TRW pioneered the polypropylene cap for the Power Electronics industry, introducing them at POWERCON 5 or so. They are just about ideal electrically, with none of the esr vs. temperature problems of electrolytics. The only drawback was/is size: They could never get the polypropylene film thin enough to make a part below 100VDC, so the parts were not so successful as they had hoped. However, for your application, 100 Volt parts might be just right! At 28VAC, bridge rectified, you're looking at a 40 volt rail, and some overhead never hurts in a power design. Capacities are small, though, due to the volume just being the thickness of the dielectric plus that of the metallization (a current/heating constraint) times the area, which translates directly to capacitance. But the question would be whether you really need 1000uF. The largest part TRW made was 30uF, which would support maybe 1/4 amp if you could tolerate some AC ripple on the output; or were regulating down to a lower DC voltage. Multiple caps would of course multiply this current by N, the number of caps. As you suggest, these parts are solid body epoxy, and are only at hazard from a soldering iron as they are not encased. Finding them is not easy, as TRW sold off the line a while back. I think Cornell Dubilier is where to look! Mouser is the right distributor. So far as esl (series inductance) is concerned, these things resonate up in the hundreds of KHz, vs. 2KHz for electrolytics, so it's no problem! Sorry to be so long winded....