Positively the rotor, been there several times myself. Some test fine but open up or short at speed from centrifugal force, not all of the wire core is solidly glued on some of them. That means the wires can move around to abrade. I personally got stranded 200 miles from home one evening by one that ran slowly down until I was pushstarting it, then I slowed down from 70 mph to 45 and by the time I was home the battery had come back up enough to start bike again. Go back up faster riding local and the battery then runs down again. After checking all and no issues found I changed rotor alone and problem gone. As well on the DOHC sites others have reported it. I think it's worse on rotor rewinds as many of them are epoxy coated at the center and on the outside only and again the inner wire loops move around if not wound tight enough.
Here in Texas pretty much lower corrosion issues, I've not had any other DOHC alt trouble other than needing to change brushes from wear. The only issue I've had is with the rotors and why I go to them first every time. I changed maybe 5 of them over the years. I always check at the slip rings only and clean too. I had some stators cook pretty good to lose some epoxy but I simply patched them up with new glue and ran them some more, I rebuild all my own car alts as well. A bit of trivia here, the Ford PWM controlled alts are well-known for testing and retesting fine all day long and then still being bad. I tested plenty of them when I was in parts there for a bit. On those a similar issue to these Honda but a different cause, the diode leads stick way up to be soldered into a tinfoil thin gathering connector for the A/C, the tall leads and no support lets the heat combine with vibration to crack the lead to then touch/not touch in use and it drives people crazy when the regulators melt from pushing the field too high trying to get back lost charge. Ford only cares about selling more parts and their thing now and the rebuilders were forced to re-engineer that diode plate to hold the leads in a much more substantial way and problem pretty much gone then. I've rigged a couple of my own instead of buying the $200 alt by simply resoldering the cracked leads (often more than one is broke) to get them back on line for nothing. Years more of free use.
Every design of everything pretty much has a primary failure mode.
I did somewhat mis-speak above, I meant that the rotor was the most common fail, not that the rotor checking within tolerance was........