PWM for 110/220V lamps

One can find lots of schematics for triac-based front-phase cutting dimmers. But there is neither schematics, nor actual products which would do proper PWM, where the circuit is switched on and off many times per mains cycle and amount of light is controlled by the width of the pulse (hence, PWM).

In my opinion, front-phase (triac) dimmers introduce too much noise into mains and cause filaments to make noticeable acoustic noise. I understand that PWM will be free of those problems, am I right? So, what is the reason nobody uses PWM for dimming mains lamps? Is it costs or power dissipation or EMI/RF noise or something else?

Thank you for your opinion!

I understand that PWM will be free of those problems, am I right?

Not if you are trying to control AC power. A PWM signal (basically a digital on/off signal can not directly switch AC power, the final control element will still have to be a thyristor (or back to back SCRs).

PWM control of AC is very possible if the PWM frequency is much slower then the AC frequency. It is used a lot for temperature control (which is generally a slow controlled process) using the PWM signal to activate a solid state relay. However a solid state relay is just an embedded thyristor with an optical isolated input.

Keep in mind that thyristor devices can hold off AC voltage and then trigger on at anytime in the cycle, however the device cannot turn off the AC current once triggered, only when the AC voltage drops to zero crossing will the thyristor turn off and wait for another trigger command. Trying to PWM directly at faster then 60/50hz is not possible with typical power AC switching devices.

Lefty

Thank you very much for your replies!

All the problems you cite for simple phase-angle switching are resolved by simply inserting an inductor in the load line to limit the dv/dt.

The way I understand it, the inductor should be selected based on the load. Too little load – and inductor will smooth away too much; too big load – and the noise will be back: just look at simple triac-based dimmers (Lutron) from Home Depot: the bulbs’ filaments buzz in my bathroom when load > ~200W. I.e. if the load is not known beforehand, selecting the inductor is a crapshoot. I understand that reverse phase-angle switching can reduce this problem, but not completely eliminate it, and I’m hoping PWM can eliminate it.

thyristor devices can hold off AC voltage and then trigger on at anytime in the cycle, however the device cannot turn off the AC current once triggered, only when the AC voltage drops to zero crossing will the thyristor turn off and wait for another trigger command. Trying to PWM directly at faster then 60/50hz is not possible with typical power AC switching devices.

Thyristors can’t turn off, but MOSFETs and IGBTs can. They’re more than capable of switching on and off > 1000 times/sec, i.e. > 10 times/halfcycle. What can be wrong with this approach?

Thyristors can't turn off, but MOSFETs and IGBTs can. They're more than capable of switching on and off > 1000 times/sec, i.e. > 10 times/halfcycle. What can be wrong with this approach?

Your correct. However the number of MOSFETs or IGBT used will double because the bi-polarity of AC power, plus high voltage high current mosfets and igbt tend to be a lot more expensive then thyristor devices.

Indeed some large industrial UPS systems use mosfets or igbt for generating the AC power from a DC link, however the large ones I worked with still used high power SCRs.

Lefty

If you want true PWM for 110/220, do what electronic ballasts do in fluorescent lamps. Run the AC through a diode bridge, filter it a little bit and chop it.