I found that the output was more impacted by changes on the input than I like, and you already know I'm fussy about ripple. I also found that if I had a part of the load (~100 mA) that was switched on and off repeatedly, the effect of that load change did appear on the supply output.

The thing I used, and do, occasionally use, still, was the little 2 cm x 6 cm simple-switcher sample board (I'll send you a picture if I see the thing any time soon) that they (NSC) distributed back in the early '90's (?) when the simple switchers were new. Since it's a 5-volt regulator, and since that's what I wanted, I didn't try it as a pre-regulator. Switchers generally don't like largely capacitive loads, however, so I wouldn't rely entirely on a cap to suppress load changes. A resistor in series with the switch and supported by a cap would probably work, though, to suppress fairly significant (20%) load peaks.

RE