March 19, 201115 yr I'm looking to replace a 92mm fan on the back of my Norco backplane to reduce noise and improve airflow in my system. I have a Supermicro X8SIL motherboard which has connections for 5 fans onboard, of which I'm only using one for the CPU fan. Are there any pros/cons to one or the other? I was considering the following two fans: Non-PWM fan: Noctua NF-B9 http://www.amazon.com/Noctua-NF-B9-Case-fan-92/dp/B0014I9K30/ref=pd_rhf_p_t_1 PWM fan: Cooler Master BladeMaster 92mm PWM http://www.amazon.com/Cooler-Master-BladeMaster-Silent-R4-BM9S-28PK-R0/dp/tech-data/B0034AFDL4/ref=de_a_smtd Or certainly value any other suggestions... Thank you very much!
March 19, 201115 yr At that size there's not much to choose in terms of cooling effectiveness. Some PWM control systems may allow the fan to run slower if temperatures permit, but there's no rule one way or the other. The main reason that PWM has been introduced is that voltage control from the motherboard becomes a problem with larger more powerful fans that result in more heat being dissipated in the control circuits. PWM, by contrast, switches the fan electronics on and off at high speed and there is very little power lost. From that point of view, PWM fans are more efficient. (edited for typos)
March 19, 201115 yr Given those two options I think it really depends on which matters more to you, airflow or noise, and how these options stack up to what you are replacing. From the NF-B9 and CM Blademaster specs Model PWM CFM Noise ~~~~~~~~~~~~~~~ ~~~ ~~~~~~~~~ ~~~~~~~~~ NF-B9 N 24.2-37.8 7.9-17.6 CM Blademaster Y 15.7-54.8 17.0-35.0 The Norco will give you good flow while being quiet. The CM will give you better flow if needed but will be much louder. The CM at its slowest speed is as loud as the Norco at its fastest speed. PWM gives the motherboard the ability to dynamically adjust the fans speed based on system temperature readings while the Norco requires changing a pigtail with a series resistor to adjust between its three speeds (you could also rig up a potentiometer and control this yourself). edit: this is a moot point since the motherboard should be able to adjust the supply voltage and change the fan speeds. One other consideration is the bearings and longevity. The CM uses a sleeve bearing rated at 40,000 hours (4.5 years). The Norco uses a Self-stabilizing oil-pressure bearing rated at >150,000 hours (17 years) though the warranty is 6 years so I wouldn't expect them to last longer than that. All bearings will get louder as they wear, but this seems to be more noticeable in sleeve bearings than other types. My personal recommendation is the Norco. I've been very impressed with their performance and quietness since I bought them and unless budget constraints get in the way I favor them. In this scenario, you're talking about $7 so it shouldn't break the bank.
March 20, 201115 yr this is a moot point since the motherboard should be able to adjust the supply voltage and change the fan speeds. No. This is a common misconception. Mobos do not vary the voltage... they use PWM on pin 2 to vary fan speed for 2 or 3 pin fans, or send a PWM clock signal on pin 4 for 4-pin PWM fans.
March 20, 201115 yr I would get the PWM fan and use the script from this post to control your fan speed based on the HDD temps in your array. I just started controlling my fans using this method it works very good. It's better than having the MB do it because the MB does not adjust the fan speed based on the HDD temps instead it uses some system temp on the MB.
March 20, 201115 yr Author My personal recommendation is the Norco. I've been very impressed with their performance and quietness since I bought them and unless budget constraints get in the way I favor them. In this scenario, you're talking about $7 so it shouldn't break the bank. Thank you for your very thoughtful and well researched answer... it is very much appreciated! I would get the PWM fan and use the script from this post to control your fan speed based on the HDD temps in your array. This is a very useful script - thank you for posting it here!
March 20, 201115 yr this is a moot point since the motherboard should be able to adjust the supply voltage and change the fan speeds. No. This is a common misconception. Mobos do not vary the voltage... they use PWM on pin 2 to vary fan speed for 2 or 3 pin fans, or send a PWM clock signal on pin 4 for 4-pin PWM fans. Well, when you PWM a DC voltage you vary the average voltage so the motherboard does vary the voltage on 3-pin connectors. Peter
March 20, 201115 yr Well, when you PWM a DC voltage you vary the average voltage so the motherboard does vary the voltage on 3-pin connectors. Nope. It is 12v on, and then off, in pulses. 12v on for 20ms, then off for 20ms, does not give you an average voltage of 6, except statistically.
March 20, 201115 yr Well, when you PWM a DC voltage you vary the average voltage so the motherboard does vary the voltage on 3-pin connectors. Nope. It is 12v on, and then off, in pulses. 12v on for 20ms, then off for 20ms, does not give you an average voltage of 6, except statistically. You REALLY need to learn what average means. Sorry to burst your bubble, but pulse width modulation (PWM) is used on the motherboard to vary the average fan voltage on the 3-pin fan connectors. Your example is a 50% duty cycle. 12V x 50% + 0V x 50% = 6V average. This is simple math and has nothing to do with statistics. FYI, you will often find the BIOS has 2 settings for the CPU fan speed control, PWM or voltage. The voltage setting is to allow speed control on 3-pin fans connected to the 4-pin header via PWM voltage control done on the motherboard. The PWM mode provides full-voltage to the fan (pins #1 and #2) and a PWM signal on the 4th wire (blue). The fan uses the PWM signal to do the speed control internally. Peter
March 20, 201115 yr You REALLY need to learn what average means. I do. As my previous post said, it is a statical measurement. You have to be VERY careful applying statistical formulas to electro-mechanical devices. For example, if you have a fuse that blows at 10 volts, but you feed it an "average" of 6 volts by using 12 volts on for 10 seconds, then 0 volts for 10 seconds, you blow the fuse with your "average" 6 volts. 12 volt PWM 20ms on, and 20ms off, is indeed, a statistical average of 6 volts.... but this theoretical 6 volts means nothing to the fan. What does your "average" tell you if it was a 40 second wavelength rather than a 40 msec wavelength? Gee... the average is the same so the fan must behave the same? How about 60 volts for 20ms, then off for 80ms? Same average. A fan will run faster on a steady DC 6 volts, than it will on the above 50% PWM square wave. PWM is not a method of varying voltage to some intermediate or average level between 12 and 0 volts. It is a method of adjusting fan speed. FYI, you will often find the BIOS has 2 settings for the CPU fan speed control, PWM or voltage. The voltage setting is to allow speed control on 3-pin fans connected to the 4-pin header via PWM voltage control done on the motherboard. That is not correct. PWM is not limited to 3-pin fans on a 4-pin connector. Mobos do PWM 2 ways, as the BIOS option suggests. One way is PWM on pin 2, and this works on 2-pin, 3-pin, and 4-ping fans, and it works on 2-pin, 3-pin, and 4-pin fan headers, if the mobo is so designed (my Asus, Biostar, and GByte do pin-2 PWM on both 3-pin and 4-pin headers.) The other is steady 12v on pin 2, and the PWM setting signal on pin 4 so the fan does PWM internally. The benefit of the 4-pin arrangement is you get a predictable performance curve between 20% and 80% PWM input.
March 20, 201115 yr Since it's maybe not obvious, you have to use PWM correctly to get a useful result. Giving a bunch of dumb examples of PWM being used incorrectly doesn't prove anything. To use PWM correctly, the pulse period must be much shorter than the time constant of the load and the correct supply voltage has to be used. FYI, the typical sensor IC used in most fans has a voltage ratng of 16V so providing 60V to it would be dumb unless you want to destroy it. 12 volt PWM 20ms on, and 20ms off, is indeed, a statistical average of 6 volts.... but this theoretical 6 volts means nothing to the fan. PWM is not a method of varying voltage to some intermediate or average level between 12 and 0 volts. It is a method of adjusting fan speed. You do seem to be contradicting yourself. The PWM does "indeed" produce a varying voltage. And the average voltage does mean something to the fan, otherwise the fan speed would not vary as the PWM duty cycle or average voltage varied. Second sentence - http://en.wikipedia.org/wiki/Pulse-width_modulation Right after the bold - http://www.ecircuitcenter.com/Circuits/pwm/pwm.htm Now, if you are implying that the PWM or average voltage does not directly set the fan speed then you are correct about that. I really can't tell if you are meaning this or not. However, this has nothing to do to with the whole PWM does not produce an average voltage discussion. All the 3-pins fans I have opened were about the same. They all had a hall effect latch driving the 2 motor coils. Varying the average voltage varied the power to the motor which produces different fan speeds. The fan blades require a certain amount of power to turn a certain speed. So, if you vary the motor power then you vary the fan speed. The only thing to know is that the relationship is not linear, for example 75% power is not 75% speed and 50% power is not 50% speed. FYI, you will often find the BIOS has 2 settings for the CPU fan speed control, PWM or voltage. The voltage setting is to allow speed control on 3-pin fans connected to the 4-pin header via PWM voltage control done on the motherboard. That is not correct. PWM is not limited to 3-pin fans on a 4-pin connector. Mobos do PWM 2 ways, as the BIOS option suggests. One way is PWM on pin 2, and this works on 2-pin, 3-pin, and 4-ping fans, and it works on 2-pin, 3-pin, and 4-pin fan headers, if the mobo is so designed (my Asus, Biostar, and GByte do pin-2 PWM on both 3-pin and 4-pin headers.) The other is steady 12v on pin 2, and the PWM setting signal on pin 4 so the fan does PWM internally. The benefit of the 4-pin arrangement is you get a predictable performance curve between 20% and 80% PWM input. Wow, argumentative aren't we? You are just expanding on what I posted. Yet I see I am wrong and you are right. Odd. Unlike my post, you are actually wrong about one thing. Most if not all 4-pins fans run at minimum speeds when connected to a 3-pin header. So, you can't claim that you can run a 4-pin fan using PWM on pin 2 of a 3-pin header. You might be able to slow it below minimum speed but it will never go faster than minimum speed. You are also being argumentative about the fan connectors. The only place I've ever seen a 2-pin fan or connector was for something like an included chipset cooler, used on say a northbridge or a video card. If you read what I posted, not once did I post that you couldn't use a 2-pin fan. But, since you brought it up, I'd love to see all the prime examples of 2-pin fans that are suitable for CPU or case use. Peter
Archived
This topic is now archived and is closed to further replies.