I am still betting 10:1 on one of those diodes breaking down. The diodes you have in there if I remember correctly are 200 PIV 12 amp diodes. If I had a few lying around here I could find and I thought they would work I would mail them to you.
Something I mentioned earlier was measuring the voltage drop across the diodes one by one with it running good then measure again across the diodes when it fails.
You lose 50% of your field and each diode controls 50% of the field on alternating half cycles from your excitation windings. Now I figure either you have a diode breaking down or one of the field excitation windings breaking down. You could measure the AC voltage of the excitation windings with both diodes disconnected, Just measure across the Yellow wires measuring AC Volts. If that AC Voltage remains constant you have a diode breaking down. Not going open like a disconnect but breaking down. Now if the AC from the excitation windings fails and drops to half you could measure from each yellow wire to ground and isolate which excitation coil is breaking down and failing.
That's my story and I am sticking to it.
Ron
Something I mentioned earlier was measuring the voltage drop across the diodes one by one with it running good then measure again across the diodes when it fails.
You lose 50% of your field and each diode controls 50% of the field on alternating half cycles from your excitation windings. Now I figure either you have a diode breaking down or one of the field excitation windings breaking down. You could measure the AC voltage of the excitation windings with both diodes disconnected, Just measure across the Yellow wires measuring AC Volts. If that AC Voltage remains constant you have a diode breaking down. Not going open like a disconnect but breaking down. Now if the AC from the excitation windings fails and drops to half you could measure from each yellow wire to ground and isolate which excitation coil is breaking down and failing.
That's my story and I am sticking to it.
Ron