Jeff Liebermann wrote:

it's IT-3121, connected to an HP-1801A/180AR as display.

Schematic and operating instructions at:
http://www.vintage-radio.info/download.php?id=225

On Pg 62 of the operations manual, it shows various display anomalies.
Some are loops which look something like your photo. They suggest
using the LOOP control to reduce the effect or lower the power supply
voltage. You've probably already tried this, but I thought I would
mention it anyway.

Yes, I've tried the loop control but it doesn't change much of anything on
that display.


I've tried going from 5K to 50K and oscillation remains, of course peak
current varies. Using 100K as series resistor doesn't make the tube ignite
at all. Max voltage value on this curve tracer is around 240V.

You might be too close to the conduction threshold on the tube. Hard
to tell without specifications. If the device appears to fail to fire
at high series resistances, it might have fired, dropped the voltage
to about 15v across the tube, and stayed there. Grab a DVM and
measure the voltage across the device. My guess(tm) is that it's
about 15v and holding.

the test voltage is a half waveform of the 50 Hz (in this
part of the world) coming from a secondary winding and some transistor
buffers. The voltage waveform is sent to the test probe and the X axis
of the oscilloscope. So if the trace on the scope goes all the way to 240V,
I think it's not firing. With 100K series resistor I can't see any
vertical (current axis) movement even at 1mA division.


For fun, put a capacitor (300v or high) across the tube with a high
series resistance and see if you can get it to oscillate as a
relaxation oscillator.

another thing to test, ok.


Frank