I'm working on a project where I'd like to control the speed of an AC powered box fan using a microcontroller.

I am wondering if this can be accomplished by using a PWM output to control a solid-state relay (the same way PWM can be used to control brightness of an LED, but applied to an AC powered device). This would be controlling 120VAC power with something like a 5-12V PWM signal.

I have not worked with solid-state relays before, so I wanted to check if this was conceptually solid before diving any deeper. Alternative solutions and / or parts recommendations are also much appreciated. Thanks!

  • \$\begingroup\$ Yes, conceptually it is workable approach. Dive deeper and you shall find. Make sure SSR is zero-cross switching. An alternative is using Triac. \$\endgroup\$ – Maple Jul 10 '18 at 0:01
  • \$\begingroup\$ A box ceiling fan cannot be controlled with PWM or a SSR. They use multi pole switch and capacitors. Is that what you meant? \$\endgroup\$ – Tony Stewart Sunnyskyguy EE75 Jul 10 '18 at 1:03
  • \$\begingroup\$ I mean something like this: amazon.com/Lasko-3723-20-Inch-Premium-3-SPEED/dp/B000HHJ2I2 \$\endgroup\$ – thegrinch Jul 10 '18 at 2:40
  • \$\begingroup\$ @Tony Hmmm... I get that capacitors might be used as the standard method to control ceiling fans but why wouldn't this approach work in tandem with that? The fan should speed up when it gets AC, and not speed up when it doesn't, so switching the AC on and off with variable PWM will necessarily lead to more or less acceleration. Unless there is an electrical concept I am not understanding in which case I am happy to learn. \$\endgroup\$ – thegrinch Jul 10 '18 at 2:44
  • \$\begingroup\$ "something like this" or exactly that? The fan in the link already has 3 speed switch. You can open it up and see how they do it, which might give you a clue as to how it can be controlled. \$\endgroup\$ – Maple Jul 10 '18 at 9:20

If you pulse the supply on and off to an AC motor, then it will run slower.

However, as happens so often in technology, the problem isn't getting what you want to happen to happen, it's getting it to happen without unwanted side-effects.

There are only 100 or 120 half-cycles (depending on locality) of AC supply. If you use a zero-cross SSR and start dropping out every 3rd cycle, then as well as turning slower, the fan will also likely to emit a loud humming noise, louder than it normally does. This may or may not be acceptable to the listeners in the room. Only odd cycle-drop sequences are permissible, if you go for even sequences, you will supply net DC to the motor, which will saturate it and quickly burn it or blow a fuse.

If you use phase-shift control, then the hum it emits will acquire lots of high harmonics, much more audible and annoying than the low hum.

If you can actually chop the mains PWM fashion, then it will emit an audible whine at PWM frequency, unless you chop at a high enough frequency. Depending on the motor quality, it may get very hot with PWM, especially high frequency, remember these are built down to a price not anticipating use like this.

  • \$\begingroup\$ Okay, so it will create problems if the PWM cuts out one half of the cycle more than the other? That makes sense. Does the PWM have to be synced with AC cycle frequency, or will having a high frequency PWM signal (the phase of which relative to AC will shift over time, since the two frequencies are likely unrelated) solve that particular issue? \$\endgroup\$ – thegrinch Jul 10 '18 at 16:34
  • \$\begingroup\$ Is that what causes a cheap motor to get hot when fast PWM is used? Or is that another issue? \$\endgroup\$ – thegrinch Jul 10 '18 at 16:35
  • \$\begingroup\$ Thick laminations. A motor designed for use with raw PWM will have thin laminations, to reduce eddy current losses, which increase with frequency. \$\endgroup\$ – Neil_UK Jul 10 '18 at 16:37
  • \$\begingroup\$ And what does a phase-shift control mean for this application? \$\endgroup\$ – thegrinch Jul 10 '18 at 16:43
  • \$\begingroup\$ When I said phase shift, I meant the lighting dimmer type control, where the leading edge of conduction is delayed to control the output by and adjustable phase shift network. \$\endgroup\$ – Neil_UK Jul 10 '18 at 18:16

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.