CFL Driver Circuit TESTING STAGE

[Flux]

Trying to drive fluorescents from a forward converter without ballast is asking for trouble. You may manage with enough leakage reactance in the transformer to act as the ballast but it is pure chance.

The single ended things that work as flyback converters are safe, but they are not efficient and blacken the tube ends.

It is possible to make push pull self oscillating converters  that are resonant, they are reasonably efficient but they murder the transistors if the tube fails to strike. I have had some running for 20 years but I wouldn't build any more that way.

I now use driven forward converters with resonant series ballast ( 3525 does this ok at 12v. IR lamp driver chips are better for 24v)

These work very well and with excellent efficiency but the resonant balast is critically dependent on the components and without specifying an exact inductor and winding scheme it would not be reproducible.

Unless you get these lamp drivers right the actual efficiency in lumens/watt can be so bad that you would be better off with halogen lamps. It is fairly easy to make something that will light but to do it properly with good lamp life and efficiency is a different matter.

Many of the low voltage drivers I have seen for marine and camping use leave more to be desired than I am prepared to tolerate.

Flux

[Slingshot]

If you provide so much overvoltage, what limits the current after the bulb starts?

[s4w2099]

the ballast

[s4w2099]

Ok here it is, work in progress.

http://rabilco.com/ahome/cflVideoV1.MPG

My ferrite core transformer just didnt worked so I am using a standard transformer while the one that I ordered arrives.

As you can see it is a good start since you can already power large tubes with a ballast. For testing I am using a 14Watt bulb (in the video I said it was 10W coz I was expecting some loss but when I read the bulb it said 14Watts)and the current was 1.2Amps in the DC side. The frequency is about 8-10khz and the xformer is vibrating and making a buzzing sound naturally.

Lucky me that the bulb was outputing full intensity. I dont have a tube with a ballast so I will see if I can get one this afternoon to actually test with a ballast.

It should work as well. At least for now its a full wave. This same circuit with a ferrite transformer can be used as a boost converter. It would be nice.

[alancorey]

Around here at Home Depot they're $10 for a package of 6, at least the 14 watt kind.  They've been at that price for over a year.  In my local store those are on the end of an aisle, not with the other CFLs.

  Alan

[s4w2099]

I will check that out I hope they have it here too.

Thanx

[s4w2099]

Ok, flux. I have done it. I got a ballast and made the circuitry. There is a little discrepancy though. The DC side says that I am using up 17 Watts and the AC side says 18 Watts so I dont know how to calculate the efficiency on this one.

Both of my meters are different models so maybe thats the problem. I am still waiting on the ferrite core transformer so that I can actually run it at 20Khz.

I have a question. will the ballast tolerate that kind of frequency? in the label it specifies that its to be used with 60Hz. Injecting 60 Hz to it and it works wonderfully.

[Slingshot]

Where is "the ballast" in this circuit?

[s4w2099]

The ballast is inside the bulb. Its an electronic ballast if you are talking about the circuit in the video. I have another video actually powering a larger tube with an inductive ballast that you can get at the home depot.

[Slingshot]

I realize you are calling this entire circuit an "electronic ballast", but I'm asking what is the mechanism for current limiting, since you say you design for a large amount of overvoltage.  What component(s), in the posted schematic, set the steady-state bulb current?

[s4w2099]

Ohh sorry you are talking about the first circuit posted. The first circuit does not have a ballast and the only current limiting factor in this one is the horribly small transformer and the losses from running it at high frequency.

In a decent setup actually the ballast (which is another part of the lamp separate from the driver circuitry) will limit the current to the bulb. It will initially  send a high voltage pulse to get the bulb on and after that the current and the voltage is lowered just to keep it on.

There is no control with the 555 circuitry that I made, plus its only a half wave design. I THINK that not having a current limiting factor such as a ballast will increase the chances of blackening the tube's ends.

The version that I am working on right now has all the features from a very reliable system. I will post as soon as its finished.