As someone who did electronics design and prototyping for many years - and wrote for most of the magazines in the US and Australia - I shouldn't have to ask this sort of formless question... but it's been a number of years since I designed any power circuits and I am not sure where to start. A little bit of navigation and current SOTA would be helpful.

I'd like to experiment with AC power control using a "whole wave" approach, which I am sure is the wrong term - but if there is a term, and it's not full-wave, I don't know what it is. Basically, I'd like to do dimming and motor speed control by allowing selective full AC waves through - e.g. for 50% power suppress every other wave, for 75% power suppress every fourth wave, etc. This may be more complex than standard triac/full-wave AC power modulation, but seems like it would have advantages in RF noise, motor hum, light flicker etc.

I doubt this is any new idea, but searching my old references and the best I can online, I can't find anything that doesn't point to the partial-wave "full wave" control that's so standard. So...

1. Is this a standard technique, and if so, what's it called?
2. Are there standard circuits to achieve it, on a low-power (3-500W) level?
3. Are there any chips that simplify the design?

I'd be most interested in an "open" circuit design where I can specifically tell the circuit which waves to pass and which ones to skip, so that I can try various combinations of net power settings. I'm thinking a combination of zero-crossing detection and possibly two time-delay circuits that would enable X number of waves and pause for Y number, with a decent power stage.

Thanks for all useful input to what no doubt sounds like a hugely ignorant question...

 

I can't speak to the wisdom of your approach and I doubt it will bring any advantages. But it should be no problem to have a microprocessor follow the waveform and produce any output pattern you want. The microprocessor signal will then control the switching of the power.

I can't think of an analog, non-micro approach that would provide the flexibility you want.

 

One term for that is duty-cycle AC control.

Don't know if there any off-the-shelf devices for that but a Google search may turn some up.

Otherwise you could use one 555 timer running in an astable mode with both an adjustable duty-cycle (PWM) and frequency to control a zero-crossing AC SSR (solid-state-relay).
Below is an example 555 circuit for that.

 

Thanks to both of you. Yes, I can work out the theoretical model for the circuit - zero-crossing trigger, two 555's to control on and off time, triac output stage. I just wondered if there was any existing art in this direction, up to a single-chip solution.

It seems to me that it would be more efficient in driving motors and lamps more smoothly, with less noise/hum and possibly less flicker, since all switching is done at the zero point. Since this is not rocket science and I'm not claiming to be Tesla's heir, I assume it's been investigated and set aside for good reason. However, as we all know, those "good reasons" often involve trivial cost savings and the ready availability of standard full-wave control.

Further input solicited; I'm going to start ginning up a breadboard circuit. Now, where did I put my suicide cord?

 

It's commonly used for for close control of the heat in environmental ovens, where the heating element has a high thermal mass to average out the pulses of power.
The likely reason it isn't used more is that the low frequency pulsing of the power causes mechanical noise in motors due to the large torque fluctuations, and flickering in lamps.
Remember that the power pulse frequency of such a system must always be (generally significantly) lower than the power line frequency.

 

Skipping cycles would probably be a bad idea on rotational loads. We usually control Duty Cycle of each "wave". Is the noise "chop noise" or power hum?

Are you willing to pay more for less noise? One could use an old analog VFD.....but you pay for it with heat generation.

And there is always the old pre-digital standby.....controllable reactors or magnetic amp circuits.

 

I made a cycle-skipping controller for my floor fans. Only every-other full cycle was used. The motor was constantly running in "spin up". It wasn't until the thermal fuse in the second fan blew that I understood what was happening.


Output waveform


Circuit timing


Block diagram of circuit


Schematic of the circuit


The board that killed two floor fans.

Note: This circuit is connected to the AC Lline and carries lethal voltages. If you really feel the need to probe the circuit while in operation, please insert an isolation transformer between the AC Line and the circuit.

By the way, everything was working beautifully -there was a gentle breeze from the fan but little noise, until the thermal fuse blew which caused a certain amount of misery as I had to find a replacement and install it. In the end I used motor run capacitors in series with the fan to slow it down and those have been working beautifully for more than a year.

 

'Burst mode' is another term for what you propose. It is usable with heating systems having high thermal inertia but, as noted above, can be problematic with motors. I certainly wouldn't use it with lamps, since flicker can be annoying even when every mains cycle is used.

 

Remember that the power pulse frequency of such a system must always be (generally significantly) lower than the power line frequency.

Aren't airplane power systems 400Hz? The technique might work better in that sort of power system.

 

Here are two recent threads that include some discussion of burst fire control:

https://forum.allaboutcircuits.com/...ter-with-a-5v-pwm-signal.150211/#post-1285213

https://forum.allaboutcircuits.com/...g-ssr-heat-sink-s-potentiometer.146071/page-2

This is a link to a Crydom page which describes the pros and cons of various switching methods, including burst fire:
http://www.crydom.com/en/tech/newsletters/solid statements - ssrs switching types.pdf

And here's a link to a microcontroller burst fire solution:
http://ww1.microchip.com/downloads/en/AppNotes/00958A.pdf

 

I'd be most interested in an "open" circuit design where I can specifically tell the circuit which waves to pass and which ones to skip, so that I can try various combinations of net power settings. I'm thinking a combination of zero-crossing detection and possibly two time-delay circuits that would enable X number of waves and pause for Y number, with a decent power stage.

Thanks for all useful input to what no doubt sounds like a hugely ignorant question...

One approach that might be pretty simple to implement (and might work well for some loads and not for others) is to set a reference that establishes the target RMS value applied to the load over some reasonable time window. Then have a monitoring circuit that measures a proxy for the RMS voltage applied to the load. Use a comparator to arm/disarm your firing circuit which allows one full cycle of the AC waveform through at time (when armed). Thus at each cycle the circuit decides whether to let that cycle's power be applied depending on whether the average is above or below the reference target.

 

At first my thoughts went to simply generate the desired waveform in a µC and be done with it... Simple enough. But...

Still using a µC, and in this case I'm thinking of a Microchip PIC device with built in ZCD (Zero Crossing Detector); using an appropriate "switch", watch and count the ZC'ings and when you want, turn the switch off, killing the output for as many cycles as you want, turn back on as desired. No need to do anything but count, and if I'm not mistaken, I think a PIC (w/ZCD built-in) and a resistor and you're done. If it's 60 cycle, you've got (in µC time) eons to do counting and stuff between cycles...

Good luck!

Mike Tripoli

 

Another approach is to chop the mains AC into equal segments say at 50 Khz rate and PWM the signal, this effectively varies the power from 0 to 100%, example concept below for a high power battery charger using a large toroidal core that I wanted to do something with, rather than use it as a door stop.
This is what the AC will look like



Cheers
Mike