THOR: SMPS warmup behaviour followup

["Tesla list" <tesla-at-pupman-dot-com>]

Original poster: "Marco Denicolai by way of Terry Fritz <twftesla-at-uswest-dot-net>" <Marco.Denicolai-at-tellabs.fi>

Hello all.

Maybe some of you recall that over a month ago I had a problem with Thor
and its
SMPS. Namely, after 15-20 seconds of running the performance got worse. It was
because the SMPS primary capacitor charging speed got slower.

The supposed cause was long reverse recovery time of the BYM36G diodes used to
rectify the charging pulses.

I did switch to using the UF5004 (faster, 75 ns vs 250 ns). That involved
de-soldering 80 diodes and replacing them with 112 new ones.

Here are the preliminary results:

1. Now the diodes run cool also after 1 minute: the old diodes got "can't
touch"
hot within 15 seconds. Good.
2. The performance is still getting lower after some 15-20 seconds!

I have now identified what I believe is the actual cause: there is some corona
between the SMPS transformer secondaries and the ferrite cores. As the corona
grows together with charging voltage it is a good explanation for the charging
profile bending. I can also "ear" the corona. A week ago I had also a discharge
through the secondary (SMPS, not TC) coil former and the core, resulting in one
of the driver getting damaged.

I am now going to try two remedies:

1. ground the ferrite core: at the moment it is floating. This should
prevent it
from collecting charge (if it does).
2. as the core is double-U shaped, smooth the ferrite sharp corners below the
secondary coil former. This should lower the electric field density in that
area.

Can anybody give some hints about ferrite grinding?

Regards