The transformer has two sections: a core section around the primary winding, and a section around the secondary. These are coupled somewhat looser than a regular transformer, so (without the capacitor) the secondary voltage would be lower and "squishier" than a regular transformer (i.e., higher series inductance). Adding the correct value capacitor cancels the series inductance, forming a resonant tank. But the increased voltage difference across the barrier forces the transformer deeper into saturation, thus limiting voltage.
The transformer needs to be designed so that, over the design range of input voltage (usually +/-10%) and load (0-100% current), the transformer must remain in saturation, without overheating. Finally, because the secondary voltage is generally lower (in terms of volts/turn), a lot more copper is needed. Iron operated in saturation also has high core losses. This makes these transformers particularly large and low in effeciency.
One upside: the effective LC filter between primary and secondary isolates harmonics and transients; although the input and output are still a bit distorted, they make reasonable sine waves, so it doesn't matter much if your primary waveform is part square wave, or that your load has similar behavior (like a capacitor-input rectifier).
Tim
--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms
"amdx" wrote in message
news:2d5eb$50203c99$18d67c20$18756@KNOLOGY.NET...
>I recently repaired a marine battery charger. It was set up to charge 3
>separate 12v batteries. The repair was simple the charger had a *shorted
> 40uf 660v ac capacitor. The capacitor is connected across a winding on
> the transformer. This is the usual constant voltage transformer
> configuration. This is the extent of what I think I know.
>
> Can someone describe the operation of a constant voltage transformer
> that has 3 windings, input, output, and a third winding that is connected
> across a capacitor.
> I've seen these for decades but never have understood the operation.
>
> Mikek
>
> *bonus, why didn't the shorted capacitor (**measured 1.5 ohms) blow a fuse
> or overheat the winding on the transformer.
>
> **1.5 ohms plus 0.7 ohms lead resistance.