What is the first computer?

From: John Lawson <jpl15_at_netcom.com>
Date: Sat Jun 20 11:10:39 1998

On Sat, 20 Jun 1998, Ward Donald Griffiths III wrote:

> Allison J Parent wrote:
> >
> > <> Mechanical: man or motor powered, bars cams, wheels
> > <> Electric: Relays, steppers, solonoids and contacts has logic tree.
> > <> Electronic: uses active devices, tubes, transistors, ICs some types of
> > <> diodes and neon filled tubes.
> > <
> > <A good definition, I could agree, althrough I've seen
> > <Electric and Electronic the same, since a relay isn't
> > <diferent from a transistor or a tube for the effekt
> > <(beside the current).
> >
> > The distinction for the last two is significant from a design and speed
> > standpoint. the design process is very different at the detail level.
>
> Yah, with a relay-based system, a one-hertz clock speed would be
> setting a record.
> --
> Ward Griffiths
>

  Weeelllll.... let's do some Research.

  From "The Annals of the Computation Laboratory of Harvard
University", vol XVI, 1948.... one can find a few references to
what we would term 'clock speed' in the descriptions of the various
machines (the IBM ASCC Mark I, the ENIAC, Bell Telephone's Relay
Computing System, the IBM Dahlgren Calculator Mark II).

 The Mark I "...operates at a speed of 200 cycles per minute."

     I make this to be over three cycles per second... and by
cycles is meant *machine* cycles... I would posit an
engineering-educated guess that the actual timing pulses were coming
along a few multiples faster than this, though the description does
not call out the actual master timing freq.

  The ENIAC, of course, was all electronic, but for example it's
clock ran at 100 kHtz, with a duration of 2uSec.

  Bell Labs does not specifically say a clock rate, or any
operational speeds, but the I/O were Teletype tape transmitters and
Model 15 page printers. You do the math. (note: this article is full
of cool pictures and examples of coding instructions and tables. If
there's interest I could maybe do some scanning and uploads to my
webpage.....)


 The Mark II Calculator was faster than the Mark I (and much bigger).
 Some timing references: "There are 100 storage registers in the
entire machine. Any quantity standing in one of these is available
to the computing parts of the machine in 33 milliseconds."

 "There are about 13,500 electro-mechanical relays in the Dahlgren
calculator. The operating times of these relays in either direction
are from six to ten milliseconds."

 "....cam operated contacts, about 800 in number, are used for this
purpose. [the cams sync the machine and carry the heavy currents,
sparing the delicate relays] The impulses...... are of 16 2/3
milliseconds in duration. A large margin of safety is thereby
provided.."


   SO: The first American relay machines operated at speeds
considerably in excess of 1 Htz. I have no data on the Zuse.


 neener neener neener, Ward ;}


Trivia: The first Binary Adder was built by two Bell Labs engineers
in the mid thirties, and called the "K2" because it was assembled
one a Kitchen table in thier free time. I will look up the reference
if anyone's interested.

  I really must scan the pix of the Mark II... they have a fish-eye
shot of the whole machine, and also the various units, and even one
of inside the relay cubicles... tres cool, no? Oui?


Cheers

John
Received on Sat Jun 20 1998 - 11:10:39 BST