#analog computing

ANALOG Computing June 1984

Sure were!  One of the simplest examples is a slide rule, which is a mechanical analog computer.  Electronic analog computers were made out of operational amplifier circuits, and they were really good at giving you real-time feedback on complex calculations that digital computers simply weren’t fast enough to handle.

Here’s a big example, designed by George Philbrick, made specifically for M.I.T. in 1958.  It doesn’t have a game, but they call it George.  It’s entirely vacuum tube based.

Then you have this little one from 1959: the Heathkit EC-1, which was sold until 1971.  Again, totally analog, and totally vacuum tube based.   However, this one is smaller and significantly more simplistic.  You use patch cords to tie various stages together to create a program.

Here’s more information on the EC-1, if you’re curious: https://www.nutsvolts.com/magazine/article/May2016_Heathkit_Restoration

The PACE TR-10 was a desktop sized analog computer from 1960, and it featured plug boards which could be pre-wired to contain a program, and swapped with minimal effort.  I got to hold one back in high school.

Everything about an electronic analog computer is expressed as a voltage.  Input voltages and output voltages representing all sorts of information.  I won’t lie, alot of it goes over my head, but there is one really good example I always remember:  say you want to simulate a car’s suspension, perhaps to design an appropriate spring to use.  You can model the aspects of the spring in terms of how it will dampen motion (like the length of the spring, or the stiffness), as well as the speed of the height of the bump it will encounter, and the speed at which the simulated vehicle will hit that bump.  Then you can watch how fast it recovers from the bump in real time on an oscilloscope, and tweak the parameters as you wish to see how it responds.  Stiffer spring?  Lower the gain of a specific op amp.  Bigger bump? Increase the representative input voltage.  I’m probably not painting this in the most accurate way, but the hope you get the idea.  And you can see the results about as fast as you turn the knobs and start the program over again which was really useful. 

A digital computer of the 50s or 60s attempting the same task would have been at it for a significantly longer period of time, and require significantly more sophistication to yield a result, and thus it would be way more expensive.  Plus, you would have to spend more time programming such a simulation, and you sure as hell wouldn’t have gotten a real time computed answer.  If anything, you would have submitted a program as a batch job, and waited to see what the result was a day or so later, after they ran it through a large mainframe  -- assuming it ran correctly in the first place.

Military computers for controlling the trajectories of artillery, or bombs falling from an aircraft back in WW II were mostly mechanical analog computers.  Like the Norden bomb sight..  I seem to recall this video explaining mechanical analog computers (as opposed to the electronic analog computers I’ve shown above): https://www.youtube.com/watch?v=s1i-dnAH9Y4