[vcf-midatlantic] vacuum tube computers

Dan Roganti ragooman at gmail.com
Mon Dec 4 14:26:48 EST 2017

On Mon, Dec 4, 2017 at 1:41 PM, william degnan via vcf-midatlantic <
vcf-midatlantic at lists.vintagecomputerfederation.org> wrote:

> >
> > <snip>
> > >
> > > PS: a vacuum tube, controls electrons going from a hot metal to a metal
> > > plate. The flow of electrons is a current, like water from a faucet.
> The
> > > current is controlled by the electrons going through a grid. a grid is
> > > wires strung across the current-flow, which repel the electrons like
> > > magnets do. Just as you can control water by turning a faucet, you can
> > > control electric currents with a vacuum tube, by changing voltages on
> > > its grid. Old radios and TV's, very old computers, used vacuum tubes.
> > >
> > > --
> >
> >
> One notable exception - Many of the best new production guitar and bass
> amps still use tubes.      It's not a nostalgia thing, they often sound
> better and there is a market for the "tube sound" because solid state amps
> can't seem to reproduce it.  I use both tube and solid state amplifiers.
> Many have tubes as pre-amps, there are also guitar pedals with tubes in
> them, same purpose.  They warm up the sound before the post-gain power
> amplifier pushes the sound signal to the speakers.
> Bill

The big reason that tube sound still has a strong following in the music
industry has to do with the characteristics of the components.​ The vacuum
tube has a natural characteristic in the audio spectrum to produce even
harmonics which is soothing to the human ear. While solid-state amplifiers
have a natural characteristics to produce odd harmonics - which isn't very
pleasing. This is before you run any guitar through some sort of effects

That feature doesn't apply to analog computers. But it does apply to
digital vacuum tube computers. The main reason CPU speeds couldn't reach
beyond 500khz or more is because they were still using linear vacuum tubes
in the beginning. Even though they had a large frequency bandwidth, this
was a linear measurement [ie. sinusoidal waves as in a transmitter - not
square waves as in a computer]. In general, square waves in a computer have
odd harmonics, and so you needed even a larger bandwidth for vacuum tubes.
This is because the circuit needs enough energy to output the main
frequency plus the odd harmonics without the signal getting distorted and
not resemble a square wave anymore. Another component characteristic which
impeded this bandwidth is called the Miller Effect, excessive internal
capacitance of the component limited it's bandwidth.

Which is why they innovated a new form of vacuum tube in the mid-50's
called computer-grade tubes which contained a larger frequency bandwidth
for square waves - ie higher switching speeds. This would have pushed the
envelope so to speak on next generation vacuum tube computers. I saved old
vacuum tube computing books with examples running at 1mhz and 2mhz. But
then look what else happened back then. Transistors came into our world -
much smaller and less power - not much faster at first. But it didn't take
long before they were running a 1mhz and more in the digital realm - the
same reasons from before applied regarding the linear bandwidth.

More information about the vcf-midatlantic mailing list