Thursday, November 22, 2012

A Nerd’s Walk Down Memory Lane … or Computer History for the Young-uns


Today is a day to recall blessings, and I have been most blessed with family, friends, health, and wealth. I live in the land of the free and the home of the brave, and I’ve been able to spend my life in pursuit of my goals and interests. Some have described the mid-life crisis as when you first realize you aren’t going to be what you wanted when you grew up. I never had that crisis. From early childhood my goal was to be a scientist. Well, I ended up an engineer, but that is pretty close.

I think I was about eight or ten years old when I first subscribed to Popular Electronics. I would consume each issue and dream of the projects printed on the pages. I would order parts from mail order houses like Lafayette Electronics and Allied (no Radio Shack), and I built several projects out of the magazine such as a battery rejuvenator.

My dad bought me a used Hallicrafters short wave radio, and soon I was deep into amateur radio (WPE7BDC).

Hallicrafters S-85




My first radio was an S-40B. This is my second, an S-85 that I bought used with money earned working at my dad’s store. I paid $100, which seemed like a lot of money … and, in those days, it was. I later bought my first motorcycle for only $200.

The S-85s were made from 1955 - 1959 and provided continuous coverage from 538 Kc to 34 Mc.  (That’s Kilocycles or Megacycles. Today we use the term Hertz or KHz and MHz.) This radio has AM and CW but no Single Side Band. It has the usual band spread dial and a "pitch" control for CW.  The pitch control is a variable Beat Frequency Oscillator.

Like my original S-40B the set has 8 tubes with the exact same tube complement. So Hallicrafters made very little change, if any, to several models of receivers, but packaged them differently to keep up with the times.

It was several years later before I got into computers. My first computer was in 1965. For those of you who started with PCs or Apple computers, this is an ancient ancestor. This was before computers were personal, and I met this ancient digital machine at the Montana School of Mines during my freshman year.

IBM 709




It was the last of IBM's first generation of big scientific vacuum tube computers; this machine was built in 1959. It was the first computer with data channels for I/O. This picture shows the central processing unit, which can be opened up like a book to access the wiring. The complete system consists of seven different units including memory, data channels and power supplies plus a card reader, printer and many tape drives. The entire system originally cost $2.5 million although you could also lease it with complete IBM service support. Large corporations, government labs, and big universities would have one or two for the entire institution. The School of Mines used it to run their entire IT operation, and students weren't allowed to program the computer. We did get to operate it, however.

Tektronix 536 Oscilloscope




The most important instrument used in computer development and maintenance is the oscilloscope. Although Hewlett-Packard made the best general purpose test equipment, their oscilloscopes were not as good as the Tektronix ‘scopes, like this model 536. Note the "plug-in" drawers so you could put different vertical and horizontal control units with the 'scope. I had plans after getting out of the Navy to move to Beaverton, Oregon and work at Tektronix. Ironically, my parents ended up in Hillsboro, Oregon, which is next door to Beaverton, in 1975, and my dad still lives there. We visit frequently and several of his friends work at Tektronix, although it is only a shadow of its former corporate self.


IBM 7094




I met this computer about five years later in the Navy. I had been trained as a precision electronic test equipment calibration technician. (Say that fast three times!) The Navy version of this computer had specially designed, compact tape drives. In order to reduce space to fit onto a ship, the tape drives had very short vacuum columns. The vacuum columns held the magnetic tape and acted as a shock absorber when the tape rapidly started and stopped. The shorter columns didn’t work very well, and the tapes would often break.

At one point the DP department (that’s “Data Processing” … a dated term) had only one alignment tape and it was broken in half. Due to the short length of the remaining tape, they weren’t able to get their oscilloscope to sync before the tape ended. So they came up to the calibration lab and borrowed our much fancier Tektronix 547. It could sync much faster than the 536 they were using. So I got to take some pictures of both their scope and their computer.

The IBM 7094 Model I was a middle member of IBM's second generation of scientific computers, built with discrete transistors. It was upward compatible with the 709, so that the same software could be used for both. This picture shows the card reader, two different data-channel consoles, and the system printer. The back row shows the CPUs, with the memory (blue module without doors) at far right. The memory had a total capacity of 32K 36-bit words—about 1 Mb, total. (The magnetic core memory was manufactured at the IBM plant in Boulder, Colorado.)

We used punch cards for our shop’s “Maintenance Data Collection System” and the computer also kept track of all spare parts inventory on the ship producing 14” wide, “green bar” printouts that stacked up into “books” several inches thick. The perforated printout paper was designed to fold, and you would read it by flipping it over vertically, rather than the horizontal flip of a book. Sort of like the iPhone app “Flipboard.” The printouts were put in large binders, some over twelve inches thick.

IBM 026 Keypunch




Keypunch machines predated computers. Herman Hollerith invented punched card data processing for the 1890 census. From the 20's through the 50's electromechanical machines like this cardpunch and card sorter were the mainstay of business data processing. Using keypunch cards for input into the first digital computers was a natural progression. Programmers would enter their software into the IBM 026 Keypunch using the keyboard...

IBM 082 Sorter




... the IBM 082 Sorter was handy for, among other things, reordering stacks of cards that had been dropped

These are relatively modern pictures taken at the Denver Department of Social Services where my friend William worked. The equipment was out-of-date then, but you would still see it being used in government offices that lacked adequate budgets. Besides … they worked!

IBM 2302 Disk Drive




Here’s an IBM 2302 disk drive for the 7094. The disk platters are 24-in. in diameter and the head assembly is positioned with compressed air. It is one of the last models this size and can store 300 MB. IBM invented the hard disk at their facility in San Jose, California.

Although I was assigned to the Calibration Lab on board the ship, I was a First Class Electronics Technician (E-6), and I had wide responsibility. On duty nights … that’s when everyone else went home, but you stayed to stand watches and fix any critical equipment failures … I would often get involved in our telecommunications systems.

Our ship, the USS Vulcan (AR-5), was a “man-of-war,” and we were in constant communications with the “net.” (That was before the Internet, but this was still a form of digital communications.) We used special Teletype machines that sent and received encrypted (top secret) military messages and orders. The machines were basic teletypes with additional secret sauce to handle the spy stuff.

The teletypes could be connected to radios and used while underway, but, in port, we would just connect to good old AT&T landlines. The telephone company provided a line to the pier. We would connect to a box on the shore and that was our telecomm connection. It was a 60ma current loop interface and there was a mechanical relay in the box on the pier that served as the interface.

If we lost communications during the night, I would troubleshoot the problem and determine if it was our equipment or did we need to call the telephone man to fix the issue. Often the problem was just the electromechanical relay in the box on the pier, and I would usually start by swapping out the relay to see if that would fix it. The relays were packaged in a little metal box with a base like a vacuum tube so you could plug them in and out. This little instrument would test the relays.

Teletype Test Set




This is a test set for polar relays used with teletypes. The polar relays amplified the signal as it came off the wire. The central hole in the middle of the instrument is where the relay probe plugs in and the dials allow the user to feed in different voltages and currents. The user could determine whether the signal was balanced and get a rough idea of speed.

I mentioned the Internet. Its roots were in early network and communications work started in the ‘60s. Simultaneous work on secure packet switching networks took place in 1964 at MIT, the RAND Corporation, and the National Physical Laboratory in Great Britain. 

At the same time, IBM’s new System 360 computers come onto the market and set the de facto worldwide standard of the 8-bit byte, making the 12-bit and 36-bit word machines almost instantly obsolete. The $5 billion investment by IBM into this family of six mutually compatible computers pays off, and within two years orders for the System 360 reach 1,000 per month.

On-line transaction processing debuts with IBM’s SABRE air travel reservation system for American Airlines. SABRE (Semi-Automatic Business Research Environment) links 2,000 terminals in sixty cities via telephone lines.


PDP-8




DEC unveiled the PDP-8 in 1965, the first commercially successful minicomputer. Small enough to sit on a desktop, it sold for $18,000 — one-fifth the cost of a low-end IBM/360 mainframe. The combination of speed, size, and cost enables the establishment of the minicomputer in thousands of manufacturing plants, offices, and scientific laboratories. 

With ARPA funding, Larry Roberts and Thomas Marill create the first wide-area network connection. They connect the TX-2 at MIT to the Q-32 in Santa Monica via a dedicated telephone line with acoustic couplers. (You may have seen these in old movies. You placed a telephone into cups on the coupler and it “talked” and listened via the telephone handset.)

Later the ARPA-funded JOSS (Johnniac Open Shop System) at the RAND Corporation goes on line. The JOSS system permits online computational problem solving at a number of remote electric typewriter consoles. The standard IBM Model 868 electric typewriters are modified with a small box with indicator lights and activating switches. The user input appears in green, and JOSS responds with the output in black.

Built with discrete plastic "flip chip" transistors, this was the first computer for less than $25,000. It was cheap enough to be purchased for individual projects. Its 1.5-µs core memory made it fast enough to be useful on a wide range of applications from scientific research to typesetting. It quickly became a popular node on the still evolving DARPA Internet. Here is a close-up of the front panel lights and switches. This was our computer at Electronics Technical Institute where I taught in the '70s.

IBM System/3




IBM introduced its System/3 computer (IBM 5410) on July 30, 1969 to meet the computing needs of small businesses. It was IBM's first mini computer. It was the first system totally developed in-house by the company's laboratory in Rochester, Minn., and the most significant IBM product announcement since the IBM System/360 in 1964. The System/3 was not compatible with the System/360, as it featured a smaller punched card that could encode up to 96 characters per card. That's right ... it had "state-of-the-art" PUNCHCARDS! The System/3 used IBM's new monolithic integrated circuits, and rented for less than $1,000 a month -- about half the cost of a System/360 Model 20.

On October 28, 1970, the company rolled out the IBM System/3 Model 6 (IBM 5406). Rochester's Advanced Unit Record Systems Programming group had developed the Report Programming Generator II programming language intended for commercial applications on the Model 6. RPG was a simple language and a forerunner of VisiCalc ... at least in function.

On July 10, 1973, IBM introduced a new, larger-capacity System/3 -- the Model 15 (IBM 5415) -- with added function and versatility. Developed in Rochester, the Model 15 was manufactured in Boca Raton, Fla., and in Vimercate, Italy. I visited Italy several times as the site was also a manufacturer of the Series III Copier. I also spent several months at the plant in Rochester, MN, too; home of the Mayo Clinic. A beautiful little city in the summer ... a bit cold in the winter ... but some great restaurants and very friendly natives.

By July 1974, more than 25,000 System/3s had been installed around the world, and another version, Model 8, jointly designed by Rochester and Boca Raton, debuted that September.

The IBM System/38 ultimately succeeded the System/3 in 1978, and all System/3 models had been withdrawn from marketing by June 1985. Today the System/3 legacy is carried on by the IBM AS/400 family sharing hardware with the RS/6000 line and produced in Austin, TX. Many a small office or company still uses this versatile computer for day-to-day "data processing" needs.

During this hey-day of the System/3, I was working as a test engineer, first on IBM’s Series III (name coincidental) Copier and then on the ESTAR – Eight Station Test and Repair facility at IBM’s US production facility for Diskette Drives, Boulder, Colorado. I was responsible for testing all diskette drives produced by IBM. I developed and maintained the ESTARs to perform that function. A Unified Test System computer based on the System/3 ran the ESTAR. We manufactured six different models of eight-inch floppy drives, as well as 5.25 inch and a special 4 inch drive that we developed, but it was never a success, foreclosed by the 3.5 inch drive developed by SONY.

During that time I wrote a lot of code in both PL/I and Assembler, and I also earned several patents for data formatting algorithms on disks. It was during this time that I wrote code that was released with the new IBM PC. I wrote parts of the diskette access driver for DOS 1.1 … very nerdish. I also did extensive data analysis of diskette wear characteristics. All of that work was done on this 3741 Data Station, although I also used 3276 and 3278 terminals connected to an IBM mainframe for my mathematical and design algorithm work. Some of the circuit board layouts were designed on Tektronix terminals that had display storage CRTs originally developed for oscilloscopes.
 

IBM 3741 Data Station




I did all the programming using this simple data entry station that was part of the ESTAR, the 3741. The introduction by IBM of the Data Station, in 1973, made it clear that the end of the punched card era was coming. The Data Station was designed to eliminate keypunch equipment like the IBM Model 129, using the newly invented diskette (floppy disk) as storage medium. The original IBM floppy disk was intended to restore the operating system on large computers, but it was quickly recognized as a good, inexpensive, and very portable data storage medium.

(That's an IBM Model 129 pictured at the start of this essay.)

Indeed, the "Data Station" was the first IBM product to use read-write diskettes. The original floppies were read-only. These single-sided diskettes were 8 inches in diameter, with 77 tracks, hard sectoring (8 sector holes), and room for 240 kB (equivalent to about 3000 punched cards). We quickly added a second side and then doubled the capacity of the drives, ultimately producing a juke-box like drive that held thirteen individual disks that could be loaded under computer control.

Other novelties of the Data Station were that you could see what you were doing by looking at a CRT screen, visible in the opening at the left through a mirror, and that you could make corrections. The little, green, eight-line by forty-column display is that little box to the left of the keyboard. I would put an eight-inch floppy into the system and write code that was then cross-compiled on an IBM Series/1 computer in a lab. 


I later added bar code readers and other improvements to the data input of the ESTAR, eliminating manual data entry and reducing errors. I won a $10,000 IBM prize for that work which saved IBM ten times that amount in less than a year.

It was shortly after that that I got my first IBM PC and … as they are wont to say … the rest is history.


5 comments:

  1. Hi Mickey, thanks for sharing your memories, I was googling for a weird IBM Data Entry Station we had at IN.CO.RES. in Rome in 1985 (a programming teaching company that's disappeared in the nineties) for which I can't remember the model number.

    May be you've had to do with something similar, looking at your photo of the 3741 Data Station, it looks very very similar, CRT display of 8 or 10 rows, 8" floppy drive and the best (mechanical) keyboard I've had the chance to use in the last 30 years.
    But unlike the 3741, the one we used to type COBOL was in fact a pair of units sharing the same desk. Two operators sat on the opposite sides of the desk, each with its own keyboard and floppy drive, but sharing (very clever) the same CRT screen which was placed horizontally and tilted by 45 degrees at the center of the table. The CRT unit had two mirrors mounted above it allowing each operator reading it's half screen display anc corrected the angle of view. In this way a 40 rows display appeared to each user as a 20 rows display!
    I imagine there was some magic in the video controller that flipped the upper half screen upside down, allowing the person in front of you reading his text normally. At the time this trick really hit me.
    This is one (among hundreds) example of how expensive those machines were, and how a technical trick could effectively cut costs both for the manufacturer and the end user as well, allowing more people to work with them.
    There at IN.CO.RES. we had two or three of these units, and that was the only place where I could use an 8" floppy disk, the 51/4" had been introduced some years before.
    I had much fun (and frustration) typing my first COBOL programs on it :) but the ergonomics in that keyboard (as in other IBM products) were really the finest I've found.

    ReplyDelete
  2. "Unknown," Sorry, I don't recognize the dual workstation you described. I worked at the Boulder IBM and our focus was on copiers and disk drives, so I didn't see all the various IBM mainframe workstations that were available. I just worked with a few products which I described in this note.

    Sorry I couldn't help

    ReplyDelete
  3. Hi.
    For "Unknow" like Ypu I search for this machine and only found this:
    http://www.computerhistory.org/collections/catalog/102622180

    Hope Usefull for You.

    Regards

    ReplyDelete
  4. Yeah I think mao is right, the 3742 was like the 5252 terminal - one crt split into two via two mirrors - real mickey mouse!

    ReplyDelete