What type of motor AC fans use and how can I control them using an MCU?

[ivan747]

Hi guys, quick question. What type of motor is the AC motor used in domestic fans? Both ceiling fans and standing/tabletop fans.

I need a way to change the speed of those with a microcontroller. I guess my only option is an opto-isolator detectying zero crossings and having an opto-triac triggered after a certain phase delay. WIll that method produce humming? I suspect it does.

Thanks a lot.
-Ivan

[Paul Moir]

Mostly shaded pole I think.

Commercial (fairly cheap) ceiling fan speed controllers operate essentially like a light dimmer.  I don't know why you would get the micro in the actual AC zero point control loop; why not just make a speed control (or hack apart a commercial one), then control it's operator with the micro?  There was someone trying to do something similar here on the Forum for a soft start incandescent, you could just replace the pot with an LDR/LED optocoupler for example with very little effort.

PS - for applications on random fans, there's a risk that running them too slow will overheat the motor. 

[ivan747]

Thanks for the reply.

Well, I wanted to be more or less formal since I would be working on someone else's college project. He's a programmer, so I get to do the hardware. It involves computer control of an AC fan.

I first thought of just slapping together a tiny hobby servo to a commercial light dimmer, but that's not a professional solution.

Edit: searching for "shaded pole motor speed control" delivered interesting results, thanks.

[Paul Moir]

Oh, I get it - that makes sense now.   

I wonder if there's something interesting you can do with firing the fan more intelligently then; turn it on for a few phases, etc for quieter operation or starting torque.  If that's the case that's the way to go, a zero point (or any handy point) detector feeding the MCU and back again into a TRIAC.  Give the software the most freedom, and some safety with the optocouplers.

[ivan747]

Yes. The prototype will probably be an Arduino with an appropriate shield to communicate to the computer wirelessly and a custom board with the opto-couplers. It will probably use a standard wall-wart internally to make things simple. I want a simple, not too expensive, quick prototype that doesn't look slapped together.

[pickle9000]

A servo to control a switch is not unprecedented about 10 years ago a retrofit furnace control did just that. Moved the actual thermostat lever on the wall. It was billed as a 5 minute diy project for the homeowner. It's not a fail it's a feature!

[orion242]

I would not even mess with this in a homebrew fashion.

For small shaded pole motors use this.

http://controlresources.com/ac-fan-control-ac-motor-control-smartfan-nimbus/

Seems to me they are under $30usd.  Control speed with 0-10vdc you could pwm from MCU, set speed via dip switch, or via direct input from a thermistor.  This is UL listed and piss cheap.

I have used this on large 115vac fish tank pumps and exhaust fans, can't argue with its simplicity no issues.

[ivan747]

A servo to control a switch is not unprecedented about 10 years ago a retrofit furnace control did just that. Moved the actual thermostat lever on the wall. It was billed as a 5 minute diy project for the homeowner. It's not a fail it's a feature!

If it seems stupid but it works, it ain't stupid. I just don't want people to hear a servo moving inside the black box. And it's kind of embarrassing to present a project with a servo moving a pot. You can surely get away ith it in a retrofit though.

I would not even mess with this in a homebrew fashion.

For small shaded pole motors use this.

http://controlresources.com/ac-fan-control-ac-motor-control-smartfan-nimbus/

Seems to me they are under $30usd.  Control speed with 0-10vdc you could pwm from MCU, set speed via dip switch, or via direct input from a thermistor.  This is UL listed and piss cheap.

I have used this on large 115vac fish tank pumps and exhaust fans, can't argue with its simplicity no issues.

That's brilliant!

Perfect for my prototype. As I said before, it ticks all the boxes: quick, simple and cheap. I could even do away with the Arduino as the wireless solution I was about to use has some basic I/O, I could implement something simple to get 0-10V with 6 to 8 bit resolution.

Edit: uhm, I hope they sell them in 1-of quantities. They are talking about 50 piece orders and such. I think my "clients" want to sell some units, but we need a prototype.

[pickle9000]

Sorry, I didn't intend for you to use just saying that it has been done. The same can be said about any design it can be bad or good it just depends on the application.

[orion242]

They will sell in singles or its available online for not alot more in singles from memory.

If you have trouble finding it, shoot me a pm.

[NiHaoMike]

Most modern fans use PSC induction motors. You can get pretty good results phase angle controlling them, but a VFD is really the proper way to do it.

[ivan747]

I should be able to control both PSC and shaded pole with phase angle control. Maybe it's not the optimal solution but is definitely the cheapest. What I can't do is start them with a very low speed setting because they might not start rotating and might overheat. ALso I should keep the minimum speed relatively high, so I don't burn the coils.

[orion242]

The nimbus starts the fan at full speed or something like 70% at first to get it spinning before it turns over speed control to the inputs.

[ivan747]

Do domestic fans use TRIAC speed control for their typical 3 speed or 5 speed selectors?

The other day we installed a KDK ceiling fan and wow, the speed control was incredibly light, which makes me suspect it is solid state.

[orion242]

Speed selectors like your talking about are typically using multi-tap motors and nothing more than dumb switches.

[macboy]

If multiple discrete speed levels are acceptable, as opposed to "continuously" variable, then you may try using series capactors and switching them in/out with relays (normal or SS). If you use power-of-two capacitor values, then you can get 2^N speeds for N capacitors and relays.

[ivan747]

We don't need anything close to say, 12 bit resolution, but I'm sure they will want more speeds than what's currently offered. Perosnally I think 10 speeds would be a nice compromise. As you say, I could do something similar to an R2R ladder but with relays and caps with 3 or 4 bit resolution.

 That's a good way to save $30 in the board Orion was talking about earlier. The relays would have to stand the current consumption of the fan, and the caps must be safe for mains use. I think that's the way ceiling fan 5 speed controllers work. I took one apart recently, but it was all potted inside.