17 Jun
2006
17 Jun
'06
10:03 p.m.
The odd thing about gas-discharge tubes is that they
work at all by sheer
probability.
If you can exclude all external and internal radiation (e.g. gamma
particles, light), a neon bulb won't fire at a voltage anywhere near its
spec-sheet level. The ionization avalanche needs an energetic particle or
Remidns me of the time I had a randomly flickering power-on neon on one
of the PSUs in my 11/45. As soon as I pulled the CPU out to get to it, it
stopped flickering. I thought 'Aha, a loose connection, shaken up by the
CPU fans when the CPU is in the rack', and started to invetigate. Nothing
wrong. I then discoveredm that even with the CPU in the rack, it would
stop filckering if I shone a torch (flashlight) on it. Hmmm....
photon to get it started. That's why neon bulbs
fire at higher voltages in
a dark room than in a lit one. Manufacturers put a bit of radioactive
I think one of the early computers (EDSAC?) used a matrix of neons and a
photomultipler as one of the ROMs. They had to 'flash' a light on the
matrix jut before each read (and gate out the resulting pulse from the PM
tube) to ensure a neon in the matrix would strike when required.
-tony