I made a microcontroller circuit to control speed of a small fan with TRIAC phase control. It works fairly well, but the fan never completely stops, among other things, so I am wondering if the TRIAC is being triggered the way I expect.
I don't have a CRO, so I want to sample the output voltage from the TRIAC using the microcontroller's ADC.
I could do this using a simple resistor divider, but then I have to connect circuit ground to Neutral, and circuit ground is my computer ground. Neutral is supposedly connected to earth, which in theory would make this safe (I would test first), but it seems like bad practice to do this...
I have a transformer, 240V->24V 20VA 50/60Hz. If I put it's primary on the TRIAC output and a resistor divider on the secondary, that would probably be safe, but will it work? Will I be able to see the sharp edges where the TRIAC turns on in a half cycle?
Another question about whether the transformer will pass the signal: To detect zero crossing, I am using an optocoupler, but other circuits show using the AC output of the power supply transformer. Because the transformer is inductive, wouldn't there be a delay between the actual zero crossing and when the zero crossing occurs on the secondary?
Hi, David. Did you expect a good answer without providing much of any pertinent information? Maybe you could start by answering a few questions:
What type of motor do we have here? Is it an AC induction motor?
What's the current rating (or watts rating) on the nameplate?
Are you using an optocoupler to drive your triac? (Hope so!) If so, which one? Is it a zero-crossing triac trigger? (If it is, throw it out and get one that isn't.)
Are you timing a delay from the zero crossing of the AC line?
Have you provided a snubber (R-C or other) across the triac? Could you sketch out your circuit, using an ASCII Circuit program? (possibly use Andy's ASCII Circuit v1.28
formatting link
)
To start with, there are a few things here. First, you should know that phase control of AC induction motors is difficult, and sometimes it just can't be reliably done. Problems do tend to occur at the low end, and can actually be more difficult with smaller motors.
However, there are several things you can try here.
If you haven't already, put an R-C snubber across the triac (start with 220 ohms 1/2 watt in series with 0.1uF line service self-healing cap).
Obviously the biggest problem here is the inductance. You're going to have difficulty turning the triac off if the curent is going great guns while the voltage crosses zero. One good solution for this is using back-to-back SCRs, with separate optocouplers for each SCR. Obviously the reverse-biased one won't trigger, and the other one will stop when the voltage reverses. There are many solid-state relays which are made like this, and all you have to do is provide the 3mA or so to drive the opto input (make sure you don't get an SSR with zero-crossing triggering). One cheaper solution, which *might* help is to get a Teccor Alternistor, which has many of the characteristics of back-to-back SCRs. It provides better turn-off behavior than a standard triac. I've found I had better luck driving inductive loads with these:
formatting link
If you're desperate, and the motor's small enough, you can start trying to change the power factor of the load by adding a power resistor in parallel with the motor. Adding a capacitor across the motor is a spectacularly bad idea, because you'll cause di/dt failure in the thyristor (too fast a change in current causes the triac to smoke). Of course, you're just burning power this way, but it can help.
If worse comes to worst, you can always do it like the dimmer switches used for ceiling fan controls -- they reduce the speed as you dial down, and then use a switch to turn off. You can always just use a relay in addition to your phase control.
Having said all of this, I'd guess that, especially without a scope, you're going to have an easier time just working on solving your problem than trying to diagnose it by reading the voltage across the thyristor or the load with your uC. You're going down a rabbit hole here -- you just want to control the motor load. Take the steps you need to do that, and worry about the other stuff later.
Actually, there are many applications where metering the voltage (or more frequently, the current) on an AC load *is* a good idea. It can be a good part of the control circuit, and there are many ways to do this -- but that's actually another problem. In fact, as long as you don't load the transformer secondary, you can get a pretty good idea where the zero crossing is, and what's going on with the load voltage. And there's no lag to speak of between transformer primary and secondary. You are also adding inductance, and changing the circuit you're measuring, too. But let's work on the fan first here.
Feel free to take the time to express your problem a little better -- the few minutes it might take to actually describe your problem well is a lot easier to spend than hours thrashing around in the lab.
It's likely to trip an ELCB/GFI. but if you have an isolating transformer on hand you could run your circuit from that and connect its "neutral" to the PC's ground without risk.
The 24V transformer would be OK to measure the output voltage and there would be no "delay". Also, the easiest place to measure the zero crossings is on the secondary of the power supply transformer. All it needs is a couple of diodes. john
I want to make this for a hobby community I am involved in, I don't know what fan people might use but it will probably be in the 30-200W range. Capacitor run will probably be the most common, but some people use shaded pole types.
I am testing with 3 different fans. One is '240VAC 0.25A 50Hz' shaded pole. One is '230V 50Hz 173W' capacitor run. One is a self-oscillating room fan, '240VAC 50Hz ONLY 50W', not sure what motor type but it has 3 speeds and there are 4 wires going up into the motor.
I am using a MOC3021, which is nto a zero-crossing type.
I detect zero crossing with a 4N25 and start a timer in my microcontroller, driving the MOC3021 at a variable time after the zero crossing.
This schematic doesn't show it, but I am using a 1/2W 68 ohm and a 100nF X2 across MT1 and MT2.
I didn't have one when I first started because there appears to be one in the schematic above, which I copied from a MOC3021 datasheet appnote. Apparently that is something to do with the TRIAC gate so now I have a 68R and 100nF extra across MT1 and MT2.
Is an alternistor what ST would call a 'snubberless' triac? If so, don't I need the ability to trigger in the 4th quadrant to do power control?
I have a very old (Nov 1980 date code) fan speed controller here, which is almost identical to the circuit in figure AN1003.15 in
formatting link
This controls all 3 fans perfectly, but not to a complete stop. It is triac based so I should be able to get away without SCRs. In fact it doesn't even have a snubber...
OK, well since I first posted I have found a few bugs, mostly in my zero crossing detection circuit. That part is working properly now and matching my calculations.
I tested my triac firing output from the microcontroller by putting an RC circuit on it so that I can get a duty cycle reading. It also matches expectations, giving 0-100% over 0-90deg firing angles.
Unfortunately fan speed control is not so predictable.
The 173W capacitor run fan:
Starting with 90deg firing angle (zero power), works as expected, fan stopped, no buzzing and doesn't get hot. Speed increases as expected up until 45deg when it doesn't get any faster. However, at 16deg, the speed starts jumping around, 1s at full, 1s at slightly slower, and so on. At 14 deg it slows right down, starts buzzing and generally just gets very hot. This continues until 2deg, when the fan stops and I cannot hear any buzzing at all.
Starting at 0deg and increasing power, starts silent, then starts buzzing at
2 deg, buzzes until 19deg, then at 21deg it takes off full speed. Stays full speed until about 45deg, when I can detect a bit of a slow down (by ear). Can control speed right down to zero.
The shaded pole 0.25A fan: Barely worth writing about. At 90deg, no buzzing as expected, but as power is increased, starts to spin but vibrates soflty at what sounds like about
10Hz. On the way from 90 to 0, fan starts to spin then stops at 3 or 4 places in this range. Eventually burns out snubber resistor.
This ^^^ is with 68R and 100nF snubber.
I have tried 3 different triacs (in brackets, critical rate of rise of off state voltage): T405-600 (20V/us) BT137X-600 (100V/us) BT139-600 (250V/us)
** The LED inside the MOC3021 must be driven on and off correctly to avoid " half waving" (*)
Do not pulse the LED - drive it on throughout the time you want the triac to be on only turning it off 0.5 mS prior to the next zero crossing so the triac can turn off at the next current minima.
Half waving is common problem when PULSE fired triacs are used to phase control an inductive load. The triac may turn off at an unexpected point in the AC cycle since, at that point, current flow has gone through zero value. The result is that the output waveform is unsymmetrical and has a large DC component. Motors and transformers will make a growling noise and burn out if this continues.
The solution is to continuously fire the triac for the whole time it should be conducting.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.