Hey Ash, yeah i have 120v here! I always wondered this, can a regular fluorescent sustain arc at 120 volts if it is ignited with a few kV like an inductance kick, or does the arc need more juice hence auto transformer? This is interesting! I did open one old fixture up but didn't see a separate cap, would I have had to open the black ballast case to find the capacitor? Thanks man!
The discharges usually need moreless constant current source to let the arc control the voltage. The current source have to have enough voltage room, so the current is not influenced by the eventual arc voltage fluctuations, otherwise the arc become unstable.
But with AC supply, the current have to sometimes change the polarity. That mean in one moment it drop to zero, so the arc extinguish. But to remain lit continuously (as perceived by the users), it have to immediately restrike for the opposite polarity. For this reignition, the ballast should provide an OCV about twice the arc voltage just in the moment, when the current disappear, or as soon as possible. This is the reason, why the inductance is so important in the ballast impedance - the inductance change current only, when there is a voltage across it. So during the zero cross, the rest of the circuit should be at an instant with higher voltage (to allow the inductor current to change, so drop to zero), while the same voltage is then available for the lamp reignition just after the zero cross. As the ionization in the lamp does stay some while after the current disappear (but it is decaying), such reignition by far does not require so high voltage as when the lamp is turned ON, usually ~20..30% above the arc voltage reignite it. But for the reliability, the available voltage have to be twice the arcV, as if it does not reignite, it mean the lamp goes out and need to be normally restarted (one misfire in 1 million already mean the lamp extinguish after few hours)
At the same time the current should not vary as much with mains and/or arc voltage fluctuations. So when assuming the ballast is a linear impedance and the OCV is twice the lamp arc voltage, the current change about 7% with 5% mains voltage change, what is quite a lot. If usual core nonlinearity is added, the change is usually about 10% for 5% voltage change, what is just marginal to be accepted.
At the same time the ballast should provide clear differerence in some parameter for the ignition device to distinguish running from not existent arc.
Normal glowbottle starters and HPS electronic ignitors use the voltage across the lamp (higher than some threshold mean arc is OFF, so the starter kick in, lower voltage mean the lamp is running, so starter stand by). But that require the ballast to provide enough OCV, to allow the starter threshold to be set with sufficient margin above the lamp arc voltage (to not interfere with running lamp), yet the OCV to be way above the starter trhreshold (to kick in after power ON). This ask for at least twice the arc voltage as well.
Now this could be accomplished by few ways:
The easiest is to use the lamp of maximum half of the mains voltage and simple series inductor
Use an autotransformer with output about twice the lamp arc voltage and a series inductance.
Or one option remain for situations, where the mains is at least 20% higher than the arc voltage: Use series LC, operating near the resonance at low current, but the coil softly saturating at the desired operating current. The inductor saturation then move the resionance further above the mains frequency, increasing the impedance, so lowering the current. This mechanism then regulate pretty constant current even when the voltage headroom of the mains above the arc voltage is quite small.
The components should be designed so, the voltage across the capacitor on the normal current is about the mains voltage.
In this configuration during the current zero cross the capacitor's voltage add up to the rather low mains, where this sum is the voltage available for the past zero cross reignition.
But as the capacitor provide it's voltage only after there were any current, it is not present at power ON. So this circuit does not provide sufficient voltage for the starter to recognize the present/missing arc by the voltage, so it require the circuit current to be used as the control "signal" for the starter. So it need a thermal starter, glowbottles won't work here.