Hey everyone -

I've done some googling on this, but my google-fu is weak and I've not found the answers that explain this in such a way that I'm able to grasp it. So I apologize if this information is available elsewhere and I'm just missing it. If it is, please just link it and I'll leave you alone

I'm struggling with understanding the different ins and outs of air cooling - specifically, PWM vs. voltage control for regulating fan speed. For example, I'm looking at two different ASUS motherboards, the Maximus Gene-Z (http://www.newegg.com/Product/Product.a ... 6813131759) and the P8Z68-V PRO/GEN3 (http://www.newegg.com/Product/Product.a ... 6813131790) to go with my 2600K (or 2500K, possibly).

The Gene-Z appears to have 5(!) PWM fan controls. The P8Z68 has 3 PWM fan controls (2 for the CPU, 1 for the chassis). There is conflicting information about whether the chassis PWM fan control on the P8Z68 is actually PWM, but let's just assume that it is. On both boards, it appears that the fan control will not go below 60% of max RPM.

So if I were to get a Scythe SY1225SL12LM-P 120mm (which maxes at 1300 RPM - http://www.newegg.com/Product/Product.a ... 6835185144), I'd be looking at a minimum fan speed of 780 (1300 x 60%). I have concerns that this might result in too much noise after I put 3 in my case.

First, why would a motherboard have a lower limit on fan speed in the first place if the fan could support that low voltage?

Second, since I'm concerned about this, does it make more sense for me to get a lower max RPM fan so (say, this 800 RPM model http://www.newegg.com/Product/Product.a ... 6835185057) so that I have more downward room to play with?

Third, are all motherboards this way? Are there other boards with multiple PWM connections that don't have the 60% restriction? I'm not sold on ASUS...if MSI, Intel or someone else has a reliable motherboard with good PWM connections and modest overclockability and heat sinks that don't get in the way, that's fine with me. My issue with all the typical recommended system build guides out there is that none of them really pay too much attention to noise (especially noise at idle, which is my primary concern).

Finally, is it even worth it to go overboard and try to get all PWM fans? What are the advantages of PWM vs a typical fan with a fan control? Is it really just that you can control it via software to preset temperatures? I just don't know.

I was hoping there'd be a detailed FAQ somewhere, but if there is, I just can't find it.

Thanks!
Hausertrey

The Q-Fan BIOS controls in relation to the PWM CPU and PWM CPU-opt headers on the P8Z68 allow the setting of a minimum fan rpm down to 200 rpm or 'ignore'. If you don't want to get involved in setting profiles manually there is a choice of three pre-set fan profiles - Silent, Standard and Turbo. If you want to set a manual profile, then you can set CPU lower and upper temperatures on a 20 - 70C range, and the PWM fan duty cycle on a 20% to 100% range. When the CPU temp is 40C or less, the minimum fan duty cycle will be applied. The P8Z68 also gives you some Q-Fan BIOS controls for the chassis fan(s).

As PWM fan control is thermal, based on a combination of overall CPU temperature and individual core temps, what fan speed you actually get will depend on those factors along with ambient temperature and the effect of any heat sources within the PC itself such as graphics cards or hard drives.

I am not sure where you get '....the fan control will not go below 60% of max RPM....' from; there is no such restriction as far as I know. Clearly, what any particular duty cycle setting translates to in practice will also depend on the fan specified, as well as the temperature factors already mentioned. So for the Scythe SY1225SL12LM-P the minimum 20% duty cycle CPU fan setting would be around 250 rpm but in practice this could only be achieved at idle in a very cool room.

Generally with Asus boards, I would say it is easiest to set Q-Fan control to enabled and use the 'Silent' BIOS profile for the CPU and chassis fans. Only if this is not good enough is it IMO worth the complication of setting manual profiles.

Oh, boy. A bunch of things are getting mixed up here.

First is the basic difference between 3-pin and 4-pin fans. 4-pin fans have a built-in circuit that controls the motor speed based on a signal carried by the fourth pin (the other three are ground, +12V, and sensor). They are generically called PWM fans, because the pin-4 signal is a variable-width square wave (PWM stands for pulse width modulation) controlled by a chip on the motherboard. That chip in turn is controlled from software, either BIOS or user software such as SpeedFan. The width of the on-pulse sets the duty cycle and thus the average voltage supplied to the fan motor.

By contrast, 3-pin fans have only ground, +nV, and sense. A few years ago, high-end motherboards (including many ASUS boards) had hardware to control the average voltage of the second pin, and therefore the speed of the fan. This was typically (but not always) also PWM, which has led to much confusion. Anyway, such boards also used software to set the average voltage.

Very few recent motherboards support control of 3-pin fan headers. Those headers now always supply a constant +12V. The same applies if you plug a 3-pin fan into a 4-pin header: the fan will always receive 12V (again with a small number of exceptions, where the motherboard can control either pin 3 or pin 4, although I haven't seen any boards like that in the last couple of years).

To get fan control on the current crop of motherboards, you need to either use 4-pin fans, or some kind of external controller (eg, FanMate, NMT-3) with 3-pin fans.

Anyway, assuming you choose to use 4-pin fans, the remaining questions center around how they are actually controlled. The answer in every case is software; again either BIOS or user. The software involved has several inputs: various temperatures (CPU, GPU, motherboard, HDD, etc), and fan speed sensors. It also has targets: temperature ranges or speeds.

BIOS controls tend to be simplified, limiting their usefulness, especially when the goal is quiet as opposed to cool operation. Typical BIOS targets are a fixed minimum speed or a fixed maximum temperature, with a very small pallet of values.

By contrast, user-level fan control software such as SpeedFan allows you to set a broad range of targets (in 1% or 1 degree or 1 RPM increments) and arbitrarily link temperature inputs to fan speed outputs, for each temperature that can be sensed and each fan header that can be controlled. For example, a typical mid-range SandyBridge motherboard has two controllable 4-pin headers and a few fixed-voltage 3-pin headers; with some fans, you can daisy-chain the 4-pin wiring allowing one header to control multiple fans (all at the same speed of course).

One downside to SpeedFan (and others): the BIOS settings typically have to be "full on", which means that the fans run full tilt each time you reboot until you log in and SpeedFan starts up.

That's true of many motherboards but Asus do seem to be making concerted efforts to improve the situation. I have built systems using splitter cables to create PWM chains, but I would prefer more PWM headers on the board itself. The Asus P8Z68 board under discussion here is a step in that direction. It has two CPU headers - so push-pull CPU cooling or a CPU/exhaust combination can be had without the need or cost of a splitter.

BIOS versus software is probably more a case of personal preference, with Asus boards the supplied Fan Xpert software is pretty good. I have used SpeedFan - on some motherboards it worked very well, on some it didn't work at all. Where it didn't work the manufacturers fan control software invariably did, and equally that software relied on PWM fans in the same way that the BIOS did. I would always say check and adjust the default BIOS settings first before deploying any software solution. For example the more recent Asus motherboards have Q-Fan set to enabled in the BIOS with the Standard profile as default. This is combined with a minimum CPU speed value of 600 rpm. This is unlikely to be the quietest setting, so just changing the profile to Silent and the minimum CPU speed to 'ignore' in the BIOS could make a significant difference.

You don't always need a PWM fan to get proper control of speed. My Gigabyte has PWM options for CPU and a front PWM fan connector, but will happily drop speeds down using only 3 pin fans.

I use 3 pin fans on this board.
Using their own software you can modify the bios fan speeds, and temp control to a fine degree.
It's worth downloading board manuals you are looking at to check to see what they support.

Generally I've found PWM fans run faster at max speed than 3 pin ones from various makers, though if you adjust the bios/software you should be able to ramp them down to quiet levels.
There is no reason why regular 3 pin fans cannot meet the needs of most it's simply a case of voltage supply to the fan.

First, let me say that the responses to my question were extremely informative and helped a ton. I went back and re-read some of the sites and they made a lot more sense. And surprisingly as I did more research on it, this specific thread showed up on page 1 and page 2 of a lot of google searches (due to your replies, I'm sure).

My new understanding is that the ASUS P8Z68-V has two PWM fan headers associated with the CPU. Period. Both operate based on the temperature of the CPU, of course, and there are two because it's convenient for certain aftermarket coolers that happen to have a pair of PWM fans in a push/pull setup so the cable doesn't have to be split.

The two chassis fan headers are 4 pin, but the 4th pin is +5v. So they're not PWM, but for the other 3 pins, they can be adjusted. It is these two fan connections that have a minimum of 60% voltage.

Now, of course, what this ultimately means for me...and I again welcome your extremely informed opinions. I figure I have two options for my Antec P280 (which will arrive Monday). In both cases, I want to have modest positive pressure, if I can. I'll also always connect my rear exhaust and CPU cooler fan to the PWM CPU headers. (likely to be a Thermalright True Spirit 120, incidentally) I'll probably remove the two upper exhaust fans and cover the exhaust port.

Option 1: I mount two of the Scythe 1200 RPM fans (3-pin Scythe SY1225SL12M) and connect them to the Chassis fan on the motherboard. I run some tests and set the speed (down to that 60% limit...which should be plenty low) that best balances thermal and noise using ASUS's Fan Xpert or SpeedFan.

Option 2: Instead of the 3 pin fans, I split the 2nd CPU (PWM) connection and power the two intake fans (Scythe SY1225SL12LM-P). So the CPU 1 will power the CPU fan and rear exhaust while CPU 2 will power the intakes.

Are these pretty much my two options? Even though the HIS 6950 IceQ X's cooler is really quiet at idle (about 31 db according to TomsHardware, so I assume it's relatively cool), I suspect it's going to produce more heat than my CPU will at load. If I have my fans tied to the CPU's temperature, my GPU may get hotter more quickly than the CPU does. (which doesn't sound like a good thing) But really, I'm not going to know until I experiment and find out.

Therefore, the most likely scenario has me purchasing 3 Scythe PWM fans and optionally using 2 them in 3 pin mode so I can try both options. I'm told that at 800 RPM they're very, very quiet - so the 1300 RPM fans should be fine.

I'm sure I messed something up in here, but is my logic sound? Is that likely going to be enough air flow?

Thanks again.

(EDIT: typos)

Its weird with Asus, on my Gene Z i also have 60% restriction on Chassis fans, but i seen comments from other people on that they even can control 3pin fans below 60% with Asus AISuite / FanXpert, check **Official** ASUS ROG Maximus IV GENE-Z68 & Z68(Gen3) Owners Club, check the following posts, post #4204 and post #4156, there are others, just dont have the time to search, but check the thread.

I did had an Asus P8B WS briefly connected 6 Scythe Kama Flex PWM fans, and all were running at 450rpm or so under idle on Chassis fan connectors, i didnt toy with it too much as i had trouble with the board being stable for my server and return it in couple of days, but what im trying to say is that not all mobos are like Gene Z, or not all Gene Z have the problem with Chassis Fans.... its better to research more about this subject. I think Asus hasn't invested the time because doesn't seem a wide problem for all, but a bios revision without the 60% restriction could easily fix the problem, but then again this is just my guess.

Now i also own 2 intel Sandy bridge mobos, in mini build an H61 and one board that i was using for my server Q67 but change my mind after some issues with WHSv1 (i had a nightmare getting up my sever, but now with supermicro and xeon its alright). Now on both mobos, i have tested with Scythe Kama Flex 120mm PWM Fan (SA1225FDB12H-P), and in both mobos i can drop them down to 300rpm on case fan connectors, intel has a restriction of 30% for case fans and 20% for CPU fans, much better imo than Asus 60% policy. What i haven't seen is if he intel mobos will increase the speed as CPU or other component rise in temps, but i have huge heatsinks and the cpu temps stay very low on either case. Now on Asus both Gene Z and P8B WS the CPU temps did increase the PWM fan, i manage to get under WHSv1 for a day, and ran prime 95 on it to test for stability as i had some issues, and the kama Flex did move from 450 to almost 700rpm, this i havent seen with Intel mobos, but different bios, different policies into how the bios manages fans can justify this.

Again this are just my personal experiences, take my comments with a grain of salt, and research more about the mobos you like. Just as comment, most of the MSI mobos that i seen come with only 1 or 2 PWM fans connectors, the rest seems 3pin, weather or not you can control the 3pins (as some mobos can) idk, but just throwing it at you, so you check the mobo layouts also for fan connectors.

Now getting into the options..... i like Option 2, Gene Z can go lower than 60% in bios with the CPU headers, check post #4184, around 370rpm using Scythe Slipstreams PWM fans (which should be inaudible). So you could just get Akasa Flexa FP5 PWM 5-Way Splitter - Smart Fan Cable (AK-CBFA03-45) $7.95 + 3x Scythe SLIP-STREAM PWM SY1225SL12LM-P $9.95, the extension will not overload the mobo header as it draws power directly from the PSU with a 4pin molex, just uses the header for PWM and rpm signal (this is the way i connected my Scythe Kama Flex when i was using the Asus P8B WS as the mobo didnt have enough 4pin PWM headers for 6 fans), you can easily hide the cables on the back of the case.

That's in the UEFI, usually it is possible to go lower with Fan Xpert. A quote from the SPCR review of the Asus P8H67-I Deluxe:

Chassis fans on the Asus P8P67 can be set to 20% of max RPM with Fan Xpert, so it varies between different Asus motherboards.

Tephras - how very weird. But also great to know. That's great news.

Abula - that 5 split powered by a molex is perfect. I'm definitely getting one and trying the positive pressure in my case. I think I'm going to start with 1 exhaust and 1 intake (blocking the top vents). I'll add an intake if necessary.

Thanks again.

If you are looking for positive preasure you will need more airflow in than out, most of the time frontal fans have filters which will lower the airflow inside, so with even fans intake with a filter and the outake with just a honycomb, my guess is you will have negative preassure, this is best for you to test, but i would get 3 fans if you want to try mulitple configs and see how he temps vary, the akasa fan splitter will allow you up to 5 fans, so you have room to test a lot of different configs. Btw what case you using?

Good luck, and if you can post some pics and results of what happen with your testing.