I've got a small problem I could use some help with. I am using an Arduino to control a cooling system I am building for a DSLR. Primary goals are to use PWM and some pots to control TEC power, as well as fan power. There will likely be two TECs and two fans, both attached to a long 98mmx40mm heatsink. Another goal is to keep the whole cooler unit as simple as possible, so I would like to use a single power connection, and preferably a standard connection (i.e. 5.5mmx2.5mm power jack if at all possible) that is frequently used with astrophotography gear.

The problem is that if I power the fans off the same +12V as the Arduino, could the fans inject a lot of noise onto the power lines? I have done some electronics tinkering, but I've rarely used motors, so I don't know a lot.

I do know that I need a diode (I plan to use Schottky diodes, since the fans will be controlled via PWM) across the positive an negative leads to the fans in order to avoid voltage spikes when power to the fans is cut off.

I also have read that a small capacitor parallel to the diode and fan can also be used to filter the noise a bit. It sounds like the capacitor should be QUITE small...anywhere from 100pF to 10nF? I don't know if there is an easy way to determine what size capacitor should be used. I am also not sure if a simple small capacitor like that would be sufficient to filter the noise generated by the fans, especially when using an Arduino.

I know that an opto-isolator (or opto-coupler) could be used instead of a MOSFET with the PWM signal to totally isolate the fan power circuit from the Arduino. I believe that would require using a totally independent power connection for the fans, though. That goes against my design goals of standardized simplicity. So, I am wondering if there is another way to isolate the fans from the Arduino enough that an opto-isolator would work well enough. If that is not possible, would using an LC or LRC filter on the fans be enough to filter out the noise and protect the Arduino?


  • \$\begingroup\$ Is there a reason you need to control the speed of the fans? Why not run them at full-speed all the time? Then you could use a simple relay/SSR/MOSFET and even a different power source if you're worried about noise. Another test is to simply grab your oscilloscope and watch for noise on the power rails when running the fans. \$\endgroup\$ Dec 31, 2015 at 4:55
  • \$\begingroup\$ Check out this article on motor capacitors: robotics.stackexchange.com/questions/267/… \$\endgroup\$ Dec 31, 2015 at 4:57
  • \$\begingroup\$ I wish I had an oscilloscope. :P I tinker, but there are long spells between tinkering. I've never done it consistently enough where I thought I could justify an oscilloscope. The cheapest ones (and I never like to buy "cheapest") seem to be around a $100, decent ones seem to be a few hundred at least, and very good ones easily seem to top $1000. \$\endgroup\$
    – jrista
    Dec 31, 2015 at 7:54
  • \$\begingroup\$ One of my goals is to power this cooler with a single power connection. I don't know of any way of supplying the fans with a different power source without having two power connections, and I would really like to avoid that. Leaving the fans on full power is probably an option...but couldn't they still introduce noise into the system regardless? \$\endgroup\$
    – jrista
    Dec 31, 2015 at 7:55
  • \$\begingroup\$ Is motor noise as much of a problem if the motor being used is brushless? It sounds like a lot of the noise problems have to do with brushed motors, sparking, and radio frequency emissions. With a brushless motor, there wouldn't be any sparking...I assume there would still be some problems with noise, but the fans I am going to be using are pretty low power (12V 100mA tops, likely as little as 65mA) and definitely brushless... \$\endgroup\$
    – jrista
    Dec 31, 2015 at 9:43


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.