The inductance formulae make parallel branches our friend, as the Warsaw machine’s cap bank demonstrates. This idea also pre-supposes that the switching problems have been reasonably solved so it can re-direct the capacitive fraction (which would have been returned to the cap bank) and all outputs into firing the next stage while providing some useful output. Q would have to be >1.5, I’m guessing. But if we’re ever up against a cap lifetime issue, this may be a workaround to some degree.
