Fluorescent Lamp Driver

I would place a power on button and detune a little one of the 47nF divider at right side to help startup! Nice circuit even EC does not provide multitapped transformers.

Yes the startup circuit is missing. I would use a diac for that as it is common practice in commercial cfl circuits. And yes a way to build magnetic circuits with simple inductors would be nice.

Is this circuit for CFL or for CCFL?

What's the specs on the bulb here?

CFL it is. The filaments are on both sides of the 1nF capacitor in series

I haven't tried to build this circuit yet. Thus I have no idea what lamp you could use for this. The ballast coil also needs to be adjusted to the particular lamp. In real life you would probably also increase the ignition capacitor from 1nF to 6.5nF or something an tinker with the emitter resistors and windings to fine tune the operating frequency. There is a quite complicated numerical calculation behind all this. I'm reading this up soon.

Does it needs a heating phase at startup? Or does it run constant in this mode?

A heating phase is Requierd but very short. If the filaments are warm enough from the resonance current the tube will start to conduct which bridges over the capacitor and filaments which changes the resonance frequency and clamps the power output to the tubes specific value.

Starting phase > 1kV, burn phase ~ 84V. The current depends on the tube resistance of course.

The startup is identical to normal FL , only difference is the frequency? You know, the heating phase at start is done by connecting both filaments in series and wait a while. Then it opens and a chocke coils makes a back emf 1 to 3kV which actually starts the tube to conduct and the voltage drops across tube to around 100V additional ~100V are dropped across the ballast chocke 4H and ~100Ohm. This is the old methode. I try to understand what different new variants exsist ;-) from heating (if needed) to start the plasma and operating.

Well there is also a resonance capacitor across the switch. The switch is a neon lamp in which one Elektrode is a bimetal strip.

In many cfl circuits there is a ptc for this job

???

Never mind

In old circuits the warming via filaments decreased the resistance too slowly for a high voltage peak to form. Thats why they had a starter switch bulb which would normall be open which heated up and instantly closed the filament circuit. the closed switch would then cool off and again instantly snap back in to its starting position which opened the circuit fast enough to create a voltage spike which would then ionize the preheated tube.

I know how the old systems work, even i dont get what you mean with “decreased the resistance too slowly for a high voltage peak“. Filaments do have a resistance of about 10s of Ohm. Sure the starter is always needed in old system to generate 2kV or 3kV. Cuz the 325V are to low for ionisation/lawien effect. But this is independent need from the filament resistance. What i looking for is a good explanation how new technic is doing the complete process of heating (if needed as separated step or precondition) and start with 2kV or continues running at 2kV high frequency 50-100kHz? The complete process from.off to on, step by step.

Hier kannste haben: http://educypedia.karadimov.info/library/AN00048.pdf

A continuous run in 2kV mode would eat up the filaments like crazy because of the high temperatures and ionic diffusion or something. It would also make the walls of the tube black in the process via condensation. Not to say that the efficiency would be terrible. Real ccfl tubes have a different material for the electrodes and a different gas mixture. Never the less they can be a terrible waste of power.

Mir sind vor einigen Wochen drei von den Dingern in die Hand gedrückt worden, mit dem Hinweis "der Scheiss ist schon kaputt". Da war dann die Frage, was ist denn eigentlich kaputt? Die Röhre, oder die Elektronik? Habe dann alle drei zerlegt und zwei davon konnte ich mit der alten Balast und Starter Methode zum Leben erwecken. Eine wollte aber garnicht. Also grundsätzlich kann man bei den Dingern einen normalen Startvorgang mit 2kV machen und dann über Balast betreiben. Wie lange weiss ich natürlich nicht. Aber danke für die AppNote, die sieht gut aus!

Ja bei mir ist es genau umgekehrt. Ich hab nen Haufen von den alten Dingern bekommen und wollte mal gucken ob ich die upgraden kann. MJE13009 ist ein passender transistor der verfügbar ist habe ich festgestellt. Das einzige was mir fehlt sind die DIACs aber da lass ich mir was einfallen.

Ich hab auch schon Schaltungen gesehen bei denen nur eine Seite geheizt wurde und die andere überbrückt war. Vielleicht kannst du damit die dritte zum laufen bringen.

Mit max 400V Uce erscheint mir der MJE13009 etwas zu anfällig für derartige Schaltungen. Aber warum überhaupt ein upgrade und auf was upgrade?

Wenn man die feedback Windungen vergrößert kann man die Betriebsspannung verringern. Ich forsche noch. Im Endeffekt möchte ich einen leistungsstarken 12V cfl Treiber entwerfen der nicht die schlimmen Verluste der bereits in Taschenlampen verbauten Inverterschaltung besitzt. Zumal diese Inverterschalungen die Leuchtstofflampen einseiting stark erodieren und somit eine geringere lebensdauer haben.

Die 400V reichen aus da erst hinter der Drossel die großen resonansüberhöhungen auftreten.

Ich habe vor kurzem eine 12V CFL gesehen. Sehen identisch aus wie die 230V dinger. Also die scheint es kaufbar zu geben. Naja, selber basteln macht natürlich mehr Spaß ;-)

Genau. Frage keinen Bastler warum, die Antwort ist immer "Weil ich es kann!" ;-)

Das grenzt an Philosophy!