Regarding ripple current specifications, I have designed a high ripple application, indeed it is akin to KHz frequency spot welding given the currents pulsing to 800A or so. I have found that the Illinois Capacitor PJ series do handle the duty which runs them at up to 60 C continuous (25C ambient). They do better than Rubycon etc.
They are in parallel banks of 3 @ 2200uF with solder braid augmented PCB traces spec'd to sub 2 m¦¸ per trace. I pre-qualify the capacitors with an ESR meter to sub milliohm? grading and binning for bank balancing.
As I look at the data sheet again, I note they are specified for SMPS duty.
These capacitors are hammered continuously for 100s of hours per month. I have not had one fail, except when I put in one backward which then pops it in about 30 seconds.
Here is an O'scope 'blue' trace of a pulse event using a 50 ¦¸ impedance @ 40A/V? Ion Physics (purpose built) current transformer monitor:
I have systems running in Africa where the ambient temperature in the shade is 45C.
Ancel
