Ron H said:
What about a transformer driving a half wave rectifier?
Hey, hey, hey...
The EMF will still be there but the current will flow in only one direction since impendance of a positive polarised dieode is much much smaller than imp. of the negative polarised diode. that's just it, but the emf will be there.
Regardless of the quality of diode used (meaning low low leakage current) you'll not get any DC buildup because winding is almost a short circuit for any DC component.
If it appears then You've got a problem elsewhere.
But then again if the pulses are short and far away allowing the atmosphere in the lamp to cool down before next cycle then you'll get cold lamp on each positive half-cycle, which means it will have a higher current rush later in the cycle. That will certainly damage one of the electrodes.
Another thing. and i'm on the edge of shooting my foot now:
-Since you do intend to get the best of the transformer, you'll use the frequency as high as convenient, too low the freq. and you'll waste money, too high and you'll heat up the core thus wasting energy.
-If you follow this, you'll see that usually voltage rise-up and fall-down times of the oscillator are much smaller than current rise up and fall off times. (Which is of course governed by the Z of the tansformer.) And current rise/fall times of the transformer are equal, that's it's characteristic. (The strange thing is that I'm lying right now, there is a slight difference due to hysteresis of the core, since we're driving dhe flux in only one direction, but believe me in these cases you'll not see it.)
So the point is JUST TUNE THE DAMN THING WELL AND IT WILL WORK.
P.S.
Thanks guys, just for laughs I've checked my cellar and found there something beautiful: The MAGNETIC AMPLIFIER, I myself would never have had heard of it if not for some old centrifuges, but that's another story now.