#classiccomputing

🎄💾🗓️ Day 14: Retrocomputing Advent Calendar - Atari 400/800🎄💾🗓️

The Atari 8-bit computer line was launched in 1979 with the Atari 400 and Atari 800. Both were advanced for home computers at the time. Both were driven by a 1.79 MHz MOS Technology 6502 CPU, with ANTIC and CTIA/GTIA custom chips for advanced graphics and the POKEY chip handling sound and input/output duties. The Atari 800 was the premium model, featuring a full mechanical keyboard, user-expandable to 48 KB of memory, more metal shielding, and more durability. The Atari 400 was a lower-cost alternative with a membrane keyboard, limited to 16 KB of RAM, and a simplified construction.

With hardware-accelerated scrolling, Player/Missile graphics, and rich sound, the Atari 8-bit systems were known for gaming and creative applications. They were a versatile platform with cartridge-based software, cassette, and floppy disk storage. Atari's proprietary SIO (Serial Input/Output) port allowed daisy-chaining peripherals such as printers, modems, and disk drives, making connecting them easier than with other systems.

Newer models were more compact, combined memory expansion to as much as 128KB, and compatibility with developing software and peripherals improved. Atari's 8-bit computers are remembered for having groundbreaking hardware and a very colorful game library, and they are still being used by the retrocomputing community.

While doing research for this, I saw the XE Game System, never saw it before, very 80s for sure!

The Atari XEGS (XE Game System) was launched in 1987. A repackaged 65XE with a removable keyboard, it boots to the 1981 port of Missile Command instead of BASIC if the keyboard is disconnected.

This old computer is comprised of four big boxes, three of which are ever actually used.

The UNIVAC 1219. This is the brains of the system. It controls the operations of every other device. This is what I'm referring to when I'm not gesturing to the UNIVAC 1219 as a whole.

The UNIVAC 1540. This is the DDR, or Digital Data Recorder. It holds, writes, and reads the magnetic tape operators load into the machine.

The Digital to Analog Converter. The UNIVAC 1219 was the first digital computer on most U.S. Navy ships, most of which had analog weapons systems. This hulking mass of steel translated the digital signals from the computer to the analog signals of the weapon systems and vice versa in regards to the radar.

The UNIVAC 1532. The I/O console managed the...you guessed it, input and output of the UNIVAC 1219. You can load and punch paper tape for programs more bite-sized than would be used for magnetic tape.

In addition, we have two teletype machines. You can think of them like typewriters that don't receive human input (except the one that can if we want), but instead output what the computer tells it to.  We have a Teletype Corporation teletype that is optimized for character compatability and a Kleinschmidt teletype that is optimized for speed. Both rely on the I/O console to send and receive data.

The real ingenuity begins with the floppy drive. Duane, who's career revolved around this system, developed a way for a floppy drive to imitate the I/O console. The computer thinks it is reading and writing to a paper tape, when it is in fact reading and writing to a 5.25in floppy inside an ancient CNC machine floppy drive.

And this, dear reader, is where the magic happens. This framework was originally built for interfacing with the 1219 via BIN files over Serial port and was easily changed to support BIN files over floppy. Duane has been working on an off adapting our purple converter box with a raspi to let the 1219 read and write BIN files over Bluetooth.

Make no mistake, you cannot simply SSH into this machine as tons of setup and channel changes must be performed to ready it to receive and send data. That being said, I don't see any other UNIVAC mainframes with Bluetooth [or any other running UNIVAC 1219s at all :(], so I will take what I can get.

Also, I was wondering if you have any fun stories surrounding it! Strange quirks it has or anything like that.

I'd love to see more photos if you're allowed to post them!

Thanks for the question! These are my favorite part about my blog by far.

Not exactly, the UNIVAC 1219 doesn’t have a nickname. I did realize recently that I should specify the pronunciation (Twelve-Nineteen), but it doesn’t have any nicknames. Apart from ‘the 1219’, it’s also regularly referred to as the CPU or just ‘the computer’.

Fun stories or weird quirks? Boy, I could fill a book with this machine’s weird quirks (or as we say, intermittent issues), but I’ll try to blitz through the most common ones:

Sometimes the computer will stop running and enter a WAIT mode. No reason, it just needs a break. We can’t fix it, it just has to decide to go back into operating mode.

The computer will often start attempting to communicate on IO channel 13. We’re not telling it to talk to anything, it just decides to try to.

One of our teletypes (the Kleinshmidt) stamps ink splotches into the paper rather than characters most of the time. However, this weekend it worked for the first time in 10 months! We didn’t change anything, it just had an extra cup of coffee or something.

The Digital Data Recorder, or the tape drive, has the most gremlins out of any of our units. The top handler works fairly well, but the bottom handler won’t properly read data, write data, move the tape forward, initialize the tape, or any number of other issues.

There’s more but hopefully this satisfies your curiosity.

Fun stories? Well, I can’t name any specific ones, but I can say it’s a very endearing machine. It’s the very last of its kind and being one of three individuals in the world responsible for it makes every issue that more frustrating. There is no real forum for it, the subject matter experts sit next to me and are often just as exasperated as I am.

But the unique nature of this situation make every successful diagnostic test that much sweeter. Every new addition (5.25” floppy drive via serial) that much cooler. I have an IBM PC-XT clone at home, but I thank my lucky stars every day that this big iron is what I get to specialize in.

As for more photos, I have none that are as grandiose as you would probably expect. I do have my working photos though. I took all my photos when I first started working on it and now I am more dedicated to fixes than photo-ops.

This is a photo of our finicky Kleinshmidt teletype. Still has blotches but it actually printed!

This is the back of the bottom handler. Pictured is the vacuum pump in the bottom left (so sudden stops just yank magnetic tape slack rather than ripping tape). The big cylinder in the center is a motor for running the magnetic tape handler itself. The big black ‘hose’ of wires coming out of the steel plate contains all the cables that come right off the handler’s head for reading and writing data!

This is the forward pinch roller of the bottom handler. It was replaced after this photo was taken as you can see the rubber has deteriorated in the 55 years this machine has been operating.

As for being allowed to post photos, that’s not an issue. The last 1219 was decommissioned in 2014 and now you can find all of its documentation online at http://www.bitsavers.org/pdf/univac/military/1219/

🎄💾🗓️ Day 11: Retrocomputing Advent Calendar - The SEL 840A🎄💾🗓️

Systems Engineering Laboratories (SEL) introduced the SEL 840A in 1965. This is a deep cut folks, buckle in. It was designed as a high-performance, 24-bit general-purpose digital computer, particularly well-suited for scientific and industrial real-time applications.

Notable for using silicon monolithic integrated circuits and a modular architecture. Supported advanced computation with features like concurrent floating-point arithmetic via an optional Extended Arithmetic Unit (EAU), which allowed independent arithmetic processing in single or double precision. With a core memory cycle time of 1.75 microseconds and a capacity of up to 32,768 directly addressable words, the SEL 840A had impressive computational speed and versatility for its time.

Its instruction set covered arithmetic operations, branching, and program control. The computer had fairly robust I/O capabilities, supporting up to 128 input/output units and optional block transfer control for high-speed data movement. SEL 840A had real-time applications, such as data acquisition, industrial automation, and control systems, with features like multi-level priority interrupts and a real-time clock with millisecond resolution.

Software support included a FORTRAN IV compiler, mnemonic assembler, and a library of scientific subroutines, making it accessible for scientific and engineering use. The operator’s console provided immediate access to registers, control functions, and user interaction! Designed to be maintained, its modular design had serviceability you do often not see today, with swing-out circuit pages and accessible test points.

And here's a personal… personal computer history from Adafruit team member, Dan…

== The first computer I used was an SEL-840A, PDF:

I learned Fortran on it in eight grade, in 1970. It was at Oak Ridge National Laboratory, where my parents worked, and was used to take data from cyclotron experiments and perform calculations. I later patched the Fortran compiler on it to take single-quoted strings, like 'HELLO', in Fortran FORMAT statements, instead of having to use Hollerith counts, like 5HHELLO.

In 1971-1972, in high school, I used a PDP-10 (model KA10) timesharing system, run by BOCES LIRICS on Long Island, NY, while we were there for one year on an exchange.

This is the front panel of the actual computer I used. I worked at the computer center in the summer. I know the fellow in the picture: he was an older high school student at the time.

The first "personal" computers I used were Xerox Alto, Xerox Dorado, Xerox Dandelion (Xerox Star 8010), Apple Lisa, and Apple Mac, and an original IBM PC. Later I used DEC VAXstations.

Adafruit FruitJam is a Mac 384K 🍏💾

The FruitJam board has an RP2350, dual USB ports over PIO, DVI output from HSTX, and the ability to emulate some classic computers. In this demo, we're using it with the Mac 128K-But-With-384K-Instead

It's not quite a 512K or Mac Plus—which had 1 MB of RAM—but it does have the same 68000 core and is running a Mac Plus ROM, so once we get PSRAM going, we could easily expand to much more RAM. Right now, we can run Finder 5.1 and test out the mouse and keyboard, which both work flawlessly. This is super fun—we can't wait to get some of our favorite old games working! FruitJam is still a work in progress; we have to revise the PCB to fix some errors, but you can sign up here to be notified when we get them in stock

Oh my goodness, my first question! Well, in a sentence, the UNIVAC 1219 was used to point and fire missiles at enemy aircraft. Now, we can break this down:

The UNIVAC 1219 is a militarized version of the UNIVAC 418. This thing has better resistance to salty air and other things you’d want a seafaring computer going into battle to have, but at their hearts they’re the same. There were three different missile systems the UNIVAC was a part of: Talos, Terrier, and Tartar. These three systems were among the first sea-to-air missile systems fielded by the U.S. Navy.

The UNIVAC 1219 would receive data from the AN/SPG-55 radar and analyze the data it received. Running one of dozens of programs lost to time (or on a single last-of-its-kind magnetic tape we are worried about damaging), it will pass on the results of its calculations to the missile system itself. The missile will go where it is told, and hit an enemy plane.

But our UNIVAC 1219 is unique! The specific machines here are from the John Hopkins Applied Physics Laboratory. At John Hopkins, they developed the fire control programs ran on Navy ships! They had a radar and everything, and I only wish we had a radar and missile battery too ;D.

As for what we do for demonstrations, we have assortments of cool programs we run, and I encourage you to swing by and see for yourself. Our oldest party trick is printing ASCII art of Spock and George Washington, but maybe if this post performs well I will divulge some other cool things we do ;).

With the UNIVAC 1219 being 62 years old now, you would be forgiven for thinking we spend our days figuring our how to get old programs to run and what they did.

Oh no. We make our own programs. We make A LOT of our own programs!

Duane, the person whose career revolved around the 1219 pumps out a new or updated program about every two months for us to try on the UNIVAC 1219.

See, while our tapes for diagnosing issues may date back to when the machine still served, we often find ourselves dissatisfied with them. We are realistically limited to about 100,000 addresses in the machine. The first 537 are reserved, then there are random gaps up to 70k where our utility programs end. Sometimes though, old programs don't play nice with OTHER old programs or programs that Duane has written in the past few years. Bill relays this to Duane and Duane gets to work rewriting these conflicting programs.

Duane has developed an entire 1219 EMULATOR to make sure that these things work properly. I can't give you the modern version, but a 2013 version can be found here under 1219files.zip. Duane also owns a tape punch and punches his own paper tape and mails it to Bill. Bill and I then set off to the museum and make sure it works as expected.