Is this normal behaviour for this fan?

Whiffle · 19 Feb 2018 at 09:51

I have just bought two new PWM fans for my server (both Arctic F, Revision 2).

One is a 120mm exhaust fan at the back and the other is a 140mm, installed as an intake.

I have plugged them both into my CPU header with a three-ways splitter lead.

When I turn the server on, the 140mm does about a half rotation but then stops again but the other two start rotating as expected.

My BIOS is set for the CPU fan to rotate at a minimum of 12.5%.

It is only when I up this to 50% that the 140mm kicks in and works fine.

Is this normal or is there a problem with the 140mm fan?

pastymuncher · 19 Feb 2018 at 10:35

It's probably sending a voltage that is too low for the fan to start. Many fans will not start below a certain voltage and it varies from fan to fan. Looking at Arctic Coolings website the F14 PWM needs 6v minimum to start and won't work at all if the PWM signal is less than 40%. See here.

chroniclard · 19 Feb 2018 at 10:39

Probably as above but some bios let you run fan calibration which will check the minimum levels a fan will run at and you can adjust fan curves accordingly. Might struggle in that there are different fans on one header.

Can you plug the 140 into a different header? As its bigger it would probably spin slower than the others.

bimbleuk · 19 Feb 2018 at 10:40

PWM especially require a minimum voltage to start and operate. What I've done in the past to stop erratic fan speeds at low voltage is to put a resistor inline so the effective voltage range is bumped up and you get better more stable fan speed control. You will loose the upper 20-30% of the fans speed range but for me that meant ~1400 RPM on the Corsair 120mm fans I tend to use which was fine as they were too noisy anyway above that.

I do use a fan controller hub though for custom curves etc.

Whiffle · 19 Feb 2018 at 11:29

Thanks very much for all that - really helpful (and good to know that the fan isn't knackered).

In a way, things have accidentally worked out pretty well on this.

My server pretty much sits doing nothing other than recording TV and streaming video to other clients - it rarely goes above a few percentage, CPU-wise.

I wasn't too sure that I even needed an in-take as the case really does not need that much cooling.

As things currently stand, therefore, the front intake can kick in in the rare circumstances that the CPU ramps up enough to warrant it (I never know what I might have my server do in the future, either).

Thanks again for your help.

doyll · 19 Feb 2018 at 11:47

If you are really using PWM control there is no way your fans are not receiving 12v constant power. PWM control is 12v power to fan housing 100% of the time. This constant 12v power is then pulsed in the fan's PWM controller by the PWM on / off signal pulse to pulse the 12v power to fan motor. With PWM control the PWM control circuit in fan sends a low voltage power to pin-4 on PWM header and the PWM controller in motherboard then pulses that power back to fan circuitry. The PWM signal pulses work backwards to fan 12v power pulses. More PWM pulses / time on give less 12v power pulses / time on. This is why when you plug a 4 pin PWM fan into a 3-pin variable voltage controlled fan header the fan will run function the same as a variable voltage fan, but if you plug a 3-pin variable voltage fan into a PWM controled 4-pin fan header the fan will only run at full speed.

Easy way to find out if your 'PWM' fan header is really PWM is plug a 3-pin fan into your splitter. If 3-pin fan speed changes speeds then your 4-pin fan header is not functioning on PWM, but is a functioning 3-pin variable voltage with no PWM control on 4th pin and variable voltage on 2nd pin.

PWM fan header pin-out is:
pin-1 = ground
pin-2 = 12 volt constant power
pin-3 = RPM signal to header
pin-4 = PWM signal power from fan, pulse control from pin-4

Variable voltage header pin-out is:"
pin-1 = ground
pin-2 = variable voltage
pin-3 = RPM signal to header

Whiffle · 1 May 2018 at 13:24

I thought would pop back on here as my setup turned out not be as good as I had hoped.

Further to the above, my front fan never comes on and the exhaust fan is definitely audible enough to be irritating - is it possible that air flow from the front of the case being blocked by the non-rotating fan?

It could be the fan themselves, but I have used Arctic before and they always seem to be rather good.

I think I need to go down a different route and have the CPU only connected to the CPU header.

I considered just plugging the two fans into the the 4-pin fan-header that is on my board but, having read the manual, it seems that only three pins are connected - what a swizz on MSI's part.

What are my options now?

It seems pointless to have bought two PWM fans and not now being able to utilise them.

doyll · 1 May 2018 at 13:43

You can run PWM fans on 3-pin variable voltage fan headers. They probably won't idle at as low a rpm as the will on PWM control, but they will run with no problems.

Whiffle · 1 May 2018 at 14:36

Thanks Soldato.

Sorry if this is a stupid question but is there any difference between connecting a PWM fan to a 4-pin header (where only three of the pins work) from connecting a 3-pin to the same header?

I assume that, fundamentally, they would work the same - the speed of both would just be controlled by voltage.

However, would the PWM still have any advantage over a 3-pin fan in this situation?

Tealc · 1 May 2018 at 14:52

In my 'very limited 'experience 4 pin PWM fans don't work as well off a variable voltage source as actual 3 pin fans.

Best to stick with PWM for PWM fans and DC voltage control for 3 pin fans, as that's what they are designed for.

PWM fans, you see, have quite different electronics than 3 pin fans.

Whiffle · 1 May 2018 at 15:06

Thanks - this is all making a lot more sense now.

I can see, now, why MSI used a 4-pin casefan header - it gives you the option of either using a 3-pin or a 4-pin PWM fan (though to the detriment of the latter).

I seem to remember that this motherboard was pretty good value, in the circumstances, so I cannot really complain.

I think I might just go back to my original 3-pin fans (they are both Sharkoon Golfballs, so even quieter than the Arctics).

One other quick question - if I now connect my two 3-pin fans to the header, which one should I use as the lead fan (i.e. the one that sends its speed back to the BIOS) - the 120mm exhaust fan or the 140mm intake?

doyll · 1 May 2018 at 15:30

Tealc said:
In my experience 4 pin PWM fans don't work as well off a variable voltage source as actual 3 pin fans.

Best to stick with PWM for PWM fans and DC voltage control for 3 pin fans, as that's what they are designed for.

PWM fans, you see, have quite different electronics than 3 pin fans.

Not doubting your word, but could you name some PWM fans that do no work well on variable voltage? I ask because all the PWM fans I've ran on variable voltage have worked fine other than some won't idle down to as low rpm (but some have idled at lower rpm). Thermalright TY-14x series, be quiet! Silent Wings 2 & 3, Phanteks, HP, MP & XP, .. and others I can't think of off the top of my head gave me no issues/problems on variable voltage. Several even say they work on variable voltage, and PWM circuitry defaults to constant power if there is no PWM control on pin-4. Thermalbench, CoolingTechnique and others used to always test PWM fans on variable voltage .. probably because they didn't have a PWM controller. Again, not doubting our word, just trying to verify where/if I'm incorrect.

Tealc · 1 May 2018 at 18:15

It was just a couple of fans, and was many moons ago. I no longer have them.

I found they didn't work too well at lower voltages, doing nothing until around 8v, where all of a sudden they'd kick into life at a relatively high rpm. Tweaking below this threshold caused it to stop almost immediately. A 3 pin fan would typically continue to spin below its starting voltage, often down to 3-4v.

Only in my experience. My experience of PWM fans is not great. I've always used 3 pin fans, except for my Thermalright TY-14x something which I've always used as a PWM fan and never stuffed DC into it.

I didn't buy many PWM fans as my motherboard of the last so many years has just the one PWM signal. No point in buying PWM fans when you don't have a way to control them, and can easily control any of my dozen or so 3 pin fans off any PWM source I take a fancy to.

Motherboard makers often install 3/4 pin connectors as opposed to 3 or 4 pin connectors so that all fan types will fit. Usually CPU headers do afford PWM control or DC control, typically selectable in BIOS.
If the header is set to DC and the profile set too low your fans will stall/not start. If on PWM then the fan will run at a relatively slow speed at idle, but may initially speed up to 100% momentarily as some controllers do on start up.

doyll · 1 May 2018 at 19:48

Makes sense. I've found most don't idle as well on variable voltage too, but some users have used variable voltage instead of their motherboard controllers gave them lower rpm with variable voltage than PWM, I assume because minimum PWM was higher % minimum. Silent Loop 3 PWM on aquaero go from 200-650rpm from 50% to 85% and from 650rpm to 1100rpm from 85% to 100% with 200rpm increase in 95%-100%, but work fine on variable voltage. http://thermalbench.com/2016/09/22/be-quiet-silent-wings-3-140-mm-fan/3/

But like I said above most PWM do not idle as low on variable voltage as they will on PWM control.