I know, bad form to follow up to myself....
[...'scope which doesn't seem to have the beam
on-screen...]
...I turned out to be in luck.
On closer examination, this 'scope has a bank of screw terminals which
are labeled (in black-on-dark-brown, not obvious except on close
inspection) as being suitable for jumpering in certain ways to get the
X and Y deflection plates as coming direct from inputs or coming from
the X and Y deflection amplifier circuitry. I removed the insulator
panel on which these terminals are mounted, and it is clear that four
of the terminals go directly to the deflection plates (two for each
axis), with a resistor (4.7M) hardwired from there to the
deflection-amplifier terminals. (The direct-input terminals are
capacitor-coupled to the plates if the jumpers are in the direct-input
positions; if the jumpers are in the amplifier-fed positions, they
short out these resistors, with the direct-input terminals doing
nothing at all.)
So I removed the jumpers entirely and, using alligator clip leads,
grounded both X deflector plates (to the chassis, that being where all
other grounds go). Voila, I can find the trace! (Well, spot, under
the circumstances.) Using a voltmeter on the amplifier-output
terminals indicates that one is driving about +100V and the other about
+400V (of course, these may be reading low because of meter impedance,
but they clearly are nowhere near equal). Grounding the deflection
plate corresponding to the +400V line and leaving the other alone gives
me a spot near one edge of the display; the other way around and I get
the symptom I wrote about.
So now it's just a matter of tracing circuitry back until I find the
problem. (Heh. "Just".) Sorry to bother everyone when the answer was
right there waiting to be discovered all the time.
/~\ The ASCII der Mouse
\ / Ribbon Campaign
X Against HTML mouse at rodents.montreal.qc.ca
/ \ Email! 7D C8 61 52 5D E7 2D 39 4E F1 31 3E E8 B3 27 4B