Please do a PWM fans roundup!

comomolo · Post by **comomolo** » Thu Nov 20, 2008 11:43 pm

This is a call for SPCR staff: would you guys be so kind to prepare a PWM fans roundup? Things are pretty well documented here for ordinary fans, but there's little info on the PWM kind. I'm sure many of us would appreciate it if you could add these to your recommended lists.

Thanks!

QuietOC · Post by **QuietOC** » Fri Nov 21, 2008 5:41 am

I don't know. PWM makes fans noiser, so I am not sure there is much point in SPCR testing PWM fans. I'd rather find out what motherboards/video cards have the option to use voltage modulation instead of PWM.

DaveLessnau · Post by **DaveLessnau** » Fri Nov 21, 2008 6:54 am

But, how do we KNOW that PWM makes fans noisier? There's anecdotal reports for specific fans. But, we don't know how things REALLY are. For instance, for all I can tell from the sound, my Scythe Kama PWM 120mm CPU fan could be on Mars. I hear nothing that I can ascribe specifically to it: no whirring, no clicking, no nothing. To me, it looks like PWM is becoming more and more prevalent. It would be nice to have a real comparison so we can pick the best option.

frenchie · Post by **frenchie** » Fri Nov 21, 2008 7:21 am

A review would be nice !!

But I'm wondering if the noise a PWM fan makes doesn't depend on the way the PWM is implemented by the board (quality of PWM control circutry ?). So I don't know how really relevant a review would be of it is the case.

QuietOC · Post by **QuietOC** » Fri Nov 21, 2008 7:23 am

DaveLessnau wrote:But, how do we KNOW that PWM makes fans noisier?

Physics. Voltage pulses + electric motor = sound pulses--especially with a fan that is designed to move air!

Now capacitors could smooth the pulses out, but ideally that turns out to be constant DC voltage.

PWM as a control scheme does have the avantage of being cheap and easy. Also motors start with voltage pulses better than equivalent constant DC, but the noise is unavoidable.

comomolo · Post by **comomolo** » Fri Nov 21, 2008 8:46 am

frenchie wrote:A review would be nice !!

But I'm wondering if the noise a PWM fan makes doesn't depend on the way the PWM is implemented by the board (quality of PWM control circutry ?). So I don't know how really relevant a review would be of it is the case.

Well, that's the kind of details I'd love to read at SPCR about. How and how much a board's design influences a PWM fan performance might be determined by using a few motherboards with the same fan. This is exactly the kind of thing that needs to be reviewed by professionals in order to end with speculation and anecdotal testimonies in forums.

Please MikeC?

xan_user · Post by **xan_user** » Fri Nov 21, 2008 8:59 am

QuietOC wrote:
DaveLessnau wrote:But, how do we KNOW that PWM makes fans noisier?
Physics. Voltage pulses + electric motor = sound pulses--especially with a fan that is designed to move air!

Now capacitors could smooth the pulses out, but ideally that turns out to be constant DC voltage.

PWM as a control scheme does have the avantage of being cheap and easy. Also motors start with voltage pulses better than equivalent constant DC, but the noise is unavoidable.

What would be nice is a PWM relay controller.

It would plug into the mobo header and the PSU 12v, getting PWM info and adjusting the dc output from the PSU to match the mobo's intended PWM fan output.

Then we could run multiple constant dc fans at via mobo PWM control.

Ch0z3n · Post by **Ch0z3n** » Fri Nov 21, 2008 9:32 am

My motherboard can already voltage control a fan based on temperature. At least, I think that is what you are talking about. Now, if you could just set up all the fan headers to do it based on various temperature diodes. Running them all off the CPU header wouldn't work so well since my processor goes to .8V 1ghz at idle so all the fans would almost turn off.

pony-tail · Post by **pony-tail** » Fri Nov 21, 2008 8:00 pm

Physics. Voltage pulses + electric motor = sound pulses--especially with a fan that is designed to move air!

All electric motors have pulses of current ! if you applied a fixed DC current to the armature of an electric motor it would not spin .
The issue is that firstly PWM is a sqare wave pulse ( as opposed to a sinewave signal ) and secondly that the frequency is too low - This especially in a light weight ( low mass ) means that the fan blades accelerate and decelerate , and in some fans even flex a little with each pulse . To make the fan quieter you run it slower , to do this you lower the frequency of the pulse ( the voltage remaining at 12 ) thus lowering the frequency of any buzzing or clicking putting it in the middle ( the most sensitive part ) of your hearing range .
If a fan was designed from the ground up as a pwm unit they could ( and some most likely have ) worked around the issues and a decently quiet fan would result !
Personal experience ( no scientific basis ) - It appears to me that smaller fans that require less torque are less prone to the noises caused by pwm . this could be because their operating frequencies are higher or that the required pulse is less violent - I do not know but it is someting I have noticed .

Riffer · Post by **Riffer** » Sun Nov 23, 2008 7:35 am

Interesting thread.

I have all my fans connected to my DP45SG motherboard. In addition to the PWM control for the CPU fan, there is another one for the front fan. Regardless of which fan I plug into this connector, it has a noticable whine. Plug the same fn into a three pin connector and the whine goes away. I thought it might be due to the fans connector being three pin, and I was thinking of getting a PWM fan, but this thread sheds further light. Perhaps a PMW fan will be just as bad?

cpemma · Post by **cpemma** » Sun Nov 23, 2008 9:22 am

QuietOC wrote:I don't know. PWM makes fans noiser

4-wire PWM fans use a frequency well above the audio range, there's no logical reason for the "kick" effect of traditional LF PWM to be audible (unless you're a bat) but, as with linear control, bearing and motor noise are still possible. Comments from a trusted source like SPCR (and the comments from its experienced members) would be very welcome.

Like DaveLesnau, I'm well pleased with my Scythe PWM fan but it was bought as a gamble. Let's reduce the odds on buying a lemon.

pony-tail · Post by **pony-tail** » Sun Nov 23, 2008 11:49 am

I have been using Arctic Cooling PWM fans for a year now (4pin variety) They are , for the most part decently quiet ( not Silent ) but when I get hot weather ( 38c + ) they do spin up and can be very loud . The upside is that they keep the temps down on those hot days . I have also had a Zalman NT 8700 with 4 pin pwm and it was obnoxious at even the slowest speed . I have also used PWM fan controllers which with careful choice of fans can be decently quiet . I use PWM ( and put up with it's quirks ) because I live in an area where there is a sizable temperature fluctuation and I want my system as quiet as possible but still able to stay cool in 40c+ heat and it allows the fans to run (in my case ) between 550 rpm and 1500 rpm as is required without my intervention .

Olle P · Post by **Olle P** » Mon Nov 24, 2008 7:33 am

I'd like to read some more in depth technical description of those 4-lead fans and the PWM circuitry.

I know that modern digital model railroads use PWM for speed control. The default setting use a fairly high frequency, but if you want more control and smoother run at low speed (typically for shunting operations) there's a setting with the modulation running at halved frequency.
Perhaps some similar solution is used here?

Cheers
Olle

QuietOC · Post by **QuietOC** » Mon Nov 24, 2008 8:24 am

cpemma wrote:
QuietOC wrote:I don't know. PWM makes fans noiser
4-wire PWM fans use a frequency well above the audio range, there's no logical reason for the "kick" effect of traditional LF PWM to be audible (unless you're a bat)

Since you seem to know, please let us know what frequency is used. I don't see how such high frequencies (>>20kHz) are going to allow fans to start any easier than with constant DC.

Okay, I found the frequency myself from Intel:

25kHz nominal, 21-28kHz is acceptable

Also, it seems the fan header is supplying constant 12V with the PWM signal being 5V maximum on the 4th wire. The fan must be doing its own switching---I need to look at this more.

cpemma · Post by **cpemma** » Mon Nov 24, 2008 10:57 am

QuietOC wrote:25kHz nominal, 21-28kHz is acceptable

Also, it seems the fan header is supplying constant 12V with the PWM signal being 5V maximum on the 4th wire. The fan must be doing its own switching---I need to look at this more.

The fan does indeed use that 5V signal to switch the 12V line to the motor coils, so the other fan electronics are "always on", unlike when low-frequency PWM is used to drive a 2/3-wire fan. This allows more sophisticated control of the power pulse shape, eg,

Standard wrote:...This specified minimum RPM shall be 30% of maximum RPM or less. The fan shall be able to start and run at this RPM. To allow a lower specified minimum RPM, it is acceptable to provide a higher PWM duty cycle to the fan motor for a short period of time for startup conditions. This pulse should not exceed 30% maximum RPM and should last no longer than 2 seconds.

It also means one PWM motherboard header can control several PWM fans as the 5V signal current is negligible (a molex is used for the higher current 12V line).

A few more comments comparing old & new fans here.

guises · Post by **guises** » Mon Nov 24, 2008 7:56 pm

For you people talking about the Scythe PWM fan, there is a review here:

http://www.silentpcreview.com/article83 ... l#KAMA-PWM

and it was quite positive. Though they were using a three pin, not PWM, controller for testing.

DaveLessnau · Post by **DaveLessnau** » Tue Nov 25, 2008 5:02 am

Yep. It's been a while since I did the research, but I believe that's the sole PWM fan review on the site. That review and the comments in the forums are why I chose that PWM fan. Still, I went into the process knowing I wanted a PWM fan. Having only one fan review and no comparisons is not a very comfortable way to make a choice.

Olle P · Post by **Olle P** » Tue Nov 25, 2008 7:50 am

cpemma wrote:A few more comments comparing old & new fans here.

I don't understand one of their statements (below image 6):
"To deal with this noise, the newest Analog Devices fan controllers are designed to drive the fan at a frequency of 22.5 kHz, which is outside the audible range. The external control circuit is simpler with high-frequency PWM, but it can only be used with 4-wire fans."

Although those specific controllers require 4-wire fans, I see no reason why not to run 2- or 3-wire fans at high frequency PWM as well.
I and a colleague just did a quick check using an old Papst 12V 2-wire fan. (Model 8312, 12V nominal, 6-15V rated voltage.)
With straight DC it started at ~4V.
Feeding it 32kHz PWM, 5Vpeak, it would also start at ~4Vrms, and vary its speed with the PWM.

High frequencies might cause havoc with the tach signal, but as long as you don't care about that anyway there should be no problem.

Cheers
Olle

Thomas · Post by **Thomas** » Thu Nov 27, 2008 4:45 am

I would also love to see a SPCR review - and it's an interesting question: How much does the mobo PWM controller matters? Maybe that shold be common thing to test with mobo's in the future... With one, maybe 2 standard test fan?

Regarding the switching frequency - remember when the Compact Disc arrived? It didnt took that long, before oversampling became common, and thus moving the sampling frequency far above 30 kHz.

Please Mike, that would be cool

Aris · Post by **Aris** » Thu Nov 27, 2008 11:41 pm

i just use a zalman fanmate for every system i use. Hook up all the fans in the computer (usually 1 or 2) to it, and undervolt them to my desired level. It costs like 4 bucks at newegg.

i run a fairly quick setup and i've never had a problem with overheating, even with all my fans set to around 7v. just as long as you use quality aftermarket heatsinks on your CPU/GPU and set up your case airflow correctly you really shouldnt need software to change your fan speeds during load operations, which is really the only point in using PWM.

Thomas · Post by **Thomas** » Thu Nov 27, 2008 11:59 pm

Aris wrote:i run a fairly quick setup and i've never had a problem with overheating, even with all my fans set to around 7v. just as long as you use quality aftermarket heatsinks on your CPU/GPU and set up your case airflow correctly you really shouldnt need software to change your fan speeds during load operations, which is really the only point in using PWM.

Depends... @7 volt, I find most fans noisy...

Aris · Post by **Aris** » Fri Nov 28, 2008 12:04 am

Thomas wrote:
Aris wrote:i run a fairly quick setup and i've never had a problem with overheating, even with all my fans set to around 7v. just as long as you use quality aftermarket heatsinks on your CPU/GPU and set up your case airflow correctly you really shouldnt need software to change your fan speeds during load operations, which is really the only point in using PWM.
Depends... @7 volt, I find most fans noisy...

i have a scyth s-flex 1200rpm 120mm fan set to 7 volts, 700rpm. No one i've ever met can hear it inside the case from a normal seated position with the case at eye level and 3 feet away.

the home airflow from the central a/c or the fridge in the kitchen are the loudest things in my house.

i guess maybe if you were in a soundproofed room you might be able to hear it.

Thomas · Post by **Thomas** » Fri Nov 28, 2008 1:27 am

Aris wrote:i have a scyth s-flex 1200rpm 120mm fan set to 7 volts, 700rpm. No one i've ever met can hear it inside the case from a normal seated position with the case at eye level and 3 feet away.

the home airflow from the central a/c or the fridge in the kitchen are the loudest things in my house.

i guess maybe if you were in a soundproofed room you might be able to hear it.

I guess ambient noise plays a big role here... for instance, my fridge might be quiter than yours.

I've 3 S-Flex 1200 @ sub 500 RPM's - often they're close to 400. At late night, when things are quiet and no traffic outside, they can be heard. In a P180, 1 meter away.

And maybe also fan sample variances too, one of my S-Flex's definately makes more noise, than the other 2. It's still pretty quiet, though.

Okay, I admit it; I'm picky about PC noise. And you're right, the S-Flex's at 700 are pretty quiet. But why shouldnt one aim for something quiter, if it's possible? Especially, when the technology - PWM - already's avaible at low prices? I would love to have my CPU and case exhaust fan spinning around 450 RPM's most of the time, and spin up when I load my system. No worries about hot summer or cold winter.

colm · Post by **colm** » Fri Dec 12, 2008 8:11 am

the request here is a good one. I asked about the pwm stuff several years ago and ask over and over...(not just here)

ie: I have a 3 wire 92mm that has the broadest range of rpms quiet I have never had...and it isn't even advertised on its own package to be that good. On the other hand, I have paid too much for junk with claims they are good. Even the market has the pwm robbed of truth.

My favorite ended up being guts of a favorite functioning pwm, strong case of a good bladed fan, and white lithium grease good for automotive... Make my own spacers, change the little rubber rings to tighten.

For all I know, there is no better way.
I would like to go to a 120mm but the 92mm I found for cpu (sold as a case fan) was such a rarity as a decent one, I just stay where I am...

It is to buy a case fan for pwm , cpu range (8 steps or more capable- what case fan does that? I have just one in 11 years) and the world inbetween seperating the two different worlds is not defined, but by chance of purchase.

it is more than spcrs problem, but maybe users can add some input to get a heads up.

pony-tail · Post by **pony-tail** » Sun Dec 14, 2008 12:04 am

How many brands of the 4 pin PWM fans are there available ?
In Australia have only seen Sythe and Arctic Cooling retail .
I have found 4 pin PWM fans in Machines from Delta and Sunon . The delta was OK at it's lower frequency but the Sunon sounded identical to a zalman PWM fan (obnoxious ) The Delta now resides in my Shuttle SB61 - G2 rev 4 as it was not only quieter than the original but blew more air as well - but still not SPCR quiet ( a 2500 rpm max 92 mm fan never will be )
Out of the fans I have played with ( just the ones listed above ) I like the Arctic cooling ones best but would like better - much better !

LodeHacker · Post by **LodeHacker** » Mon Dec 15, 2008 11:46 pm

I'm waiting for my Nexus PWM fanto come in the mail so even Nexus has jumped to the boat, not only Scythe and Arctic Cooling. For everyone new to PWM like myself, here are some interesting products:

Akasa PWM splitter (allows connection of multiple PWM fans utilizing a single PWM connection):
http://www.akasa.co.uk/akasa_english/sp ... _cb002.htm

ZALMAN ZM-MFC2 fan control panel (4th fan socket supports PWM):
http://www.zalman.co.kr/eng/product/Pro ... sp?Idx=209

Here's a picture of the ZALMAN ZM-MFC2 fan control panel (notice the 4th fan connection is 4-pin not 3-pin and supports PWM): http://www.abchw.com/img/other/zalman/2v.jpg

----

Anyone seen any other fan control panels supporting PWM? Also the Akasa PWM splitter is the first of the kind I've ever seen. PWM is fairly new, though companies are getting quick to the market with newer products. I'll see how the Nexus does when it arrives.

- LodeHacker-

ntavlas · Post by **ntavlas** » Tue Dec 16, 2008 1:03 am

I also think PWM fans are becoming increasingly useful. For example you can adjust the fan speed/temperature curves in the bios of many recent vid cards. Imagine a radeon 4850 with an accelero and a 12cm fan that only spins up at load. And while many motherboards can control 3pin fans they often offer a wider range when working with 4pin designs.

This technology has it`s drawbacks but automatic speed control can be very useful for those in warmer countries. That or a T-balancer...

LodeHacker · Post by **LodeHacker** » Tue Dec 16, 2008 1:26 am

ntavlas wrote:I also think PWM fans are becoming increasingly useful.

That's very true. I have built my PC myself with non-stock components and find my PC very good, but my sister had bought a Fujitsu Siemens Scaleo a while back and it has a very good feature that keeps me just wondering; on startup all fans spin up very fast and are very loud of course, but once the boot process is over they get much slower and thus the PC gets more silent. In the boot process all components will get really hot so I think the feature of my sister's scaleo with the side PWM fan is fantastic.

I will get a ZM-MFC2 at some point though, it not only looks sexy but in addition to the traditional 3-pin DC fan controlling it supports PWM which is not very common among fan controlling units.

Also the Akasa PWM spliter will be of GOOD use, because most if not all motherboards have a single PWM socket meant for the CPU cooler. Now that PWM fans are increasingly coming to the market a PWM splitter will be very welcomed in everyone's collection.

drZymo · Post by **drZymo** » Tue Feb 24, 2009 3:55 am

I am also eager to see a review of multiple PWM fans, because i'm about to begin on a small (hobby) project to build a PWM fan controller. I want to make a controller for 4 fans that can be controlled automatically with a software application instead of manual rotary buttons. This way I can control the fan speeds by a simple slider or automatically based on a programmed temperature-rpm curve. I'm building this controller to reduce the noise of all my fans, so it would be nice to know which fans produce the least amount of noise

.

ACook · Post by **ACook** » Tue Feb 24, 2009 6:31 am

if you all want pwm reviews, why not send mikec some donations with a comment for pwm fan purchase ???

everyone can specify which fans they'd like to see reviewed, mike pools the money from everyone and just gets them from any store that has them, so the fans can actually by bought atm, not just sent by manuf. for review, and bob's your uncle.