After working hard to get my RIS-796A "analog" high-current source working to 500A for 10ms, for 8.3 ms single half sine-wave forward surge current tests, I realized the on-board low-esr caps, although happy delivering 500A currents, quickly run out of charge for such long pulses.
In fact, my 300-watt Xantrex bench supply, 15V, 20A, can't handle the task. For that duration, I need an honest 10-volt 500-amp power source. OK, surely stacks of hefty 18650 cells can do the job? No way, too high internal resistance.
I can stack six surplus 1kW 5V, 200A supplies, but whoa, that's a lot of heft, for only 8ms.
Aha, eight DGH407Q2R7 supercaps series-parallel 400F 2.7V can do it, 2.8mR each, for 5.6mR net, 2.8V esr drop at 500A, +0.1V dV/dt drop in 8.3ms. Don't take much space! The Xantrex supply can keep them charged.
But oops, better add some serious overvoltage protection, to limit the 10V capacitor bank to 10.8V, absolute max. 20A shunt protection to limit my Xantrex knob twirling to 10.2 volts, with shutoff, to avoid making the overvoltage protection circuit dissipate 200W.
Been there, done that? Let us know about it.