Rescuing The Data On A 1960s LGP-21 Computer’s Disk Memory

April 25, 2026 by Maya Posch 33 Comments

One of the nice things about magnetic storage is that as long as the magnetic layer remains intact, the data it contains should stay readable pretty much indefinitely. That raises the prospect of recovering data from really old computer systems featuring magnetic memory, such as the 63-year old LGP-21 that [David Lovett] of Usagi Electric is currently restoring. Its magnetic memory disk is nothing amazing by modern standards, but after initial testing it seems to spin up and read data just fine, raising the question of what was left on the drive when it was last used, meaning what was in memory at the time.

The read/write head side of the LGP-21's magnetic memory. (Credit: Usagi Electric, YouTube) — The read/write head side of the LGP-21’s magnetic memory. (Credit: Usagi Electric, YouTube)

Non-invasive data recovery here involves writing a program that will simply read the entire disk from beginning to end. Tracks 0 and 1 were found to be unreadable due to some kind of hardware issue, but track 2 could be backed up by looking at the output on the CRT, thus providing a track to use. Fascinatingly the LGP-21’s memory disks uses interleaved tracks to reduce the number of read/write heads as part of the overall cost-saving measures relative to the more expensive LGP-30. As you might expect, this slows down memory access a lot over its big brother.

Before any recovery attempt could begin, the Flexowriter typewriter that forms the user interface to the computer had to be given some serious maintenance, along with a few other components like a switch and the paper tape reader. This restored the ability to even properly enter data and receive output instructions.

The subsequent effort to recover the stored data involved a bootstrap program that got loaded into memory, after which the remainder of the program was loaded from paper tape. Following this everything worked swimmingly, though with the caveat that with not even a floppy drive to use, the raw hexadecimal data was hammered out on paper with the Flexowriter over the course of 1.5 hours.

This data will now be scanned in and OCR-ed into something that can hopefully be easily analyzed. Hopefully we’ll know before long what this system was last used for.

Continue reading “Rescuing The Data On A 1960s LGP-21 Computer’s Disk Memory” →

Hackaday Links: April 19, 2026

April 19, 2026 by Tom Nardi 5 Comments

We’ll start things off this week with a story that’s developing more than 25 billion kilometers from Earth — on Friday, NASA announced that the command had been sent to shut down Voyager 1’s Low-energy Charged Particles (LECP) instrument. As the power produced by the spacecraft’s aging radioisotope thermoelectric generator (RTG) continues to dwindle, engineers at the Jet Propulsion Laboratory have been systematically turning off various systems to extend the mission for as long as possible. It’s believed that deactivating LECP should buy them another year, during which engineers hope to implement a more ambitious power-saving routine. If this sounds a bit familiar, you’re probably thinking of Voyager 2. The plug was pulled on its LECP instrument back in March of 2025.

The JPL engineers hope that their new plan may allow them to reactivate previously disabled systems on the twin space probes, but even if everything goes according to plan, there’s no fighting the inevitable. At some point, there simply won’t be enough juice in the RTGs to keep the lights on. Although it’s going to be a sad day when we have to bring you that news, surviving a half-century in space is one hell of a run.

Speaking of ending a run, just a week after Amazon announced that pre-2012 Kindles would no longer be supported, the company is letting users know that the Kindle software for PCs will be discontinued in June. In its current form, at least. As Good e-Reader reports, Amazon is developing a new client for users who want to access the Kindle ecosystem from their computers, but it will only run on Windows 11. Since older software could be used to strip DRM from purchased ebooks, it seems likely this is another attempt to lock the platform down.

Continue reading “Hackaday Links: April 19, 2026” →

What Can You Run On A 1960s Univac? Anything You’re Willing To Wait For!

April 18, 2026 by Tyler August 14 Comments

There are two UNIVAC 1219B computers that have survived since the 1960s and one of them is even operational. [Nathan Farlow] wanted to run a Minecraft server on it, so he did. After a lot of work, of course, which is described in a detailed blog post, and, a YouTube video by [TheScienceElf] we’ve embedded below.

The UNIVAC is a seriously weird architecture by modern standards: it’s got eighteen-bit words — yeah, not even a power of two — and one’s compliment arithmatic with a weird signed zero thing going on. There’s one 36-bit and one 18-bit register, and only 40,960 words of memory. Eighteen-bit words. Yeah, it was the 1960s and they were making it up as they went along.

[Nathan] wasn’t, entirely, as this weird system is both well-documented and already had an emulator — in BASIC, of all things. [TheScienceElf] used the docs and the existing emulator to recreate his own in Rust so he could test their somewhat crazy plan without wasting cycles on real hardware. The plan? Well, there are really only two options if you want to build modern software for a niche architecture: one is to add niche support to something like GCC, and the other is to write a RISC V emulator and compile to that. We’ve seen that second one before, and that’s the route [Nathan] took.

Of course, [Nathan] is a machine learning guy, so he made the best possible use of LLMs — though it’s interesting to see that unlike Z80 Assembly, Claude Code really couldn’t wrap its virtual head around the UNIVAC’s assembly language, and [Nathan] had to bang out the RISC V emulator himself. Emulator in hand, [Nathan] and friends had code to run on the museum UNIVAC. A single frame of an NES game took 40 minutes, but hey, at least it finished before they got back from lunch.

[TheScienceElf]’s YouTube treatment teases hosting Minecraft, but it wasn’t a full server, just the login portion. That they were able to get TCP/IP over serial and set up a handshake between a 2020s laptop and a 1960s computer is still mighty impressive. Just the work the Vintage Computer Federation put in to get and keep this antique running is mighty impressive all on its own, but it’s wonderful they let people play with it.

Continue reading “What Can You Run On A 1960s Univac? Anything You’re Willing To Wait For!” →

The Kentucky Cave Wars, And Going Viral In 1925

April 16, 2026 by Al Williams 10 Comments

Floyd Collins, the unfortunate star of this post. (Public Domain)

Information, it seems, flows at the speed of media. In the old days, information traveled with people on ships or horses, so if, say, a battle was won or lost, it could be months or even years before anyone back home knew what happened. While books and movable type let people store information, they still moved at the speed people moved. Before the telegraph, there were attempts to use things like semaphores to speed the flow of information, but those were generally limited to line-of-sight operations. Carrier pigeons were handy, but don’t really move much faster than people.

The telegraph helped, but people didn’t have telegraph stations in their homes. At least not ordinary people. But radio was different. It didn’t take long for every home to have a radio, and while the means of broadcasting remained in the hands of a few, the message could go everywhere virtually instantly. This meant news could go from one side of the globe to the other in seconds. It also meant rumors, fads, and what we might think of today as memes could, too.

You might think that things “going viral” is a modern problem, but, in reality, media sensations have always been with us. All that changes is the number of them and their speed.

One of the earliest viral media sensations dealt with William Floyd Collins, an unfortunate man who was exploring caves during the Kentucky Cave Wars.

Continue reading “The Kentucky Cave Wars, And Going Viral In 1925” →

Making The Forgotten 1982 Game Adventure Canoe Run On MAME

April 12, 2026 by Maya Posch 5 Comments

A while back [Jack] came across a Taito arcade game that neither he nor any of his mates recognized. The game was Adventure Canoe and part of the collection of forty preinstalled games on a Taito Egret II mini arcade cabinet. Yet despite [Jack] and his buddies being avid 1980s arcade enthusiasts, this 1982 title for the Z80-based Taito SJ system was completely unfamiliar to them.

When even a web search turned up extremely few details, [Jack] did the only reasonable thing and borrowed the rather expensive mini arcade for hopefully some extracting of the game ROM.

As expensive as this mini arcade is, it features the typical ARM-based SoC and Linux-based firmware. Although you can totally dump the Flash, [Jack] found that the firmware update ZIP file was a much easier target to poke at and hopefully extract the ROMs from.

Of course, Taito used password-protected ZIP files within the firmware, leading to some reverse-engineering to find the passwords. The first was ‘hidden’ as plain text in the egret2 binary. For the remainder of the ZIP files the password wasn’t as readily found, but required some sleuthing. This took the form of dynamic runtime analysis with gdb, using information previously gleaned from a Ghidra analysis. Eventually this yielded the final passwords.

Extracting the game’s ROM files this way allowed for them to be adapted to the format that MAME expects, after which the game just had to be added to the emulator’s source files. With this done the game fired right up, and [Jack] was able to play the game without any trouble.

Sega Meganet: Online Gaming In 1990

April 2, 2026 by Zoe Skyforest 20 Comments

It’s easy to think of online console gaming as an invention of the 2000s. Microsoft made waves when Xbox Live dropped in 2002, with Nintendo and Sony scrambling to catch up with their own offerings that were neither as sleek or well-integrated.

However, if you were around a decade earlier, you might have experienced online console gaming much closer to the dawn of the Internet era. As far back as 1990, you could jump online with your Sega Mega Drive. But what did an online console feel like in the dial-up era?

Continue reading “Sega Meganet: Online Gaming In 1990” →

Training A Transformer With 1970s-era Technology

March 29, 2026 by Bryan Cockfield 15 Comments

Although generative language models have found little widespread, profitable adoption outside of putting artists out of work and giving tech companies an easy scapegoat for cutting staff, their their underlying technology remains a fascinating area of study. Stepping back to the more innocent time of the late 2010s, before the cultural backlash, we could examine these models in their early stages. Or, we could see how even older technology processes these types of machine learning algorithms in order to understand more about their fundamentals. [Damien Boureille] has put a 60s-era IBM as well as a PDP-11 to work training a transformer algorithm in order to take a closer look at it.

For such old hardware, the task [Damien Boureille] is training his transformer to do is to reverse a list of digits. This is a trivial problem for something like a Python program but much more difficult for a transformer. The model relies solely on self-attention and a residual connection. To fit within the 32KB memory limit of the PDP-11, it employs fixed-point arithmetic and lookup tables to replace computationally expensive functions. Training is optimized with hand-tuned learning rates and stochastic gradient descent, achieving 100% accuracy in 350 steps. In the real world, this means that he was able to get the training time down from hours or days to around five minutes.

Not only does a project like this help understand these tools, but it also goes a long way towards demonstrating that not every task needs a gigawatt datacenter to be useful. In fact, we’ve seen plenty of large language models and other generative AI running on computers no more powerful than an ESP32 or, if you need slightly more computing power, on consumer-grade PCs with or without GPUs.