On 50/60Hz sinewave this would be OK, these (motor run) are designed to be operated at the ~400V for the whole lifetime.
What I'm not sure are the spikes coming from the lamp reignition. As motot run, they are not exposed to any distortion, just the mains frequency sinewave, so they do not have to be mechanically tough enough to handle the spikes (it is the mechanical forces resulting from the peak currents, what could be the problem here).

But anyway, what would fail as first with such spikes would be the lamps, not the capacitor...

You could try a superimposed ignitor, I've had good results with these but only use it with good quality ballasts as they can arc and possibly flash over in cheap ones.

Using something like this, I wonder if you could make a better, brighter, less lamp-killing version of those LOA shoplight ballasts.  Or what about lighting a F96T12 slimline or HO lamp from 240v or 277v?

The LOA system uses a capacitor in series with the lamp itself in order to artificially increase OCV. What I attempted here was just using a capacitor in series with the cathodes to create a resonant starting system.

With that being said, if you were to mimic the LOA system, but use a smaller capacitance (just enough to make the artificial OCV high enough), then it indeed should be better for the lamp. And of course, you'd use a MUCH better choke which wouldn't burn up...

Regarding slimlines, if 240V or 277V isn't enough to naturally instant start them, then no, the capacitor won't help, because the real OCV is still only 240V or 277V. The OCV is only artificially increased once current starts flowing, so the lamp needs to be preheated in order for the whole system to work.

The capacitor in series with the lamp does bring the OCV up by a resonant effect (well, it operates off the actual resonance, but the voltage boost is there). Therefore you can not make it much smaller, it won't be just the lower current, but the OCV will drop as well.
There the capacitor forms the main ballasting impedance, with part of it being subtracted by the inductor in series (either the ballast choke, or the leakage inductance of the autotransformer).

For reasonable trade off between efficiency vs voltage boost, the system is set so, the capacitor voltage is about a sum of mains (or transformer OCV) and arc voltage.

But whatever is the desired setting, the coil inductance and capacitor must BOTH match to each other, otherwise the effect will either be too strong and fry the ballast itself, or too weak to be able to operate.


So it is really far different case than the capacitor in place of a starter switch.

Interesting. I agree, this may kill the ballast, lamp, or the cap itself.