fan blades on the 120mm OEM Panaflos. Definitely the most unusual computer fan blades I've ever seen.

OEM on left, normal 120mm L1A on right:

DrCR · Post by **DrCR** » Fri Sep 10, 2004 9:08 pm

Speaking of boat blade designs... (apparently someone was running a little out of control). lol

Seriously though, would be nice if some PC company started licensing naval prop tech.

DrCR

_________

HammerSandwich · Post by **HammerSandwich** » Fri Sep 10, 2004 9:42 pm

Like Goose, I find RPM a more useful tool for comparison. I'm sure many people here do not have a voltmeter. OTOH, almost everyone's mobo can report RPM. Finally, I expect RPM vs. voltage varies more than cooling performance/noise vs. RPM.

Gooserider · Post by **Gooserider** » Tue Sep 14, 2004 1:44 pm

Before everyone jumps on the 'naval propellor' design as evidence of bad or antiquated design on the part of the manufacturers, remember that we are dealing with two DIFFERENT mediums. Air does NOT have the same properties as water, and a design that works well in water will NOT work the same way in air.

Fan makers do compete, and noise is important to them, though possibly less so than it is to us. If they could design a quieter fan blade, they would. (Even industrial applications have noise specs, and fans have a big impact on designing stuff, a quieter fan would be a BIG competetive advantage.)

Gooserider

icancam · Post by **icancam** » Tue Sep 14, 2004 2:16 pm

DrCR wrote:Speaking of boat blade designs... (apparently someone was running a little out of control). lol

Seriously though, would be nice if some PC company started licensing naval prop tech.

DrCR
________

If I'm not mistaken, the propellers on that beached vessel are a very specialized variant of naval propulsion know as "surface piercing." They are partially immersed with one half of the propeller arc out in the air at all times in order to lessen the drag caused by the higher viscosity of water. They are designed for high speed planing vessels.

Post by **MikeC** » Tue Sep 14, 2004 3:47 pm

All I really want say here is that $$$ is behind just about everything.

Those stealth sub props. What do you figure the Navy paid for each one? A billion? Two? $500 mil? How many were made? 50? 100? 20?

I know of a good quality DC fan, direct from the manufacturer in lots of 10,000 or more, that costs something like $1.40 ea. for a 120mm model. 5 or 10 cents more and they lose the sale.

The cost of developing a better blade design? hmmmm.

Let see...

1) What works at 80mm dia does not necessarily work at 120mm (recalled from dreaded fluid mechanics course at univ decades ago). Probably, what works for 25mm depth does not work the same way for 15mm or 38mm. So if you wanted to develop a line, you'd probably have to R&D for each size.

2) You have to play with available materials and their applicability for each size -- a story from "Dorothy" the UK fan specialist is that the Panaflo 80mm fan blades design is truly unique in its streamlined aerodynamics, but it could not be replicated in the 120mm version because the larger size required much greater thickness for adequate strength.

3) If you paid 2 engineers to work on this with the requisite tools (wind tunnel, anechoic chamber, all the other tools...) and assistance for 6 months, would it be enough to develop a truly superior design? One year? What would it cost? $200,000? $300,000? How long would it take to recoup the $? Can you double the price? How much $ was spent by YSTech on their TMD Fan? IS it superior? (To answer my own question, the last one -- not really, not in terms of cfm/noise -- maybe for cfm/size)

All I am saying is that what drives innovation is profit, and it seems to me that the per unit price of DC fans is just too low for real R&D & innovation to pay for itself. Hence, we get same old same old.

Someone earlier made the comment that fans could be made to achieve high CFM w/o turbulence noise. I don't believe that is true. Even nature cannot create high airflow without making a lot of turbulence noise; what makes you think we can? Just try to blow a decent amount of air with your own lips, mouth & lungs -- can you do it with less noise than a really good fan 80~120mm fan? Probably not or at least not much better.

And air is NOT at all the same medium as water -- as someone already pointed out. Those super quiet sub propellers are probably spinning slower than any DC fan used in PCs.

IMO, high precision manufacturing can make a huge difference. Cherry picked Panaflos are still amazingly quiet at 7V and less, and do blow enough air to be useful when mated with appropriate HS. If every one could be like that...

icancam · Post by **icancam** » Tue Sep 14, 2004 4:56 pm

I agree completely with your observations, Mike. It does all come down to the bottom line but only until something manifestly better comes along for the same price. That general scenario has been repeated scores of times with one of the best examples being the American car industry that had to wake up from its second rate quality torpor when confronted by the better Japanese products. Now American cars are so much better in every way.

Admittedly, there is a world of difference between a little computer fan and an automobile, but I'd expect that there would be a ready market for a higher efficiency fan assuming that the price was competitive. The cost of computing is now reaching the point where complex fluid dynamic calculations can be done on a powerful desktop. There are so many competitors in the computer parts arena that all of them who want to survive and prosper would be happy to point out even a small but real incremental performance edge over their competition. With new materials technology breakthroughs allied with ever more sophisticated modelling programs and precision computer aided manufacturing, it would not surprise me that incremental improvements will be both possible and implemented even in simple inexpensive computer fans. Many enterprises would happily purchase computer equipment whose fans consumed a little less electricity (because they were more efficient at moving air). Indirect costs such as electrical use are a big issue at many companies. If as a by-product there is a trickle-down of quieter fans for our little sub-culture, then I say hurrah!

icancam · Post by **icancam** » Tue Sep 14, 2004 6:43 pm

I did some googling with the search object 'low noise fan technology' and found many links supporting the premise that there is a tremendous amount of activity and interest despite the low profit margins that no doubt exist. Here's a link with many references to low noise fans: http://www.electronicstalk.com/indexes/ ... wseef.html

One company claims to have invented a design (being implemented by Panasonic and NMB) with a 40-50% increase of air flow at equal noise levels compared to other computer fans. http://home.att.net/~alemont/noise1.html

I could not find anything at Panasonic's web site to substantiate their claim but there is an article among those listed above stating that NMB now uses the technology:

"NMB Minebea has developed a novel range of low-noise fans. In almost all new developments for home and office applications the noise of the fan can cause a great deal of annoyance. Ever higher packing densities and higher processor speeds result in high thermal losses, requiring cooling devices that would (ideally) generate no noise. Unfortunately the noise of a fan is traditionally linked to its speed performance. But now NMB Minebea has developed special low-noise fans in industry-standard 60 x 60 and 92 x 92mm sizes. The casing of the fan has tiny slots that damp the noise generated from the edges of the impeller. The difference between a standard fan and the low noise counterpart is almost 4dB. Another effect is that the pressure stability is much better. The fan's characteristic is similar to a diagonal fan with its large working range."

It would be interesting to put this "almost 4dB" claim to the SPCR test!

Post by **MikeC** » Wed Sep 15, 2004 9:25 am

icancam --

Spurred by your last post, I followed up on your research...

The first item was designed by the guy/company that produced the Millenium Glaciator, a company which is no longer active, I believe. Lemont Aircraft, a maker of helicopters or helicopter propellers (?) is where this comes from -- the HSF was a hobby that grew into something more for a while.

It's called augmented fan (or something like that); both Ralf Hutter and I have the Panaflo implementation of this design. Basically, slots are cut around the circumference of the frame, and the turbulence normally created by the outermost edges of the blades is converted into inflow. It may make a difference at full speed, but not at the reduced RPMs that we use fdans at. (Judging by comparisons at 5~7V against normal Panaflos on a couple HS.) I believe Panaflo cites the same noise level, but a 10-20% increased CFM at 12V.

Ditto the NMB models, which I just looked into. Found 60mm and 92mm versions. They cite 26cfm/25 dBA for one non-slotted model; 31cfm/25 dBA for the equivalent slotted one. But again, I think this design requires higher velocity at the blade tips to really make any difference.

Anther aspect of fan noise, aside from the air turbulence, is the drive system, which includes the bearings and the motor. Much attn is paid to bearings -- hence hydroflow, sintec, and other various air/sleeve/ball hybrids -- but not much attn is given to the actual motor.

In the Recommended Fans page, I cite JMC's explanation of the reason for fan buzzing:

"In traditional fans, the most dominant acoustic noise is due to the fan's motor switching noise. The stator motion is a square wave that is switched on and off before and after the peak torque position. This motion causes a small amount of undulation in motor torque, producing an audible noise caused by the lower frequency commutation operation. Each small torque causes a minute contracting of the entire fan structure and results in an audible clicking noise while the fan is operating."

What happens when drive voltage is lowered is that the frequency of the switching drops. It may be inaudible at 12V when the switching speed is at say 8kHz, but at less than half the RPM, it may be down smack in the midrange where our hearing is most sensitive; low switching speed may = clicking.

JMC has a strong incentive to point this out: They have a line of fans (used to be called Panther) with switching freq. at over 20kHz; this means that switching noise is inaudible at 12V, and still very high in freq at reduced speed so clicking is pretty much eliminated.

I have some early samples which seem to be click free but still exhibit some other noises that keep them from challenging the very quietest Panaflo, Nexus and other fans. But they've been adopted as OEM gear by big PC makers such as Dell, Gateway, HP... who are probably using them in some of their systems right now. As it has been a year since I received those early samples, I should probably look for a new batch; there may be improvements.

icancam · Post by **icancam** » Wed Sep 15, 2004 10:17 am

Thanks, Mike!

Yes, it makes sense that a helicopter designer would come up with a fan design that has superior efficiency at higher revolutions since lift is the all important criteria in flight conditions. Too bad that it does not (yet?) confer its benefits at lower speeds.

I did run across the JMC Panther information and recalled that you had been expecting some test samples. There have also been references made in other posts to Sunon's magnetic levitation bearing technology.

Somewhere in the world a manufacturer is going to bring together a synergistic mix of inspired design and superior execution at a great price and will gain an all important competitive edge. The word will get around!

We know that many manufacturers read SPCR to get ideas so would you:
Please accept the low noise efficiency challenge and become the new benchmark for computer fans!