local demolition of SAGE building

[William Donzelli]

> But inter-element shorts and leakage become more common, and my
> experience (backed up by many years of leaky/shorted tubes) is that
> these are related to power-on cycles.  (i.e. "it worked last time I
> used it, but then I turned it on and all it does is hum").
> 
> Maybe reduced sizes make physical tolerances between the elements more
> critical in the smaller tubes.  Even though a lot of the leakage
> I've seen is between adjacent pins, not between adjacent elements.

Leakage and shorts went down dramatically in the 1950s, when the tube
makers started making the tubes in (mostly) clean room environments, with
better automation. Subminis, the tubes that will-not-die, were made in
real clean rooms.

I do not know of any non-military computer system that used subminis. 

William Donzelli
aw288 at osfn.org