unusual dimmer switch design

[EinarA]

Thank you Throbscottle. But thanks to Nigel as well as he answered your main questions.

 

[Nigel Goodwin]

Nigel: I have a circuit written down on paper. I was hoping I would not have to get involved with uploading something, but it seems I will because Al is totally confused.
Al: the transistors are in a totem pole arrangement , neither has any DC bias, they are driven by the transformer thru a 10 ohm base resistor and a 1 ohm emitter R. ( Don't know why they have both.) Without the trigger diode this arrangement will never start oscillating. The output of the Osc drives the transformer primary and the main ballast inductor in series, with these in series with the lamp. ( There is also a diode back to the startup cap to keep it discharged. The startup delay is just a byproduct of this arrangement.)( The designers don't care if the tube starts or not.)

I went through all the circuits on the link I posted, and managed to find a few with diacs, particularly Osram ones

 

[EinarA]

I'll try uploading a pic of the circuit.

 

[EinarA]

Well that wasn't too hard. I see what I have is largely the same as the Osram circuit. This is a minimum parts design so I don't think the Chinese are going to stop using it anytime soon.

 

[Nigel Goodwin]

Well that wasn't too hard. I see what I have is largely the same as the Osram circuit. This is a minimum parts design so I don't think the Chinese are going to stop using it anytime soon.

There are plenty of lower parts count designs, this ones seems rather strange - why build one that needs a special start-up circuit?.

Your circuit won't work though, as it's on it's side and all the electrons will fall out

 

[MrAl]

Hi,

Yeah, haha, and i could not read the schematic unless i laid down on the desk sideways viewing the monitor

Anyways, here's a better drawing i found and now we can finally see a clearer picture of what is probably happening. It appears that the diac probably is being used to start to the oscillator because we can see the transistor it turns on is right across the lines getting to the tube. So i think i have to agree for now.

So it goes something like this:
As the device is turned on, C1 charges up quite fast, then the diac fires, then the associated transistor turns on, and when it turns on it causes the diode to conduct which then pulls the diac voltage down again below it's firing voltage, so the transistor turns off again. As it turns off the voltage across the tube circuit rises sharply and that causes the resonant circuit to operate at the resonant frequency controlled partly by the inductance and partly by the smaller series capacitor and the small cap is there just for that. The resonance causes the voltage to rise up and fire the tube, and once it fires the resistance comes down and the voltage comes down and the current is limited by the inductance (another necessary feature).

https://en.wikipedia.org/wiki/File:Fluorescent_Lamp_Inverter.png

 

[EinarA]

The transformer provides positive feedback , as soon as the transistor is on ,the base winding latches it on until the core saturates. Then the transistor loses it's base current and turns off. This allows the other one to turn on and again latch on until the core saturates in the other direction. This is a very old circuit that fell out of use 50 years ago. It is the saturation point of the transformer that primarily controls the current through the tube. I think it is this aspect of the circuit that answers Nigel's question of why.( PS: I got the phasing dots on my diagram wrong, they are correct on the Osram circuit.) The Osc continues running while the voltage builds up across the tube; simultaneously the heaters are warming up the emitters . After several cycles there is enough voltage to make the tube conduct. After 5-10 secs the PTC heats up enough to shut off the heaters and send most of the power through the gas.
I still feel that from the Mfgs point of view this is the lowest cost circuit. The two I took apart were from different Mfgs and ages but had identical circuits. Admittedly not sufficiently interested to delve further ; they are still worth salvaging no matter what you find.

 

[MrAl]

Hi,

Yeah the price has come down so much in recent years it doesnt pay to try to fix them anymore.

I had one break off at the base though
I had it plugged into a ceiling light, and of course it hung down a little from the ceiling. When i was workingg on the car one time i carried a rather large cardboard box across the room and was in a hurry. The cardboard box came into contact with the lamp and broke the thing off at the base. It was a nice curly bulb too, bigger than most. It probably still works i'd just have to epoxy it with high temp epoxy, but im just not into repairing these things.

 

[EinarA]

One thing I didn't think of until earlier today is that in the circuit I posted the thermistor is low enough in value to keep the bulb from lighting until it warms up, and in fact this lamp took a couple of seconds to light. So the thermistor heats up faster than I said above and the slower term increase in light output must be from the gas and phosphor heating up.