[Heath] IO-12 Scope poor sweep linearity

[Charles Morris]

I helped my dad assemble my IO-12 in the late 60's (the only reason I
still have it around...) and recently decided to resurrect it.
Many tubes were making poor contact so I removed and reinserted them a
few times and got a trace. I also replaced some of the ancient
electrolytics just-in-case.

Anyhow, we were always puzzled by the distorted looking sinewave (from
the scope's filament winding, routed to the front panel 1V PP
terminal) - it looks reasonable when the scope was first turned on,
but after a minute or so, as the scope warms up, it begins to shrink
and distort significantly. We never did have another scope around to
check it out with... today I put a Tek 453 on the jack and discovered
that IS what the sine wave looks like! The power transformer must be
on the verge of core saturation, or it's unable to handle the dc
component in the windings, or something similar. Any thoughts?

More importantly, the sweep linearity is, in a word, lousy.
Examination of the overly-simple sweep generator circuit, and
measurement at the horizontal amplifier input on a good scope, shows
it's not even close to a true sawtooth. The sweep is generated by a
simple RC circuit discharging, which of course gives an exponential
decay, not a linear one! An RC decay is only a reasonable
approximation over a very small portion of the curve, which this one
is not (approximately 25 volts on a power supply of about 105 volts).

Just wondering if anyone has come up with an easy-to-implement
constant-current source, without redesigning the entire timebase
section... or just live with it as the inexpensive "vintage"
instrument it is ;)

-Charles