Dimming domestic CFLs safely

I don't know of any dimmable ones. They exist in the US market but not the UK market.

You can get 4-pin compact fluorescents to run from separate electronic control gear, and there are dimmable versions of this available. However, they won't run from a standard filament lamp dimmer, and make sure you are sitting down when they quote you a price...

No.

I think Don is referring to traditional fluorescent lamps with separate control gear.

if you keep the dmimer at full its fine, but if you turn it down you can expect the cfls to die early. Not a safety issue.

NT

There was a link posted a few months ago, to ones in the UK. Only several levels of dimming, rather than variable, but still dimmable.

Osram announced the Dulux EL Dim for spring 2006:

It is already available here in Germany for about 20 EUR/pcs.

Christian.

In the US most CFLs have a clear statement on the lamp that they are NOT compatible with dimmers. A few are designed for use with dimmers and that is stated on the box or the lamp.

I don't know how lamps are labeled in the UK, but I suspect they are not suitable for use with dimmers unless they state so on the lamp or box. You run the risk of destroying the dimmer and/or the lamp if you use these lamps with a dimmer.

CFLs with electronic ballasts (just about the only type available today) never use a parallel capacitor for power factor correction. The problem is that LOW power factor CFLs use a full-wave rectifier feeding a large electrolytic energy storage capacitor. It is this capacitor that destroys the phase-cut dimmer. However, certain dimmers have a set of mechanical contacts that close whenever dimmer is full on, bypassing the Triac. If a dimmer of this type is turned to full on, it will not be damaged by a low power factor CFL. However, as the user cannot guarantee that the dimmer setting will not be changed, use of any type of dimmer is not recommended with a low power factor CFL.

All dimmable CFLs are high power factor and use active power factor correction.

If they work the way I suspect they do, there is no "power factor correction electronics". It's just a feature of the SMPSU inside them that they have pretty much unity power factor without needing any correction -- it's a side effect of needing to track the dimmer's phase angle switching.

Your assumption that the light output will be any useful function of the supply voltage is unlikely to match reality.

If you simply cut the storage capacitor out of a standard CFL, you may find you now have a dimmable one. The snag is it will likely only have something like 30-60% dimming range, with no ability to run at full output.

The first requirement is that are high power factor so the dimmer can work with Triac dimmers

The second requirement is that the inverter be designed to operate correctly over a range of DC supply voltages.

You do not need HPF to use a Variac, but if the inverter is not designed correctly, then it will not operate correctly as you change the internal DC voltage with the Variac.

SMPSU ??? The systems I have seen are FIRST designed to have a high power factor input. That enables the use of a phase-cut Triac dimmer. In these systems, the internal DC voltage will vary with the setting of the dimmer. THEN, the inverter is designed to operate correctly over a wide range of DC power supply voltage.

If you simply cut out the energy storage capacitor, the output will be highly modulated at 100 or 120 Hz and there is no assurance that the inverter will work correctly as the DC voltage is varied.

No. Basically, a CFL is a power supply. It's designed to output 0.2A (or whatever) into a voltage of 70V (example, a guess).

If you vary the input voltage (with a variac), unless it's designed to behave nicely, and properly dim the light (which when a current of 0.1A flows will want 90V), it's probably just going to give up.

Random datapoint. A "PHILIPS Genie 6Yr 11W Energy Saver 230-240V 50-60Hz", on gradually the DC supply voltage, strikes at 82V. At 150V, it's using 9W of electricity, where my PSU runs out.

It does sort-of-dim, but not very much. On dropping the voltage, it's still producing light at 60V, more or less normally, finally going out at 40V.

Interestingly, observing the heater drive, on varying the voltage, it drops, then recovers to its original level.

Spiking the voltage to 300V, with a large flying capacitor, the heater voltage is still about the same.

This _may_ suggest that it's not going to get absolutely horrible life, when underdriven. The lamp voltage, as expected, rises as the input voltage drops.

6W, measurement error.

I'll measure it at nominal tomorrow.

Thank you for the feedback guys.

So it's going to be a "no". Oh well!

The high power factor SMPSU's I've seen (not in CFLs admittedly) seem to have a first stage which is simply missing the initial storage capacitor so they draw current over most of the waveform instead of just at the peaks. Energy storage to tide over the zero crossing region is then further downline.

Just like a conventional fluorescent lamp.

This is behaviour I have observed when the storage capacitor has failed in CFLs. There is indeed no assurance, which is why I said _may_.

I still don't know what you mean by SMPSU other than I suppose it is some type of power supply unit.

Yes, you can just remove the energy storage capacitor, but that will also remove about have the efficacy advantage of high frequency operation of fluorescent lamps. There has been a lot of work done on high power factor inverters that ALSO produce relatively ripple-free DC voltage. Early designs required two power stages, which increased the cost and decreased the efficacy. Later, single-stage high power factor inverters were developed.

A good review can be found in the PhD thesis of Jinrong Qian:

Figure 6.8c of Qian's theses (page 154) is similar to the ballasts used in many high power factor CFLs. Note that the lamp current is returned to the junction of two small capacitors that are connected between the full wave rectifier and the single energy storage capacitor. This allows these small capacitors to extract energy from the power line when it is low and push it into the energy storage capacitors which have a higher voltage at that instant of time.

The above was supposed to read: Yes, you can remove the energy storage capacitor, but that will also remove about half the energy advantage of high frequency operation.

[snip]

See also US 6,998,795, which is available at

This ballast users a valley-fill circuit combined with a charge pump.

My typing is terrible. "This ballast USES a valley-fill circuit combined with a charge pump.