Not sure if I agree with the article saying there are no harmonic currents. After all, when you square off a sign wave, you have added a square component, doesn't matter if it's on the leading or trailing
edge. Oh well, time will tell.
I would agree that reverse
phase dimmers would by nature be quiet.
Now, here's a quick list of
dimmer types, pro's and cons:
Standard Dimmer: Chops off the leading
edge of the waveform.
Pros: Inexpensive, good workhorse.
Cons: Heavy due to large chokes, produces tons of noise.
Reverse Phase Dimmers: Chops off trailing
edge of the waveform.
Pros: Much quieter, no need for heavy
choke.
Cons: Newer technology so more expensive, lifespan yet to be determined.
Sine Wave Dimmers*: No chopping, varies amplitude of output. (think autotransformer-like)
Pros:
Cadillac of dimmers. No noise. Probably increases lamp lifespan due to less
filament vibration.
Cons: Creates more heat, Very expensive.
* I should add a
bit about how
Sine Wave Dimmers work- Pulse Width Modulation (
PWM) is how a lot of switching
power supplies work. The bulk of the waveform is sliced into small segments. Each segment is varied in width so that at full width, the full waveform is present. At a smaller width, only a small
spike is getting through. To "heal" the waveform and make it a
sine wave, the output is run through a
choke. The
choke stores the pulse as magnetism and then releases it back as
current. If the pulse was 100 volts, and had a width of 50%, the output
voltage from the
choke side would be 50 volts. Regular dimmers use big chokes to
help moderate the chop in their waveform and smooth it out, but they do this at 60cps, which requires a big chunk of copper and metal as the magnetic storage time is longer. Sine dimmers do it a 40,000 to 50,000cps. Because the frequency is so high, the
choke can be very small as the storage time is very short.
One last note: Heat. Transistors, Triacs, and SCRs, all are somewhat the same in core design, the difference is that Triacs and SCRs are designed to have a
runaway cascade so they go to full conduction when triggered. Transistors, don't. (Or shouldn't, although if you have worked in an amp repair shop, you have seen a few that did

) So, when fully gated, both will produce about 1
watt of heat for every amp of
current passed. Trouble is, the highest heat is generated in the "in-between" state. (why audio amps put out so much heat.) There is a time between when a transistor is told to turn on, and when it actually does. This is know as the "slew" rate. (see Slew induced
distortion on audio amps) At 60cps, you are entering the slew state 120 times per second. On a
PWM operating at 50khz, you are entering the slew state 50,000 times per second. The result is you are in the slew state 416 times longer, so the device runs hotter. That being said, it is probably well offset in sine dimmers because standard dimmers require there to be quite a length of copper
wire on the
choke, so the resistance in this
wire also throws heat.