I'm happy to know you're running again, but I'm betting you didn't get an answer fast because the facts we're incomplete and a bit confusing in how it was stated. I can't believe it slipped between the cracks and didn't get answered or commented on... if there was ever a reason to bump a post to the top I'd say it was to get help for someone. (Next time, I'd suggest a second posting of a help question if nobody responds in a day or two )
The terminology comes with experience and I don't want to discourage, because there is only one way to get experience...this is in hopes to clarify for your knowledge understanding, and for others reading too. This is basic stuff, useful to postwar & modern alike.
I'm not sure if you left a board in place, but a full wave bridge rectifier *should have worked if there is no board. It would work with a board too unless it is an ac only trolley which I highly doubt. (It is possible to use only the neg wave to "think" and only the pos. for other things.. Kinda rare in trains though)
Something was wrong. Be it component choice, bad component, or wiring, something wasn't right. A standard rectifier or diode should have worked; nothing special needed. "Special" could be the issue, all we need is a "one way check valve". (Current blocking? ... Offhand this sounds like current limiting, which would not be good for most motors. [easier to be basic and right than search to explain why yours might not be right])
*A can motor with a "grounded case" connected to a metal chassis would not be able to use a full wave rectifier, but would need half wave (what you have with the diode). It would also not reverse without a second diode added. An isolated motor would work with either.
The half wave is "lumpy dc" letting only the positive ac wave pass. Full wave flips the neg. ac wave makeing it positive, and giving two lumps of dc per cycle, so is smoother.
A capacitor can do two things here depending on the type. For power(can type) it acts as a small battery, charging during wave peaks, and letting power loose when as input goes low. (Would need to be bi-polar for two directions). Or for a filter (disk types to stop the tv/radio fuzz that motor fields cause). The smoothing cap only leaves tiny peak bumps, very close to pure DC. Half wave can have a smoothing cap too, but I would need to be very much larger to bridge the wide gap... Just not done often.
Some motors will run cooler on half wave because the missing wave allows "off time" where it coasts. Others will run hotter needing the smoother full wave to do work more efficiently. Tests are often the only way know. (We don't use 3phase)