Hi all, I recently ran some tests which involved running loads of significantly varying arc voltage on my two HPF F40T12 rapid start ballasts, my vintage CGE and my Advance R-140-1-TP. I found a side effect in running these tests which, interestingly, is handled very differently between the two ballasts.

As we all probably know, as arc voltage goes down on an HPF ballast, total system power goes slightly down, as well. Mains voltage remains constant, so that means either line current or power factor must be going down. I found that on the Advance, line current goes down and power factor remains the same, while on the CGE, line current remains the same, and power factor goes down!

Does anyone suppose how this difference in power regulation works? Perhaps a difference in where the capacitor is wired?

If your lamp circuit operates at high Q (so usually rather high OCV) and the ballast is of HPF, then all the reactive power bounces just inside the ballast and the mains provides just what the lamp extract. Because of the high OCV, the lamp appear jus as a short circuit in there, so the actual arc voltage has nearly no effect on the current.
This situation could be mimicked by an active power regulation, e.g. using saturated magnetic (the HID CWA concept; there is no reason to not use it for fluorewscents as well). This allows to keep the arc current pretty constant over really wide range of both arc, as well even mains voltages. And when the power factor correction (an air gap in the main magnetic circuit) is designed correctly, it compensates out the power factor over quite wide range of arc voltages.

On the other hand a HX autotransformer style ballasts use no active regulation at all (just a fixed impedance), the setup would have to feature really high OCV to supress the sensitivity. But that would cost losses. So as a consequence, the arc current does depend on the arc voltage, what makes the power factor correction unbalanced, once you deviate from a designed operating point (= arc voltage). So even when the ballast may be of a HPF type (because it uses a power factor correction capacitor), as it becomes unbalanced, the altering arc voltage result into just power factor reduction...

Interesting... So you say the CGE ballast which alters power factor might have the capacitor set up HX-style? That makes sense. Although that's strange because when I try to measure the OCV, the voltage reading jumps around all over the place, like a normal CWA-style ballast does.

Or it just does not utilize the magnetic shunt saturation to regulate the arc current (e.g. it has no explcit magnetic shunt at all).

The difference would be really just the dimmensionning of the small insert between the primary and secondary winding, otherwise all the other components could be exactly the same...