Expected voltage when probing a computer PWM-fan socket's sense pin?

Question

I'm using my multimeter to probe a proprietary 6 pin socket on the motherboard of a PC computer that connects to 2 small 25 mm fans. The socket's pins are unlabeled. Both the motherboard and socket are non-standard, proprietary designs.

The computer's BIOS allows independent PWM control of each of these fans, and is able to show the current speed of each fan.

With the fans disconnected, I have found:

• pin 1 is ground.
• pin 2 is 12 V.
• pin 3 and 5 measure ~4.9 V
• pin 4 and 6 measure a voltage that varies between ~ 1 and 10 V as the computer boots.

From this I guess that pins 4 and 6 are the PWM pins, and the varying voltage is due to the varying PWM duty cycle.

Further, I would guess that pins 3 and 5 are the sense/tach pins, however why would I measure 4.9 V and not 0 V (because the fans are not connected)?

Is this a consequence of typical circuits used for sensing PWM PC fan speed (that the motherboard pulls the pin high to 5 V until driven low by the fan) and if so, is a “high” for the sense signal of 12 V PC fans always 5 V instead of VCC/12 V?

Answer 1

Most fans have an open-collector tachometer signal that is pulled up by external circuitry to 3.3V or 5V or whatever you need since many fans are powered from 12 or 24V which is does not work well with most digital logic. The fan will pull the line low a certain number of times per revolution, usually 2 or 4. Your board appears to be pulling it up to 5V. It appears as 4.9 because there is probably a resistor or some other current limiting device inline since the fan presents a low impedance path to ground when it is in the middle of a pulse.