Sine-wave Light Dimmer

I've not heard of a dimmer by that name, but if it dims an incandescent light bulb with a sine wave then its not the normal phase chopped signal but a variable voltage source...

I would assume its based around a Variac or variable transformer principle.

The Variac is basically a toroidal auto-transformer with the windings being contacted by a moveable wiper to give a variable ac output, which is a sine wave...

John.

Dear Electroman

Thank you for your reply.

As far as I know there are basicaly four types of light-dimmers: 1 - that you described was one of the first to be used in theaters and, yes, it were based in variacs; 2 - then, perhaps in the 70's or 80's, XX century, with the posiibility of making cheap and reliable TRIACs, those variacs were substituted by this new type, which works by varying the conduction angle of a mains phase; 3 - more recently, IGBT has taken place, though they use a similar method (inverted phase).

The problem with TRIACs is the high amount of harmonics (mainly the 3rd one) it introduces. With IGBT this problem persists though in a small scale.

Finaly,some years ago (maybe just two years ago) a new method arised: 4 - the sine-wave dimmer. It overcomes the harmonics problem because what varies is the amplitude (not the phase) of a sine wave.

Well, sorry for this large introduction, I´m trying to build a circuit like this but I can´t find any support on this issue...

Thanks, again, for your time

Carlmat

oh,

you can use a scr and change its troggle angle for dimmer. its easy.

I can send you a simple circuit

your address.

cnpower

I too have an interest in a sine wave controller which would reduce or eliminate RF harmonic interference and hash. May I also have a copy of the circuit schematic you mention?

lshurtle@juno.com

Lonnie

Just giving the subject a bit of thought, you could do the following if you are really intent upon it. This is horrible described in words but not that difficult in actuality if you are a bit electronically inclined:

1. Use a bridge rectifier to convert the incoming AC to rectified AC, i.e., all of the sine wave cusps are positive going referred to the negative output of the rectifier bridge.

2. Build a pulse width modulator at about 50-70 KHz, powering it from a dropping resistor, filter capacitor and shunt zener diode at about 15V, using the bridge rectifier positive source as the feed and the rectifier negative as the return. There are other, lower loss ways of powering it. Write to me privately if you want to explore that further. A CMOS 555 timer can be used as an astable multivibrator to do the job, and will consume little power, so the dropping resistor disspation will be low. There are plenty of circuits published to do this, and incorporate a pot to vary the duty cycle without substantial variation of the frequency.

3. Use a complementary NPN-PNP transistor follower to drive the gate of an N-channel power FET. At 50-70 KHz the FET drive current for fast rise times and low loss will be about 500 mA minimum. You can use a 2N4401-2N4403 combination, or some higher current bipolar transistors made by Zetex (see the Digi-Key website). Be very certain to bypass or decouple the complementary follower power source (same as for the 555) with a 10 to 20 uF low impedance electrolytic, like in the Nichicon PL series (see Digi-Key website or catalog). Drive both transistor bases in parallel from the 555 output, and connect the emitters together to drive the FET gate. Connect the PNP collector to the bridge rectifier negative (also the 555 power supply common connection) and the NPN collector to the 555 positive supply pin.

4. For the FET, the International Rectifier IRFP360LC is a good bet (also carried by Digi-Key as everything in this writeup is), 400V at current high enough for a 600W dimmer (considering cold filament inrush current). Connect the FET source pin to the bridge rectifier negative output, drive the FET gate with the complementary follower. But put a 10 to 33 ohm resistor in series with the gate lead to avoid parasitic oscillation in the FET. Such oscillation is a notorious problem with FETs due to gain well into the VHF region. Connect the drain of the FET to one side of the light bulbs, and return the light bulbs to the positive of the bridge rectifier. Be sure to heat sink the FET for about 10W dissipation. Caution: The metal backing on the rear of the FET is electrically hot, connected to the drain terminal. Get a "sil-pad" heat insulator also.

5. Connect a good quality capacitor of about 2 uF across the output of the bridge rectifier to present a low impedance source to the PWM'ed load at the 50-70 KHz frequency. A so-called "safety" FILM capacitor will work, as will a common 400V DC rated inexpensive polyester capacitor at the currents you will be involved with. At 2 uF, effect on the full wave rectified sine shape at 60 Hz will be minimal.

6. You now have your dimmer complete from a functional standpoint. At a given PWM duty cycle, you will replicate the sine wave in terms of line current, with a magnitude dependent upon duty cycle pot setting. Granted, the bulbs will be driven in pulse chopped fashion, but at 50-70 KHz instead of 120 Hz as with a triac. So EMI filtering will be possible with smaller parts. You can use small inductors in the output to mitigate EMI due to FET risetime effects, since you only need to filter the PWM frequency. The same is true on the line side of the bridge rectifier to abate any excessive conducted EMI which the 2 uF capacitor does not take care of. The bulbs are RMS responding and do not care in the least if the drive waveform is not a sine wave at frequencies where themal inertia of the filament averages out the temperature. But importantly, the line current will be sinusoidal, which is what you are looking for.

7. Reason for the rectifier is that the FET is unipolar. There are other schemes using no rectifer but using back to back FETs instead, but gate drive isolation becomes more complex.

There is some evidence that bulb filament life is slightly shorter on DC than AC due to tunsten migration to one end (positive, I believe) of the filament, thinning the other end. However, since you will be dimming bulbs, you will get extended life in any event. Believe it or not, bulb life varies as the 12th power of the RMS voltage.

During breadboarding and testing, you will need an oscilloscope, which does not need to be sophisticated. Run the dimmer circuit from an isolation transformer during commissioning, because you cannot ground the negative output of the bridge rectifier through the scope proble ground.

Honestly, doing a project like this is a good way to get your feet wet, so to speak, in a relatively simple electonic design project. Good luck in whichever way you choose to implement a sine wave dimmer.

Bernie Katz

ok fine,, might you have a circuit for the sine way dimmer??

thanks

jack g

jackg @california.com

Probably, connection PFC and buck converter in series solves this problem?

PFC gives a sine-wave input current, but buck converter adjusts a voltage across a load.

Krass

Krass:

Certainly you can do power factor correction and load control in two steps. OnSemi has application notes for ICs like the NCP1653 for PFC, but generally the output voltages are very high, typically 400VDC, since they are boost convertors. There are topologies for buck type PFC circuits, but the input current occurs in high pulses which are difficult to filter from an EMI standpoint.

However, you would then need an additional or second buck convertor to control the lamps. Since the load is well defined as resistive for a given filament temperature, it is more straightforward to do it in one step. Since the load is resistive and is tolerant of PWM imposed over 120 Hz rectifed power, the PWM and 120 Hz pulsating supply which the PWM chops is high enough in frequency to not affect filament temperature on a cycle by cycle basis, variable duty cycle pulse width modulation of the lamp drive, at a high frequency will offer the ability to dim lamps, and maintain essentially sinusoidal current waveforms at the mains input.

Bernie Katz

Bernie,

From my point of view to solve two simple problems is easier than one complex. It is possible to take output voltage of PFC equals to 440V and to use two lamps on 220V in series.

Krass

Thanks to you all.

It seems there are very good ideas and I'll try to explore them. I also think that PWM is the nearest solution for what I want. In fact, I'd already made such circuit based in a IC for buck converter from IR (IRS2540, I think...).

My problem is final stage. But you gave me some interesting ideas. Hum,... the rectifier bridge one...!

Thanks!

Somebody used a PIC (i think) to isolate the low voltage on the sine curve - used for a delicate welding machine.

The results suddenly became erratic. it was found that the electricity supplier superimposed a signal on the power to switch of the geysers in a specific area during peak demand periods. it is called a ripple switch.

The introduced the signal after transformers.