Chris F. wrote:
When the b+ becomes to high, the gas-discharge tubes start
oscillating, producing a steady stream of pulses.

So start with lowering b+ to zero, and then increase, until
either back at 90 volts, or the pulses re-appear.
At that point lower the voltage a bit, and re-check with your
radiation source.
Also there should be a capacitor across the tubes somewhere,
If so, check their quality(leakage and value).

Already tried that, and I replaced all the ceramic capacitors in the
circuit. By gas-discharge tube, are you referring to the neon lamp inside?
There's another tube, with just two leads, that I can't identify or test.
See this photo:
http://s282.photobucket.com/albums/k...=detectron.jpg


No, I was referring to a detector tube filled with neon, where
the voltage on the tube is just below the sustained ignition level,
and where the charged particle causes a conducting ion trail, which
discharges the parallel capacitor a little bit, causing the click
signal you can pick off. The feed resistor slowly restores the voltage
accross the tube/cap. The amplitude of the click shows the energy level
of the detected particle. Long ago I built a cosmic ray detector/display
using these tubes.
Name is geiger muller tube I think.