How to quieten fan motor driven by a triac dimmer?

Thread Starter

MikeA

Joined Jan 20, 2013
192
I have a cap start AC motor that naturally makes noise when slowed down with a dimmer. Is there a way to silently control the motor speed? Or improve the dimmer circuit?

I know a variac would do it, but a variac itself hums. I'd love to have it be as quiet as possible.
 

Hypatia's Protege

Joined Mar 1, 2015
3,208
I have a cap start AC motor that naturally makes noise when slowed down with a dimmer. Is there a way to silently control the motor speed? Or improve the dimmer circuit?

I know a variac would do it, but a variac itself hums. I'd love to have it be as quiet as possible.
Please be advised that AV control of 'single speed' induction motors is properly achieved via VFD (featuring TSW output for 'silent' operation)...

FWIW the 'noise' experienced with 'phase control' devices (e.g. 'lamp dimmers') is down to the harmonic content of the resultant waveform.

Although a sufficiently rated variac of reasonable quality will not audibly 'hum' - please be advised that under-powering of induction motors results in excessive heating with liability to fire...

Best regards
HP
 

MrAl

Joined Jun 17, 2014
7,102
I have a cap start AC motor that naturally makes noise when slowed down with a dimmer. Is there a way to silently control the motor speed? Or improve the dimmer circuit?

I know a variac would do it, but a variac itself hums. I'd love to have it be as quiet as possible.
Hi,

As HP mentioned nicely there may be negative effects to trying to lower the speed using voltage control. A better idea would be to just find a fan with a low speed and set it and be done.

That said, i have used a variac and i can tell you it doesnt make any noise. It's rated at 20 amps 120vac. I also only slowed it down a little bit which may explain why it worked so well.

Alternate ideas would be to put a light bulb in series with it. You should check the motor for overheating though as that will be important. Most new fans have a thermal device inside that will blow out if it gets too hot and then it wont work anymore unless you dig into it and replace that thermal fuse.

The right way to do it is to reduce the frequency as you reduce the voltage. That kind of controller though will be much more expensive, so a different fan with a 'low' speed setting is a cheap way to get it done.

As a side note, it may be interesting to see what happens if we adjust the voltage AND skip every other cycle. That MIGHT emulate 30Hz for a 60Hz line to some degree. Looking over four complete cycles, the first half cycle would be allowed to pass, but the negative second half cycle would not. The third half cycle would be again positive so skip that too, but allow the fourth half cycle through. Skip the fifth half cycle (positive) skip the sixth half cycle (negative) but allow the 7th half cycle through (positive). Repeating that pattern, skip 8th and 9th, let 10th through, skip 11 and 12th, let 13th through. That produces a fixed 30Hz and if it works that would be a single set point which is not adjustable but at least lower.
A shot in the dark though which would have to be tested carefully as it wont be a pure sine wave anymore without some serious filtering.
 
Last edited:

Dodgydave

Joined Jun 22, 2012
8,797
You can change the speed by putting a smaller capacitor in series with the supply, this has the same effect as a dimmer, however the current will also be reduced, so the fan will run at a minimum speed, but may stall when loaded.
 

MrAl

Joined Jun 17, 2014
7,102
You can change the speed by putting a smaller capacitor in series with the supply, this has the same effect as a dimmer, however the current will also be reduced, so the fan will run at a minimum speed, but may stall when loaded.
Hi,

Well for a 1 amp fan you'd still need something like a 20uf capacitor rated for the full voltage and having an AC rating. But wouldnt that still be just voltage control?
 

Alec_t

Joined Sep 17, 2013
10,900
One thing to watch out for if using a series capacitor is to choose a cap value which doesn't resonate with the motor inductance at the mains frequency or a low harmonic, otherwise you can get surprisingly high voltages induced (enough to cause arcing inside the motor).
 

shortbus

Joined Sep 30, 2009
7,443
Just thinking out loud, the OP said it's a cap start motor. This means there is a centrifugal switch in it. If the motor is slowed down too much the switch will start to oscillate between engaged and not engaged, not good for the cap or the switch. How much does he need it slowed down. Lot of unknowns in the question.
 

MaxHeadRoom

Joined Jul 18, 2013
19,977
I suspect he really meant a PSC, permanent start cap motor, centrifugal sw type 1ph motors do not control well, and tend to drop out of run with the slightest load.
An alternative for the OP would be a multi secondary transformer and switch in different speeds.
Max.
 

Hypatia's Protege

Joined Mar 1, 2015
3,208
I suspect he really meant a PSC, permanent start cap motor, centrifugal sw type 1ph motors do not control well, and tend to drop out of run with the slightest load.
An alternative for the OP would be a multi secondary transformer and switch in different speeds.
Max.
I concur!

Should the stator arrangement/wiring illustrated in the attached image (below) correspond with the TS's implementation, I wouldn't mess with the Cap! - In my experience, excessive stator temperature (owing to effective under-powering?) and/or OV breakdown of winding insulation (owed to resonant transformation?) are common failure modes attending misapplication in this regard -- That said, the Cap value is appropriately 'adjusted' for different operating frequency ranges...

Disclaimer - my practical experience with induction motor control is rather limited - hence I defer to other respondents to this thread:) -- Even so - it is my stance that 'voltage throttling' of induction motors in continuous service is both unsatisfactory and potentially dangerous... My $0.02

Best regards and good luck!
HP:)

b150.PNG
 
Last edited:

Aleph(0)

Joined Mar 14, 2015
597
HP you get Coolidge tubes into everything! I say you're as bad as the basketball fanatic in Cheech and Chong tune only for you the lyrics are changed just by substituting basketball with xray tube so like this:

Then one day my mama bought me an xray tube
And I loved that xray tube
I took that xray tube with me every where I went
That xray tube was like an xray tube to me!
:D:D:D

So I say you need to find 12 step program:p
 

Aleph(0)

Joined Mar 14, 2015
597
@MikeA

MikeA It seems to me like you could just set up circuit to supply motor with 60Hz bursts as means of controlling it w/o overheating? So like half speed could be 250ms of 60hz and 250ms flatline so that's 15 cycles of line voltage then same amount of time of ov again and again. Also switching should be at zero crossing to prevent harmonics:)?
 
@MikeA

MikeA It seems to me like you could just set up circuit to supply motor with 60Hz bursts as means of controlling it w/o overheating? So like half speed could be 250ms of 60hz and 250ms flatline so that's 15 cycles of line voltage then same amount of time of ov again and again. Also switching should be at zero crossing to prevent harmonics:)?
You present an interesting concept - howbeit, inasmuch as reduced AV corresponds to reduced 'back EMF', I strongly doubt a 'time throttled' motor would run cooler than an equivalent 'reduced (continuous) EMF' scheme -- please bear in mind that 'phase control' might (likewise) accurately be regarded as control via duty cycle and, hence, 'time modulation'... --- Again, IMO (which being of possibly dubious value in this matter) control of induction motors via increased 'slip' - by whatever means - is 'a loosing game'...

Very best regards
HP
 

Aleph(0)

Joined Mar 14, 2015
597
please bear in mind that 'phase control' might (likewise) accurately be regarded as control via duty cycle and, hence, 'time modulation'...
HP I say there'd be some more heating using phase control cuz of hysteresis losses at harmonic frequencies:)!

control of induction motors via increased 'slip' - by whatever means - is 'a loosing game'...
HP I say that's right if efficiency is all that matters but if operation suits need and worst case heating stabilizes at safe level why not go for it:)?
 
HP I say there'd be some more heating using phase control cuz of hysteresis losses at harmonic frequencies:)!
Good point!:)

HP I say that's right if efficiency is all that matters but if operation suits need and worst case heating stabilizes at safe level why not go for it:)?
Because 'good enough' is never 'good enough' in the 'grand scheme' of things!

Best regards
HP:)
 

Thread Starter

MikeA

Joined Jan 20, 2013
192
Although a sufficiently rated variac of reasonable quality will not audibly 'hum'
Strange. I owned about half a dozen variacs in my life time, and used them primarily for fan control, all of them hummed, with load and no load. They were all American vintage or recent high quality variacs, Staco and such.

please be advised that under-powering of induction motors results in excessive heating with liability to fire...
Thankfully the fans I have also cool the motor by the air flow when on. The frames of the motors are about 80-90F when on to the speed needed.

A better idea would be to just find a fan with a low speed and set it and be done.
The fan I'm using does have multiple taps, but even the smallest power tap is still much too fast for my needs.

On a related subject, would it be better to lower speed using a triac or variac on the highest power tap? Or the lowest power tap? Does efficiency of the complete circuit come into the question here?

You can change the speed by putting a smaller capacitor in series with the supply, this has the same effect as a dimmer, however the current will also be reduced, so the fan will run at a minimum speed, but may stall when loaded.
I have controlled some smaller fans using a cap, but picked it by trial and error. Is there a formula to figure out what cap I need?

I suspect he really meant a PSC, permanent start cap motor
That's correct.
 

tcmtech

Joined Nov 4, 2013
2,868
The fan I'm using does have multiple taps, but even the smallest power tap is still much too fast for my needs.

On a related subject, would it be better to lower speed using a triac or variac on the highest power tap? Or the lowest power tap? Does efficiency of the complete circuit come into the question here?
Since its a PSC fan motor dropping the line voltage to limit the running torque will work just fine.

I would recommend using your variac to find what voltage best relates to the speed you want and then replace the variac with a simple buck boost transformer, either factory made or configured from any common appropriately sized transformer with turns ratios that work out to create an autotransformer with the correct voltage tap you need.

Like using a 120:24 transformer with both it's windings in series and connected to a 120 VAC line gives you a 100 volt tap or a 120:36 transformer gives you a ~92 volt tap.
 
MikeA
Is this about cost...? heat..? Efficiency...? noise..? treating the motor with the best specs.? put a piece of card board on the input side of the fan. don't cover the motor. no heat problem. no noise problem. the motor uses less current. so much for the electrical problems that could be created trying to over engineer a problem.
 

crutschow

Joined Mar 14, 2008
24,302
As a side note, it may be interesting to see what happens if we adjust the voltage AND skip every other cycle.
That's an interesting concept that should reduce the noise (or a least lower its frequency) as compared to phase-control approach.
But rather then trying to emulate 30Hz, you could just vary the number of ON and OFF cycles to give an effective reduction in voltage.
Thus two cycles on and one off should give about 3/4 effective voltage, etc.
This would just require a duty-cycle controller driving a zero-crossing SSR.
 

MrAl

Joined Jun 17, 2014
7,102
That's an interesting concept that should reduce the noise (or a least lower its frequency) as compared to phase-control approach.
But rather then trying to emulate 30Hz, you could just vary the number of ON and OFF cycles to give an effective reduction in voltage.
Thus two cycles on and one off should give about 3/4 effective voltage, etc.
This would just require a duty-cycle controller driving a zero-crossing SSR.
Hi there,

Well, doing it that way would not be as symmetrical due to the exponential part of the load response. It might work, but then again i have not done this either way since the 1980's in real life.

EDIT: Added a third graph in the third attachment showing both using a switch more like a triac.

The two graphs in the attachments show the approximate current for both methods. The red line is the current in each graph, the light violet line is the switch on/off signal (a switch was used not a triac).
AC_Motor_SpeedControl-1.gif is using your idea, skipping full cycles, and
AC_Motor_SpeedControl-2.gif is using my idea, skipping half cycles to keep currents above and below zero equal.

As we can see, in the -1 pic the positive current goes higher than the negative current goes low, thus would generate a DC offset or at least a non symmetrical drive. In the -2 pic the positive and negative current levels are the same so i would expect smoother operation, and if you finish out the red waveshape the way a triac would do it i think it would look much more like a 30Hz sine wave than with the other method. Note also that this is considered an 'odd' function for harmonic content while the other is an 'even' function, and a sine wave is an 'odd' function too.

In the graphs, the currents abruptly drop to zero but in a real circuit with a real triac that would not happen but you can see that in the -1 pic the current starts to drop before the turn off time, and in the -2 pic the drop off time doesnt matter as much because it's the same for both half cycles.

You can try this with a simulation using a triac model if you wish, i just used a switch for now that turns on or off for the first two pics, then a switch that emulates a triac more effectively for the third pic.
The turn on for the switch is shown in light violet, the pulse looking wave in each pic. That turns the switch on or off in both simulations.

Either method though is limited to the number of actual speeds that can be used due to the fact that we have to stay sync'd to the line frequency. Also, although i did something like this in the 1980's i dont remember what the outcome of it was now, so it would have to be tested carefully to make sure it worked long term. I also have to wonder about the audible noise from either method.

BTW i used 50Hz instead of 60Hz because it graphs slightly better with no repeating digit fractions for the time axis.
 

Attachments

Last edited:
Top