The leakage is not an issue of moisture migrating into the capacitor
'jellyroll' slug directly, but once there is some moisture in the slug
in combination with the sulphites in the kraft paper, they promote
aluminum ions of the foil in the slug to begin to migrate into the
paper dielectric. These ions eventually cause leakage paths to form
through the paper, or worse, cause a moderately high conductive path
that, with a high bias voltage, will cause it to carbonize and cause a
high leakage path or full short. If the cap moves to a dry location the
moisture theoretically will migrate out again, but the ionic migration
does not reverse--just slows down.
The "tropical" wax that I use in paper cap rebuilding may be somewhat
better than the 'filled' potting and slug waxes of consumer grade paper
caps, but it's main difference is the anti-fungicides it contains to
delay the formation of molds and mill-dew on the organic components.
Neil S.
n said:
Proudly stating "British Made"
What makes all those yellow tubular wax covered caps, Dubilier type 460, TCC
type 343, 645 etc 300V to 750V caps in range 1 to 100nF go so leaky ? Is it
the wax is actually hygroscopic and absorbs water vapour over time. ?
On DVM resistance scale in Meg ohms but on a Megger then as low as 50Kohm.
Strange thing this mold/midew stuff, this is one of the strangest fault
causes I've ever had
Scopex 14D10 scope
Nasty noise from the ps and nothing else.
Using an isolation transformer and observing on a scope
there was low level oscillation of the smps and one burst of
oscillator per cycle of the mains.This smps is the type with
the oscillator on the mains side of the pulse transformer.
All active components seemed ok and no leaky caps.
Putting a dc supply across the oscillator and varying
the voltage the drive to the main trannie changed state
at about 20V.Disconnecting the pulse transformer and
testing with a megger (high v insulayion tester) there
was no interwinding breakdown and the inductance of
the coils looked right (no shorted turns).Eventually found
the O/P was being loaded by a faulty opto-isolator that
gates the beam current.It was ohmic between I/P
and O/P so removed and cracked open with mole grips
(vice grip locking pliers).Looking at the transparent
bridge under a 30x microscope there were tiny circles of
mold or fungus that had grown and coalesced forming
a resistive bridge between I/P and O/P.