Finally, MB's that can shut off fans?

alexh · Post by **alexh** » Fri May 14, 2010 8:01 am

Hi all,

Given the availability of very low power platforms, I thought it would be desirable in some cases to allow fans to totally stop if power levels were low enough. On some systems, when running apps that use graphics acceleration (and assuming no encryption) the CPU utilization can be around 5% or lower.

However, I did some tinkering on one of my systems, manually setting speedfan to 0% and monitoring the fan voltage. At 0%, it reads about 5V which won't stop the fans I use (Nexus). If I manually stop the blades, then the fan may not start back up until a higher voltage is applied. My MB uses the WinBond 83791D fan controller.

I'm an EE by trade and I seem to remember designing a fan controller for a telecom app many moons ago and if I recall correctly the controller PWM output was filtered to DC and biased to 5V (so the range is 5V-12V). My guess is that the 5V bias is to prevent complete fan shutdown under any conditions.

Now I see that some are able to build systems using BIOS control, that stop the fans when power consumption is low. The person tested it using hair dryer and verfied operation. The motherboard tested is a BIOSTAR MCP6P-M2 AM2 GF6150 AMD.

Did something change regarding fan controllers or am I just behind the curve? The reason I bring this up is that I have a Media Center with Antec Fusion Remote. Even with the fan control (switch) at it's lower setting, you can easily hear the fans when there is no audio. I absolutely had to configure my system to prevent wakeups in the middle of the night and this took some doing (this is a bedroom system).

http://www.ademiller.com/blogs/tech/200 ... er-server/

frenchie · Post by **frenchie** » Fri May 14, 2010 8:14 am

I had a p5n-e Asus mobo a couple of year ago, and I could control all the fans with speedfan. If I remember correctly, the non PWM fan headers allowed the fans to be completly turned off, while the PWM fan could be lowered to it's lowest setting.

alexh · Post by **alexh** » Fri May 14, 2010 8:19 am

But the questions is, is it just luck in terms of selecting the correct fan(that stops at 5V and restarts at ~7V) or is the voltage actually going to 0V. You could say that it doesn't matter but if you want to build a system that has this capability it would be nice to know beforehand otherwise you may end up with a big pile of HW.

I thinks whats happening is that the MB fan control voltage is still a min of 5V, but the fans selected turn off at 5V and restart reliably at some higher voltage. As I recall, the fan reviews mentioned something about this so perhaps I'll wade through those articles.

I don't think many low noise fans shut down at 5V because they are designed to operate at lower voltages. It's definitely the case for Nexus.

Edit: I just repeated the experiment. I set Speedfan to 0% and the voltage at the fan is 4.75V. The fan still runs at this voltage. Manually stopping the blades and releasing still reliably restarts at 4.75V. This is a Nexus 120MM.

tjtj · Post by **tjtj** » Mon May 24, 2010 10:40 am

putting a few diodes in series should probably fix this. There is a fixed voltage drop of 0.7V for a typical diode, e g three diodes should change the interval from 5-12V to 2.9-9.9V, I havent tried this myself though...

swivelguy2 · Post by **swivelguy2** » Mon May 24, 2010 10:43 am

You're an electrical engineer! Grab a transistor and a potentiometer and you should already know how to do the rest.

Klusu · Post by **Klusu** » Tue May 25, 2010 12:02 am

CPU fan header may be limited to about 5V min. Other headers often go down to 0V with Speedfan.
There is no such thing, "restart reliably at some higher voltage". You never know at what voltage.
Low noise fans are designed for lower rpm at 12V. They stop at higher voltage than fast fans (in general).