my mother is heating her house with a air water heatpump and radiators. To further reduce the flow temperature I am building a radiator booster to force the convection. The commercial radiator boosters are expensive and not very intelligent.
I ordered 8 Arctic P8 PWM computer fans (4-pin, 12V) for each radiator. My working prototype consists of a XIAO ESP32C3 board and one Arctic P8 fan. The fan got powered by a 12V DC power supply. The microcontroller board was powered by a 5V USB-C power supply. Ground was connected together, and I connected one pin to the pwm pin of the fan. Everything worked fine and I was able to control the fan speed, turn it off, etc.
But when I connected four more fans to my prototype it stopped immediately. The microcontroller board is completely dead. The fans have a connector to daisy chain them so the wiring was fine.
Do you have any idea why it broke?
Could there have been to much current draw from the pwm pin of the microcontroller?
The datasheet says maximum IO-pin source current is 40 mA. I could not find any information about the fan regarding pwm current draw.
When a pin outputs a PWM signal it does so by turning the output on and off rapidly so the current that it can supply is the same as if it were used as a digital pin
How much current does the PWM input of the fan take and what voltage is it specified to work at, bearing in mind that the ESP32 will be outputting about 3.3V ?
In addition to that, the ESP32 board will have a 3.3V regulator on board that can only supply a certain amount of current, which you may have exceeded. The fact that the board is completely dead is an indication that this might be the problem
Thank you for your answer!
I could not find any specification regarding the pwm pin of the fans. But I found out it is 5 mA for Noctua fans, so I assume it will be more or less the same for my Arctic fans. So 5 fans x 5 mA should be theoretically in the specification of my microcontroller board.
The microcontroller is completely dead, I disconnected everything and tried to power it with the USB-C power supply without success.
Should I use something like a mosfet for the pwm pin? I would really like to control all fans with just one PWM pin because it makes wiring easier and there is no reason to control the fans individually.
From the Noctua white paper it looks right to me.
(Remember this is just the current for the PWM signal, not the fan motor)
However, the white paper shows the control circuit inside the fan is just a transistor and needs to have the current going to that transistor limited by a resistor.
See page 3
In their example, the diagram could be misleading since the fan essentially wants an open collector circuit but the "customer circuit" is creating half of it.
i.e. the customer side does the pullup and then the yanking to ground so current when the voltage is high is current limited.
You can't create a signal like that by directly using a PWM output of a microcontroller pin.
What voltage do you measure on the PWM pin with the fan powered, but no connected to the EPS32? The paper linked by @bperrybap has the following regarding the PWM pin:
If the pull-up is to 5V, the combine current from five fans could be enough to damage the ESP32.
When I connected to extra four fans the first fan was active at full speed, 12 V were connected to the fans, and 5 V were connected to the ESP32 board through USB-C. I just plugged the extra four fans to the first one and it immediately died. It is not possible to connect the fans in a wrong way because of the connectors.
I could measure 4,8 V between ground and fan PWM pin, while 12 V were applied to power and ground.
My first measurement was wrong so I edited this answer.
5V pullup is not compatible with your ESP, and might damage it. But I somehow expect some other things got wrong as well.
Can it anyhow be possible that you had one of those fans GND disconnected when failure happened?
@jim-p you are Correct in that you can't assume the Noctua is the same as the Artic fan, But....
that isn't what your comment was about when you slammed the OP.
In response to the OP saying he couldn't find any info on the Artic PWM current draw but found some info about the Noctua:
But I found out it is 5 mA for Noctua
You said:
That is wrong or your were looking at the wrong specification.
I was trying to defend the OP by pointing out that you were the one that was wrong by pointing to a white paper by Noctua that said the PWM input consumed 5ma just like the OP said.