AC fan speed control by periodic on/off switching

Not instantly, but the asymmetric output will not be good over a period. Eventually I expect the motor to die.

I wonder about putting a capacitor in series. This can be a dangerous idea, however. You don't want to make a tuned circuit that somehow rings up to a large voltage.

There is another problem with the light dimmer idea. They are not designed to work with inductive loads. When the current in the motor stops, the mains voltage won't be at the zero crossing. There will be a large dv/dt right after the current stops that would likely retrigger the triac.

That is doable as well, but the controls needed for say 10% steps like OP indicated make that approach very complex.

I think it is well within the scope of what a small microcontroller could do. For extra credit, lets see if we can figure out how to get a measure of the speed of the motor.

We have a complex waveform on the motor. If we measure the voltage and current, we should be able to figure out an impedance curve for the motor. The impedance will be very high near the synchonous speed.

I'm interested. :-)

I'm curious how you'd do it. I voted for integral-cycle, but pseudo-bangbang with zero-crossing MOCs sounds interesting as well.

Then again, is there that much difference between the two? ;-)

Thanks, Rich

Oh, OK. Never mind my other post where I asked you to post the schem. :-)

Cheers! Rich

Actually, the schematic I found using the double 555

So mr. Fields... if your solution can handle this better, if you could post the schematic anyway? Thanks!

--- Which is exactly what you want.

Lets say that you have a fan running with a couple of 555's in series so that the output into the TRIAC driver looks like this:

___________________ __ ___| |___________________|

||| | ||

where each positive pulse is 300 60Hz cycles wide and the period of the waveform is 600 60Hz cycles.

Now lets say that the fan is running too fast and that you want to slow it down. What you would do would be to either decrease the width of the positive portion of the waveform (decreasing the number of cycles into the fan) without disturbing the width of the low portion:

___________ __ ___| |___________________|

||| | ||

or increase the width of the low portion of the waveform while leaving the high portion alone:

___________ __ ___| |___________________|

||| | ||

Either way would cause the fan to stay on for less time than it was on, which is what you'd want to cause happen if the fan was running too fast.

--- Mine handles it in exactly the same way, but the dual 555 way is much less complex to implement so, IMO, that's the better way. In any case, I'll post the schematic in a day or so. :-)

-- JF

You need to use around a URL, not () See below:

Actually, the schematic I found using the double 555 only sets the absolute duty cycle. I.e. if you change the frequency, the duty cycle in seconds would stay the same, and the % duty cycle is changed. You'd need to adjust it again.

On Sat, 12 May 2007 03:37:55 -0500, John Fields wrote:

--- Aaarrghhhh!!!

I must be getting senile.

You can do it with a single 555 astable: . .VCC>--+---------------------------------+ . | | . [Ra] +---------+ | . | 7 |_ | 8 | . +----------------O|D Vcc|-----+ . | 6 | _| 4 | . +----[Rb]----+----|TH R|O----+ . | | 2 |__ | 3 | . +--[DIODE>]--+---O|TR OUT|-----|--->OUT . | | GND | | . [C] +----+----+ [0.1µF] . | | 1 | .GND>---------------+---------+----------+ .

RaC determines how long the output will remain high and RbC determines how long it will be low.

You could also use a pot (about 1 megohm) to make adjustment easy, and then C would be about 10µF to give you a period of about 10 seconds. The 1k is in there so you don't toast the chip if you crank the pot all the way up to Vcc.

. .VCC>--+---------------------------------+ . | | . | +---------+ | . | 7 |_ | 8 | . [POT]OUT . |+ | GND | | . [C] +----+----+ [0.1µF] . | | 1 | .GND>---------------+---------+----------+ .

Here's an LTSPICE circuit list you can use to run a simulation:

Version 4 SHEET 1 880 748 WIRE 208 160 -16 160 WIRE 784 160 432 160 WIRE 208 224 144 224 WIRE 512 224 432 224 WIRE 624 224 592 224 WIRE 656 224 624 224 WIRE 784 224 784 160 WIRE 784 224 736 224 WIRE 208 288 64 288 WIRE 480 288 432 288 WIRE 512 288 480 288 WIRE 624 288 624 224 WIRE 624 288 592 288 WIRE 208 352 176 352 WIRE 64 368 64 288 WIRE 480 384 480 288 WIRE 512 384 480 384 WIRE 624 384 624 288 WIRE 624 384 576 384 WIRE 144 448 144 224 WIRE 480 448 480 384 WIRE 480 448 144 448 WIRE 176 480 176 352 WIRE 784 480 784 224 WIRE 784 480 176 480 WIRE 784 496 784 480 WIRE 480 512 480 448 WIRE -16 592 -16 160 WIRE 64 592 64 448 WIRE 64 592 -16 592 WIRE 480 592 480 576 WIRE 480 592 64 592 WIRE 784 592 784 576 WIRE 784 592 480 592 WIRE -16 640 -16 592 FLAG -16 640 0 SYMBOL Misc\\\\NE555 320 256 R0 SYMATTR InstName U1 SYMBOL voltage 784 480 R0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V1 SYMATTR Value 5 SYMBOL res 608 272 R90 WINDOW 0 0 56 VBottom 0 WINDOW 3 32 58 VTop 0 SYMATTR InstName R1 SYMATTR Value 900k SYMBOL res 752 208 R90 WINDOW 0 0 56 VBottom 0 WINDOW 3 32 56 VTop 0 SYMATTR InstName R2 SYMATTR Value 100k SYMBOL res 48 352 R0 SYMATTR InstName R3 SYMATTR Value 1000 SYMBOL cap 464 512 R0 SYMATTR InstName C1 SYMATTR Value 12e-6 SYMBOL diode 576 368 R90 WINDOW 0 0 32 VBottom 0 WINDOW 3 32 32 VTop 0 SYMATTR InstName D1 SYMATTR Value 1N4148 SYMBOL res 608 208 R90 WINDOW 0 12 110 VBottom 0 WINDOW 3 -39 59 VTop 0 SYMATTR InstName R4 SYMATTR Value 1000 TEXT 16 616 Left 0 !.tran 20 uic

-- JF

Thanks for the schematic. I plan to test it on a test board; I think it's useful. As for your explanation in your last post where you said that frequency doesn't need to be altered: you are right, as a speed regulation the pulsewidth method suffices. Though I would like to be in control of frequency too (I like to think of the fan spinning up & slowing down as a sound likened to sea waves, when it has a long period of, say >=3D 8 seconds). For the method you aim at, a period of 1 second or less would work, no? (C=3D1=B5F) It would then probably go unnoticable that it's switching on/off to attain low speeds.

I'm using the MOC3041 triac driver. I have some questions regarding its use, though. This is from the data sheet: (applies to MOC3041,

Direct link to the MOC3041 data sheet (pdf):

(or Google...)

You can decrease the 360 to 180, the 330 is fine. I'd use 100 & .01 for the snubber, but it's not critical. See the url below for some snubber info.

The IR led in the MOC3041 probably has a vf closer to

Ed

I have done this a few times using different methods.

If using AC remember that you have to balance positive and negative phases otherwise the motor might overheat.

Phase control works well.

Swicthing one full mains cycle then missing one full one works too.

Thanks for the great help, John! The Fairchild datasheet really helped. And also thanks to ehsjr.

And now that we're talking about power, I think of one thing: do I need to add a fuse in series with the BT137 triac? Since the triac can handle up to 8A and probably has no protection for over-current, I'd have to use an 8A fuse, right?