| Peaking capacitor is a low value (about 100nF or so) capacitor connected parallel to the lamp output, either direftly or just in front of the series capacitor.
The idea is, as the current disappears when approaching the current zero crossing point, the voltage across the lamp is still the arc voltage of the polarity before the zero cross, but the ocv of the ballast is already of the opposite polarity. So by adding the small capacitor, the output inductance of the ballast (the high frequency behavior counts here) forms a resonant circuit with the small capacitor, causing at the instant the current zero cross the voltage to swing from the previous arc voltage (assume it was +100V) "arround the ballast ocv" (lets say -100V, to make the math in this example simple). When neglecting losses, it will really swing twice the whole voltage difference, so from +100V to -300V (instead of the -100V of the ocv alone at that point), forming a kind of voltage peak (hence the "peaking capacitor" name). In reality there are losses involved, the energy gets lost so the voltage is lower, but still it is higher than without that capacitor.
The idea is, this extra voltage then helps to reignite the arc for the opposite polarity faster.
But EYE uses in their retrofit lamps a parallel ignitor (in the form of a ferroelectric capacitor), which us supposed to generate near high voltage spike (in the 700V range) every time the voltage exceeds the FEC saturation voltage (the voltage where the complete active dielectric is polarized, so the capacitance drops, causing the charging current to suddenly drop, co cause inductive overshoot). Because these spikes are rather low energetic, the 100nF of a peaking capacitor would just plain short them out, so the FEC wont be able to ignite the lamp anymore.
Similar problem would be with the thermal "starters" some other makers use (like the glowbottles in Osram MV retrofit HPS). There the current with tye switch ON uses to be limited by a resistor, so the kickback is energy limited as well, so the "peaking capacitor" would kill that voltage kick as well.
So how to tell if a ballast has the "peaking capacitor"? You will see it connected there.
100nF (rated at at least 400VAC) won't be that small to be hidden within the winding or so, it would be a distinct component (a small box or cylinder with two wires from it).
Note many superimposed ignitors actually use internal capacitor arrangement so they actually act as that peaking capacitor. But they are connected before the HV secondary winding, so do not short the HV pulse generated by that ignitor (on the contrary, they provide good return path for the HF ignition current spike from the Neutral back to the pulse transformer HV secondary, so it does not have the ballast impedance in series).
|
