take a good look at the slip rings, they should be a bright brassy color
if they are dark, chocolate color, or greenish they need to be cleaned
you can use some 400 wet or dry paper, and scrub them up to brighten them
and then turn the paper over and use the slip rings to provide the profile to give the brushes a bit of scrubbing too.
very typical of slip rings that have been sitting around not used to get a bit oxidized and this makes for a great insulator or in your case a pretty good insulator
also check what the voltage is coming off your rectifiers, going to the field coils.
can you measure the voltage at "x" and "xx" while running?
also check all connections, everywhere that you can get to.
another thought would be to check the field coil continuity, removed either "x" or "xx" from its terminal and then check the resistance across the field wires that attach to "x" and "xx", you only need to remove one leg.
i have no idea what the resistance should be, but i would think maybe somewhere from a few 10's of ohms, to maybe a hundred? what i would not want to see is an open circuit, which would lean toward either a burned out field coil or a broken lead between coils, or lead going from "x" or "xx" to the coils.
the schematic doesn't show anything between the x connections and the fields, so not much to go wrong there, but an open wire, or bad connection, worse an open field coil.
if it turns out to be an open field coil, don't fret too much, it is much easier to replace a field coil than have to rewind the armature. a field coil a reasonably handy fellow could make with a bit of effort, or remove and take to a rewind shop to have replicated.
with the selenium rectifiers having gone up in smoke, i am concerned that maybe the fields have shorted?
also final thought, (sorry for the disjointed response)
remove both sets of field wires from "X" and "x" and do the continuity check as described
and also check between each field wire to the case ground, to see if there is any short to ground.
if you don't have a short to ground, that is a good sign, if you have a resistance reading from say ~10-100ohms or so, i think you would be ok.
when you flashed the field with 24 volts, did you get a huge arc, or just a small flash or spark?
now i am rambling
what output voltage do you get when you apply 24volts to the "x" and "xx" connections?
is that when you get the 16.35volts output on all three phases?
the fact you report and even 16.35 to all three phases, leads me to think this is a field problem or an excitation voltage problem, and not an armature problem, which is good in my opinion. not that problems are good, but an armature problem can be expensive to resolve.
does the reported 16.35 volts reading come from checking between each pair of the three phases?
can you check between the neutral connection and each of the three phases?
also so that i am clear which animal you have, is it the schematic in fig 3 page 5?
i have a few other thoughts, but lets start with these,
bob g