Thursday, March 26, 2015

Recalling The Days Of Analog Computers

1964 Ad Flashback: When Analog Computers Walked the Earth
When I went through engineering school we had both analog computers and the huge IBM mainframe. We programmed both, the mainframe with decks of cards and results back in 10 days, and analog computers which graphed out results on an xy recorder while you waited.

Analog computers were actually designed real-time by the user, with a different design every time a new set of equations was run. That was because they were a series of operational amplifiers which had feedback loops that determined their mathematical output function and these were hooked together in a manner for simulation of the solution of the equations for a specific input. For a differential equation (not really available on the mainframe), you could integrate and differentiate depending on where you inserted the capacitor. It was very cool and it was very techie for the times.


Robert Coble said...

Fond memories indeed!

I went through USAF technical school on flight simulators in 1967 - all analog, using vacuum tubes and servo motors with potentiometer stacks for calculating everything in the simulation.

In 1970, I wangled a transfer to Castle AFB in Atwater, CA because they were scheduled to get fully digital simulators. I got an additional 6 weeks of school on digital electronic circuitry. Imagine my disappointment when it turned out that everything in the new simulator was actually a hybrid digital/analog, using op amps and resistor/capacitor networks to control the simulation. Fortunately, the radio navigation aids were all digital, but were not programmable. BUMMER! I did redesign the simulation from a twin-jet utility aircraft to a B-52, so I did put my time to good use.

I graduated with an A.A. programming degree in 1977 and left the USAF. I went to work as a civilian for the US Marine Corps. My first job was on a CH-46F helicopter simulator - and it had a REAL digital computer controlling the simulation. It was a Harris super minicomputer, 24-bit words and had a full 8K (that's right: 8,192) words of real CORE memory! The simulation was written completely in assembler code.

The technicians who were there before me used to write very small machine code diagnostic routines, entered by hand through a series of paddle switches on the front console, with a "dump" program added to unload their diagnostics to PAPER TAPE, punched on an ASR-33 Teletype console, with a small boot loader to load them back in. I decided that I was NOT cut out to be a paddle-switch flipping machine code diagnostic programmer. I found the Harris resident operating system (ROS) tape and the macro assembler tape. I had to use a spare 8K core board in order to have enough memory to run the ROS operating system. I taught myself the assembler language and all the OS hooks; I had only used higher-level languages in college. We had all the source code listings for the operating system and the assembler, so I was able to figure it out without any additional training.

The only problem: there was no source code editor available! So, my first program was to design and write a small in-core text editor. Once I got it running by hand, I rewrote it in assembler, and assembled my first program. The macro assembler required four passes: two passes to assemble the object code into a format that could be linked to the operating system and loaded, and then two more passes to get a printout of the source program. The entire editor source code took up an entire 9" spool of paper tape.

Armed with my new tools, I was able to write some really excellent diagnostics for the simulator, much faster than anyone could thumb in a small program. I also made several engineering changes to the simulation while I was at it.

The Navy engineering group that was responsible for all Navy and Marine Corps engineering changes to the simulators got hold of my changes and diagnostics, and offered me a job. I moved into the "real" engineering world, continued my engineering education and eventually became a software engineer on very large systems.

BUT, I still recall what excellent education and experience I got by "coming up through the ranks" from analog to digital.

Phoenix said...

I was hoping someday (not too distant future) to receive a degree in physics or in mathematical science.But it seems engineers get to have all the fun while scientists' work are mostly theoretical and conceptual.

Robert Coble said...

@ Phoenix:

Not necessarily.

I worked in engineering with several "engineers" with degrees in physics. They were quite as skilled as those with degrees in engineering.

I also noted with interest while doing Y2K remediation for Wall Street that some of the best software developers had PhDs in physics. It seemed that they were in high demand because of the complexities involved in market forecasting.

Rikalonius said...

Robert, while I was born in 1969 and didn't arrive in Merced until the 4th grade, i lived there from 1979 to 1988 when I left for the Army.

I'm very familiar with the former Castle AFB and those simulators. My best friend in school's father was a B-52 pilot and trainer at Castle and I've been in those simulators. I used to stand on Fox road as a teenager and photograph bombers and the occasional fighter while on approach.