Hi,
I am from india. I want to control the celing fan speed in my workshop (220V AC voltage) using arduino. can any one provide me with a circuit.
I need to control the speed by sending the data to arduino from my PC.
thankyou in advance
Have a nice day
Hi Bill,
You're better at google than me - that's actually not bad a circuit at the link you found, I couldn't find any! I can't see the PCB layout (so I'm not sure on the spacing between mains and low voltage), but at least the schematic looks pretty good.
Basically, these triac type of devices are pretty much absent (i.e. glossed over in a few pages if any) in many books on semiconductor devices (or at least any books I have), I guess they come under power theory and more expensive books I don't have!. I don't like playing with mains either! but basically from limited use of them, they act like a latching switch, controlled by the gate current. Once triggered, they remain conducting, until the current through the switch drops to some threshold for the triac. Latches are usually not nice (ideally you want a switch like a normal BJT) but conveniently triacs become useful for controlling mains devices, since the sinewave AC voltage will reach zero 100 or 120 times per sec, so you know that the triac will eventually switch off anyway. If you want to dim or control speed, then you could do it in one of two ways: (a) check when the AC voltage reaches some adjustable level, and then turn the triac on. It will switch off at the end of the cycle. Then repeat. or (b) turn on the triac at the start of the cycle, and it will switch off at the end of the cycle. Then skip N cycles. Then repeat.
If you're switching a large load, then it's good to switch it on at the point when the voltage is at zero, so that there is no spike in current which generates noise. That implies that the (b) method may be better.
The other complication is that if the load is inductive, then the current through the load is not in phase with the voltage. So, when the triac switches off (i.e. when the current has dropped below the threshold for the triac), actually the voltage may be quite high. When the triac is on, the voltage across the triac is low because the load sees it (the triac resistance is low when on, so the voltage across the triac is low).
(Edit: added this quick diagram to explain this):
When the triac switches off, the voltage is across the triac (since it is no longer across the load, since the triac resistance is now high as a switch) and the fact that the voltage went from near zero across the triac to a higher voltage (due to the load being inductive) means that there was some dV/dt rate, and unfortunately large rate of change of voltage with time (i.e. dV/dt) kills triacs. This is the figure 14 from the doc mentioned earlier, where you can see that that the triac remained conducting (i.e. voltage across it was low) even after triggered, and even after the zero crossing, because the current was not in phase. When the current dropped, then the triac experienced the high voltage, and you can see the dV/dt is quite sharp.
Another thing is, how to know when there is a zero crossing. They achieve something close to that in the circuit posted, although the exact implementation depends on what they coded too, because you can see that the optocoupler won't switch on until there is a certain forward voltage for the LED inside it (and by that time the mains voltage is not zero, it is higher, i.e. many times the LED forward voltage, since they are driving the opto by a low voltage transformer). So, they could in thoery compensate for that by knowing the frequency of the mains voltage (50 or 60 Hz) and figuring out how much time would have elapsed from the true zero crossing, to the time that their optocoupler circuit reported it to the microcontroller, and taking that into account on the next cycle. There are other zero crossing circuits that won't need the compensation since they don't rely on the optocoupler, but the opto method may be cheaper and just as useful I guess, if they compensate in the code.
Mark, my friend, there is nothing that will make me want to work very hard to prove you wrong as telling me that I cannot do it. 
Ah, I think I know where William is going with this! Something like a normally-on microswitch to cut the power when the arm reaches a limit maybe? :-)
Btw, you might like this site if you've already not seen it. I'm not super-familiar with the site, but it's nice to see the circuit simulated (and animated), so if you click on the link to go to the applet (needs java), then it opens up a new window like below, and if you click on Circuits->Misc Devices->Silicon Controlled Rectifiers->SCR (which for DC is close enough and will work for DC as will a triac) then you'll see a circuit where you can click on the switches to turn them on and off. You could also modify the circuit, since it is a complete simulator, by right-clicking anyhere.
But for low voltage DC control, nothing beats the MOSFET, since it can switch lots of amps, and may be cheaper and smaller than triacs. The only problem is that most are surface-mount, but I remember someone finding some TO-92 packaged ones. I'll see if I can find the name. The discrete mosfets are good for controlling power in one direction, but for two-direction then the circuit does start to get big, so an IC becomes more practical. Recently I used an IC for the control, and used it to switch four external mosfets, using MC33033 (old IC, but still manufactured and quite nice - it's designed for brushless motors, but it can be used with DC motors, shown in figure 45 of the datasheet).
