Extracting data from an old computer system

The hack to extract the data is slow, but is probably reasonable, giving the problem. Article states the computer is possibly a TI-990.

The customer had brought in the massive rack-mounted box into the workshop, along with one of the RS232-connected dumb terminals. Aaron opened it up and found, among the dust and grot, a 120MB Maxtor hard drive (probably from the 1990s) and “a tape drive so old that the rollers had long since melted into a mushy pile of sticky liquid”.

No backups then. Nice.

There was no MS-DOS here, and certainly no Windows. The PFYs employed by the computer shop had no idea what to do with it. And so the call had gone out to the old man who knew about this stuff, a man who had cut his teeth on Commodores, Apple IIs and mouldy old IBM PC compatibles, a man who was in his early 30s and thus impossibly old in the minds of the under-25s that made up most of the staff.
So somehow Aaron was going to have to get the data off the wheezing old beast and on to something a little more modern.

The TI machine ran an archaic OS, had what appeared to be a very proprietary bus for the HD controller and no working external storage. The workshop was baffled by it, but our man “devised a plan”.


Yikes, low-tech but relatively universal:

“We printed every conceivable report that contained the data they needed,” he remembered, “hundreds of pages that I intercepted at 4800 bits/second with a ‘vampire tap’ of the serial lines, connected to a USB serial adapter on a modern PC.”

The data was then shovelled into Excel, given a good fettling and finally exported in CSV format.


I feel like Curious Marc’s SUPERTEAM could have come up with something more elegant.

I think, apart from the vampire tap, just tapping into the serial line is actually a rather elegant solution, compared to the alternative of reverse engineering an entire protocol for just a single use.
We learn, the vampire tap is the poor man’s logic analyzer.


It certainly beats trying to OCR acres of printout. Isn’t there a similar tapping of a serial line in the Cuckoo’s Egg story?

The serial port is the next best thing to [insert here anything that appears really tasty to you in your wildest dreams], when it comes to communicating with unknown systems. If in doubt, just print. It’s a pity they are gone.

(As a Mac user, what I’m even talking about? :slight_smile: )

P.S.: Is there such a thing as a universal retro printer emulator to be run on the (modern) recovery system? A simple printer spooler should probably do.

1 Like

Sorry, my favorite hack was when a group managed to hack in to an iPod and got control of the speaker. They then wrote a bit of code to compress and “download”, via the speaker, the entirety of the iPod ROM.

They then stuck it in an acoustic box and recorded it, finally recreating the ROM by demodulating the data they had recorded. Inspired.


Aha, you remind me of an experiment…
Puzzle challenge: An unusual idea, but not a new one
(Here’s the follow-up post where I explain my thinking and describe my results.)

Check CuriousMarc latest video for more of these sorts of stories:

Good god man that is both inspired and terrifying.

Meh, having a serial line to tap into is so much easier than having to build a neural network to OCR a video recording of a CRT. :-)

But yeah, the general issue of getting data out—and in!—is a constant problem. I’ve been able to record cassette tape saves from my old 8-bitters on my phone, but not had any success reloading them yet. (I need to build a small amp, I think.) Still that’s gotta be easier than Ed’s method.

And that still doesn’t help for code developed on a modern machine that I want to move over to machines with no other I/O than that; I’m thinking I need to put a microcontroller on the external expansion bus, have it pause the CPU, and do direct reads from and writes to the machine’s RAM.

1 Like

I’m reminded of some thoughts I had related to trying to dump the ROMs from an Acorn Communicator - there was a working one at TNMoC that I’d had my hands on. The Communicator has an edge connector, a printer port, and a serial port, so one idea was to try to rig up a microcontroller or similar to log the output from one of the ports. I never tried it. In fact, the next time I visited, there was an attending curator who didn’t look like they’d want me even to power the machine up.

At a later point, and much more recently, @RObC lent me his Communicator, and the same question arose. In the interim, @hoglet had developed his pixel-perfect (and very low latency) HDMI adapter based on a CPLD and a Raspberry Pi, a setup which can take pixel-perfect screenshots to the Pi’s SD Card. So I thought it mightn’t be too bad to write a few lines of Basic, dump the ROMs to a series of high resolution screens and capture each one. But, before I got around to that, I was able to ascertain that the ROMs in this machine were a version that’s previously been dumped, so no need…

1 Like

Also, mind this previous post on a very complicated way to extract the ROS (IBM term for ROM) from a IBM 5100: