"Basics of Analog Computers" book

From: ajp166 <ajp166_at_bellatlantic.net>
Date: Wed Nov 8 18:32:31 2000

From: Don Maslin <donm_at_cts.com>
>particular, I recall reading about a '60ish project by one of Britain's
>Formula One makers to use a computer to control a F1-machine driving
>around an actual circuit in lieu of the driver. As I recall the story,
>it did not make it beyond the second turn. However, an analogue
>computer handled the task quite well. Apocryphal? ???


this confuses the ideas of precision, accuracy and speed with
complexity.

Digital is absolute accuracy IE: 0101=5 and for a given machine any
variation
is error.

Analog, is mushy on accuracy, 5 Volts at some output is "approximate".

Digital to process data for a complex equation in real time has to
collect
it perform operations on it and output it. These occur at some rate that
is
fixed by hardware. The more complex the task the lower the total
Throughput will be. Also there is latentcy, time from input to processed
output, more complex means longer for a fixed speed.

Analog, assuming a fixed slew rate for the op-amps used the system
throughput IE: a change in input means near immediate change in
output that will be complete in some "setteling time". That time even
for
slow op-amps is quite fast and usually in microseconds.

Also analog does not mean the output cant be absolute (IF Vin or Fin
exceeds N (some rate) then turn relay on).

I've done some really neat things using pots realys and motors with
only limited use of amplifying devices to trans late the pots output
to something that can drive a motor. This is classic servo systems
stuff and the fundemental circuit is the Wheatstone bridge for systems
like this. If you add nonlinear devices like LOG taper pots(variable
resistors) and gears and cams you can get reactions that can mimic
things like non linear actions. The idea of balance is employed.
This kind of automation was as common as flies and widely
seen though often as simple devices.

There is another group of machines that fall dead smack in the middle
of the analog and the digital. These were the early preprocessing
boxes that might take 8 channles of analog telemetry data and
reduce it to stored data based on time, sensor and relative value
or even as a histogram. While these machines would have counters,
registers and some even core or drum(disk) they were at best not
programable and most didn't even qualify as hardwired program
(elementary state machine).

Just a few points about early computer differences.

Allison
Received on Wed Nov 08 2000 - 18:32:31 GMT