| You are wrong - electronic use coil(s) as well...
Principally the electronic work exactly in the same way as what is called "magnetic", only the frequency is "a bit" higher - few tens of kHz (iso 50 or 60Hz for "magnetic").
The energy goes first to a rectifier, what make about 300..400V DC (called "DC bus") from the mains (EU use bridge, US doubler, or both an active boost PFC - in case of 400V).
This is a supply for two transistor halfbridge, what generate 150..200V (half the DC bus voltage).
Then is connected the lamp in series with inductive reactor, while this is the main ballasting component.
The only principal difference from mains frequency serial reactor is, the operating frequency is not fixed, it might be changed. This feature is used to control the impedance of the ballasting choke, so the lamp current, either to do dimming or to compensate for supply voltage changes.
The possibility to adopt frequency to actual need's is used for ignition: In parallel to the lamp is connected a capacitor, what form with the choke series resonant circuity, what if driven on it's resonant frequency cause current and voltage to rise to really high levels - viola, we have high voltage for ignition. The control circuity has to ensure, the ballast is on (or close to) this frequency during an ignition attempt. Simple CFL ballasts use this resonance by their feedback mechanism as the frequency control components during open-load (as not ignited lamp is open load), advanced ballast (without such simple feedback) slowly sweep around the resonance, what insures high voltage, so ignition as well.
In order to preheat cathodes, filaments are either connected in series with the resonant capacitor (current mode or series heating) or to ballasting coil's auxiliary secondary windings (voltage mode or parallel heating).
Simplest ballast do operate in the resonance till the lamp strike, so apply high voltage at the same time with high heater current, what correspond to what "rapid start" ballasts do. Only the voltage is usually so high, the discharge ignites before electrodes heat up, what make them effectively instant start, with everything it mean. These ballasts use exclusively series heating, due to two reasons: At first it is the simplest, so cheapest setup. At second when filament broke (or the lamp is removed), the series resonant circuit is interrupted, what interrupt the feedback, so shut the ballast down. This work as the main and often only end-of-lamp-life ballast protection, assuming when the emission material wear off, the associated cathode fall (together with increased resonator capacitor current) would overheat the rest of the filament, so it break, so shut down. Otherwise the resonance high voltage and current condition would overheat components of the HF generator, causing them to fail. So if such simple protection has to work reliably, components in the circuity should be overrated, to ensure they will not fail (or excessively wear) before the broken filament shut it down. And this is the bottle-neck of most CFL's: they use this simplest design, while components are rated just for normal operation, so they fail immediately cathodes do so.
When preheat before strike attempt is wanted (to prolong lamp cycle life), mostly it is in the form of a PTC (positive temperature coefficient resistor) connected in paralel with the resonant capacitor. After power ON, PTC is cold, so short the arc, leaving filaments only in it, the ballast current heating them up. During this time, the PTC heat up too, till it reach the temperature of it's thermal runaway (incresed resistance increase power dissipation, what heat it up quicker, so the R increase further), where the PTC effectively interrupt the current, allowing the resonance, so ignition to happen. And viola, we have the simplest "programmed start" ballast. The "preheat time" is then the time the PTC need to reach the switch-off temperature. The draw-back of this method is, then after power off the PTC need time to cool down in order to set the proper preheat time for next ignition, usually at least 15 minutes. When this is not met, the preheat time might be insufficient to heat up electrodes, so the lamp start "hard" (as instant start). This is the most complex ballast i've seen on CFL's sold below $10 and often used on way much more expensive types.
Intelligent ballasts use frequency control for the preheat too: When the circuity is operated above, but not too far from the resonance, there is still enough current trough the capacitor (or voltage on the inductor), to heat up electrodes, but not yet enough voltage to strike the arc. After what is programmed as "preheat time" the frequency sweep down trough the resonance, where the ignition happen. As the preheat time is controlled in pure electronic, a power-on reset is easy to implement, so the preheat time is correct independently from the time the lamp was OFF before restrike, so when correctly set, the lamp is always properly preheated, so has virtually infinite cycle life.
The only mechanism limiting cycle life on intelligent ballasts is the preheat temperature accuracy: Too low cause "harder" starting (so emitter sputter), too high excessive evaporation (so again accelerated wear). Here come the advantage of voltage mode heating:
As the tungsten (as most metals) has positive temperature coefficient, the filament resistance increase with it's temperature.
So when the filament is supplied from a voltage source, the rising temperature increase it's resistance, so lower heating current, so the power input, efficiently stabilizing it's temperature (current mode heating causes the opposite, so amplify all tolerance's influence on the temperature). This stabilizing mechanism make the filament temperature quite insensitive to voltage, time and filament parameters (wear,...) tolerances, ensuring it is much closer to the optimum (yielding lowest wear) preheat temperature. Such ballast designs allow the fluorescent lamp to be switched ON and OFF as often as desired without adverse influence to it's lifetime, but it is the most complex, so expensive control scheme. On CFL's i've seen it used only when they are dimmable - during dimming the heating power from arc current is insufficient to maintain required cathode temperature, so auxiliary heating is required and the voltage mode is usable way to keep the temperature across all possible operating range (mains voltage, dimming level, lamp age,...)
|
