I've read that you can connect Arduino PWM pins directly to the 5v PWM connection of 12v computer fans, as long as the fans are powered by a separate 12v source, and grounds are tied. But then I also read that the power output of the pins, both individually and combined are limited.
From what I've been able to find, is Noctua 5v and 12v fans use a maximum 5mA for the PWM signal. I'm planning on 4x fans on 2x PWM channels, 2 on each heatsink. So, a maximum 10mA per channel, and 20mA combined. Plus 2x DallasTemperature DS18B20 sensors, a 16x2 LCD and a DS3231 RTC module.
I'm guessing that's not too much, but want to make sure. And make sure it'll work. :-p
Probably a good idea to connect a flyback diode to each fan (across 12V and ground, cathode to 12V) and a 0.1uF noise/interference capacitor for each fan (across 12V and ground).
Thank you, pretty sure I've got those lying around. I originally considered using 5v variation of the 12v fan, but it's input current is 150mA per fan. Obviously too much. But I could simply tap into the 5v feed before it passes through the Arduino. It would actually make things simpler. I'd still use the diodes and caps though.
You understood wrong.
YOU have to provide that pull up current.
A10k resistor between Ardiuino PWM pin and Arduino VCC could be enough (0.5mA).
A PWM frequency of about 25 KHz is recommended ( intel specification), but the default ~1KHz of pin 5,6 of a Nano might also work.
Leo..
It's literally in Noctua's PWM white paper, and everything else I've found while searching. The signal still requires some kind of current level, and it's maximum for 5v and 12v fans is 5mA, at 5v.
Before your topic becomes datasheet porn, just give it a go.
Otherwise I invite you to tear down the fan, so we can see how the driver circuit inside is made
Long time ago I played with one and I noticed absolutely no difference between pwm from 5V MCU and pwm from 3.3V MCU. Neither I noticed any difference between 490Hz pwm and >20kHz pwm. So I expect the driver just senses duty cycle and then drives the fan.
I think people are getting confused between the RPM output, and the PWM input.
There is nothing special about the PWM input, it is specifically designed to be driven by a 5V GPIO output pin. An arduino can easily provide the current.
Although the Intel spec says a nominal 25kHz is required, I have found that fans work fine with the default Arduino PWM frequency.
I had to google what "short circuit current" meant...
Imax=5mA:
This indicates that the device can safely handle a 5mA current flowing into it during a short circuit, but exceeding this value could lead to damage or malfunction, according to Noctua [1] and bleepitybloopity.com [3].
So... how much current is a single fan drawing from the Arduino pin then?
~25KHz is choosen in the Intel computer fan specification, because it'd outside our hearing range, so you don't hear the whiningof the PWM frequency.
30Hz is commonly used for 2-wire fans, for the same reason.
Leo..
My point was, that I didn't notice any difference (4wire fan) between those frequencies.
Maybe Noctua drives the fan at static X khz independently from PWM signal frequency.
The external PWM frequency works AFAIK just as an on/off switch for the internal 3-phase control of the BLDC motor.
You should be able to faintly hear the 490Hz/1KHz PWM frequency, but only at certain speeds. I guess about 50% PWM would be the loudest.
Noctua fans are already very quiet, so maybe try with another brand.
Leo..
