Sounds to me like he is doing the right thing, and has a supported model, any ideas about potential windows related fail here? I'd like to go in order of most likely and frankly, windows failing always seems most likely to a linux user :-)
Well, there is always more stuff to go wrong in that environment, it seems, but the clone code is the same for all the VX radios and people obviously use that in Windows all the time.
Having tried two cables, including one "real" serial port cable is what I would have tried, but Ken already did that. Trying another machine (as he said he would) is a good start -- windows users report back with things working on another machine for no apparent reason, but I never hear what changed.
Ken, can you grab the debug.log file that chirp generates *after* attempting (and failing) to do a clone? That might have something useful in it.