AC motor speed control

[[email protected]]

I have one of these (http://www.bestgrowlights.com/site/403863/page/
941189) but the bugger blows way too much air and is very loud. So I
bought one of these (http://www.plantlightinghydroponics.com/fantech-
variable-speed-fan-control-p-1487.html) and that did slow things down
a bit but not enough.

I have a 10A solid state relay and a microcontroller or two sitting
around. I was curious if I could control the speed of the AC motor by
cycling the availability of the juice like I could with a DC motor.
IE: Hook the hot wire through the relay and cycle it on for 20ms, off
for 80ms for a 20% duty cycle or some such (with the ability to adjust
the duty cycle of course, 100% for 5 seconds at start, etc).

I don't know much about AC though and less about AC motors. Is this a
valid approach? Will it burn up the motor, catch the relay on fire,
destroy my karma, etc?

TIA!

 

[Sjouke Burry]

I have one of these (http://www.bestgrowlights.com/site/403863/page/
941189) but the bugger blows way too much air and is very loud. So I
bought one of these (http://www.plantlightinghydroponics.com/fantech-
variable-speed-fan-control-p-1487.html) and that did slow things down
a bit but not enough.

I have a 10A solid state relay and a microcontroller or two sitting
around. I was curious if I could control the speed of the AC motor by
cycling the availability of the juice like I could with a DC motor.
IE: Hook the hot wire through the relay and cycle it on for 20ms, off
for 80ms for a 20% duty cycle or some such (with the ability to adjust
the duty cycle of course, 100% for 5 seconds at start, etc).

I don't know much about AC though and less about AC motors. Is this a
valid approach? Will it burn up the motor, catch the relay on fire,
destroy my karma, etc?

TIA!

You might end up with a mighty hot motor
using it that way.
It wont take kindly to fast on/off AC to control
speed, all the higher harmonics in the applied
AC are converted to heat.

 

[Kevin S.]

Instead of doing asynchronous pwm to the ac, how about keeping the
frequency at 60 hz. Turn your power on when close to the positive
peak, and then off again just past the peak, and doing the same for
the negative peak. So the main frequency is still 60 cycle, but you
can control the total power delivered by making your pulse longer or
shorter in time while keeping it centered on the peaks. A little
microcontroller with an A/D converter should be able to do that, but
you'd need a bit of analog circuitry to get the AC signal into the
right range for the A/D. I've never tried this, just an idea.

-Kevin

 

[JeffM]

[...]I don't know much about AC though and less about AC motors.

If you don't hold its hand all the way,
Google Groups does a really crappy job of posting to Usenet.
Long links in particular get smashed.
http://groups.google.com/group/sci.electronics.misc/msg/a988332e61a73ee0
If you will put each of those on a line of its own,
you have a better chance of them not getting screwed up.

In addition, if you use an email address as a username,
Google also fucks that up.
If you *sign up* for a group before posting to it
http://groups.google.com/groups/mysubs
and give yourself a username (without a domain in it),
you will look less goofy to other people using Google Groups
(aka The Usenet Archive).

I have one of these (http://www.bestgrowlights.com/site/403863/page/
941189) but the bugger blows way too much air and is very loud. So I
bought one of these (http://www.plantlightinghydroponics.com/fantech-
variable-speed-fan-control-p-1487.html) and that did slow things down
a bit but not enough.

If you just bought the properly-sized device in the first place,
you wouldn't have to jump thru hoops.
http://google.com/froogle?q=muffin-fan+cfm&num=100&scoring=p&price=between&price1=1

 

[[email protected]]

Instead of doing asynchronous pwm to the ac, how about  keeping the
frequency at 60 hz.   Turn your power on when close to the positive
peak, and then off again just past the peak, and doing the same for
the negative peak.   So the main  frequency is still 60 cycle, but you
can control the total power delivered by making your pulse longer or
shorter in time while keeping it centered on the peaks.   A little
microcontroller with an A/D converter should be able to do that, but
you'd need a bit of analog circuitry to get the AC signal into the
right range for the A/D.  I've never tried this,  just an idea.

So a transformer to bring the 120 down to 3 V peak-to-peak then boost
the baseline to 2v DC put that into the ADC input? Then when it gets
close to 4.5 or .5 then kill the motor and re-enable at the same
voltage on the other side of the peak/trough?

Or could you accomplish similar with op-amps?

 

[Wim Lewis]

So a transformer to bring the 120 down to 3 V peak-to-peak then boost
the baseline to 2v DC put that into the ADC input? Then when it gets
close to 4.5 or .5 then kill the motor and re-enable at the same
voltage on the other side of the peak/trough?

Or could you accomplish similar with op-amps?

If you just want to detect the zero-crossings of the AC supply, there's
an even easier way. Take the 120V AC, feed it through a large resistor
( > 1 MOhm ) directly into a microcontroller pin. The input protection
diodes will clamp the mains voltage to a nice square wave, and the
approx. 100 microamps that flows through the resistor is well within their
ability to handle --- check the data sheet.

However, I'm not sure that Kevin S's phase-control-dimmer-style speed control
is a good idea ... a motor is an inductive load, and treating it like a
simple resistive load is likely to burn out your switches, your motor,
or both...