While you can still find tape being used for backup storage, it’s pretty safe to say that the humble audio cassette is about as out of date as a media format can be. Still, it has a certain retro charm we’re suckers for, particularly in the shape of a Commodore Datasette. We’re also suckers for miniaturization, so how could we not fall for [bitluni] ‘s tiny datasette replica?
Aesthetically, he’s copying the Commodore original to get those sweet nostalgia juices flowing, but to make things more interesting he’s not using compact cassette tapes. Instead, [bitluni] started with a micro cassette dictaphone, which he tore down to its essentials and rebuilt into the Commodore-shaped case.
The prototyping of this project was full of hacks — like building a resistor ladder DAC in an unpopulated part of a spare PCB from an unrelated project. The DAC is of course key to getting data onto the micro cassettes. After some playing around [bitluni] decided that encoding data with FSK (frequency-shift keying), as was done back on the C-64, was the way to go. (Almost like those old engineers knew what they were doing!) The dictaphone tape transport is inferior to the old Datasette, though, so as a cheap error-correction hack, [bitluni] needed to duplicate each byte to make sure it gets read correctly.
The micro cassettes only fit a laughable amount of data by modern standards this way (about 1 MB) but, of course that’s not the point. If you jump to 11:33 in the video embedded below, you can see the point: the shout of triumph when loading PacMan (all 8 kB of it) from tape via USB. That transfer was via serial console; eventually [bitluni] intends to turn this into the world’s least-practical mass storage device, but that wasn’t necessary for proof-of-concept. The code for what’s shown is available on GitHub.
If you have an old Datasette you want to use with a modern PC, you’d better believe that we’ve got you covered. We’ve seen other cassette-mass-storage interfaces over the years, too. It might be a dead medium, but there’s just something about “sticky tape and rust” that lives on in our imaginations.
Thanks to [Stephen Walters] for the tip.
Using something like v.92 modulation you could perhaps get 9 megabytes out of a 60 minute cassette tape at 56 kbps using both stereo channels. With an even better modulation scheme you could make use of more of the audio bandwidth of about 10-15 kHz, giving 200 kbps and perhaps up to 120 megabytes for two channels for a 60-minute tape.
The very best music-rated microcassette recorder mayyybe got up to 9 or 10kHz (yes they existed! Super cool stuff!) but if it’s a dictaphone tape mechanism, if you reach 4kHz mono you should be very happy.
wowing and fluttering away…
So 4.5 megabytes per 60 minute tape then, using v.92, which was designed to transmit over POTS phone lines with around 4 kHz audio bandwidth as well.
Pactor IV might work with such an unreliable medium.
It can reach 10 KBit/s under best conditions.
But the current Dragon PTC totally is in another price league.
It’s the best there is for communications on shortwave, which is equally poor to datasette.
More information:
https://en.wikipedia.org/wiki/PACTOR
https://www.p4dragon.com/
With tape, problem is not so much noise, but varying tape speed (wov and flutter) causing phase and frequency errors. Pactor IV and other modulation methods meant for radio or phone link do not tolerate those very well.
Hi! You have fading, splatter, “echos” on shortwave and signal loss, too.
SSTV was used on both music cassette and shortwave, too and suffered on both mediums similarily.
PACTOR and the like wouldn’t work, since they use ARQ and rely on the sender repeating if the receiver doesn’t acknowledge receipt, which a stream of data from a tape can’t do (easily). Also, what OH3MVV mentioned.
Hi, but V.92 can be rather used? 🤔
@Joshua: The entire v.92 protocol? Probably not, but the underlying PCM scheme should work. Whether it’s better on tape than any other modulation scheme is another matter.
Joshua, no, v.92 is no good for cassettes and microcassettes, because their amplitude and phase output at high frequencies is so inconsistent. I worked at TEAC in its glory days, and I can tell you the cassettes’ performance at high frequencies was absolutely atrocious compared to open-reel.
v.02 was a modem thing. To reach 56k required digital connections in part of the phone line. I don’t really understand all the details but I don’t think it’s just a digital to analog modulation scheme that you can plug into just any old analog recording or transport device. It was specific to the way late 90s POTS worked.
v.92 mot v.02
The datasettes got 100 kByte per 30 minute side, I doubt you could get much more than that reliably. The fact that the builder of this had to double every bites vs adding a ECC bit for each 8, tells me he probably tried that and had too many times where there were more than a single bit flip. I bet his system is rather unreliable.
C64 with turbo did ~1MB per tape
I have no idea why Bitluni went with ADC/DAC arrangement and !FFTs! in the code. Tape is a digital medium, it stores flux flips not analog values. 8bit computers treated it as such with simple comparator on the input.
This was the most painful part of the Commodore legacy. I could not afford a floppy disk drive back than, so I was a necessity back than. I could not forget the bad user experience of the cassette deck even after 30 years. Cassette desk shall burn is hell with its designers in hell!
C64 Datasette was pretty chill and worked with all kinds of Turbo loaders out of the box. Brand new legal games came with custom turbos and loaded pretty fast. Not to mention crazy scene/pirate Tape Turbo loaders go over 10000bps
If you want nightmare fuel making you wake up at night all sweaty talk to some Atari owners.
1. Faster Atari Turbo loaders (2600 etc) absolutely required hardware modification to the tape drive! All because Atari Tape drive has build in hardware FSK decoder/encoder.
2. Atari 8-bit Tape Loading routine burned into rom has a bug that will randomly fail Tape load no matter what. No amount of nobody move for 22 minutes because Im loading a game will help you :)
C64 on the other hand worked in a much simpler and cheaper way connecting CPU IO pin directly to magnetic Head :) CPU did all the heavy lifting managing software PWM.
Always a good day when there’s a SLoM reference in the wild!
Great hack, too!
I thought someone hacked USB sticks to look like mini tapes, so this was a pleasant surprise!
There had been actual 8-Bit systems which used MCs as datasette.
Such as Sharp PC-1500 pocket computer.
https://en.wikipedia.org/wiki/Sharp_PC-1500
https://en.wikipedia.org/wiki/Microcassette
And Philips’ P2000T which used the slightly bigger Mini-Cassette.
https://en.wikipedia.org/wiki/Philips_P2000
https://philips-p2000t.nl/
That Sharp design still looks very nice. Can’t find who made that for them, lost in time…
What I think is cool about this project is that he didn’t just 3D print a mock mini version of a Datasette that is otherwise just an SD card reader, but that this is an actual, functioning tape drive with a homemade DAC. Impractical, yes. Impressive, HUGE YES!
Micro cassette not mini cassette, which is slightly large and more obscure.
Good catch, I’ll fix it in the text.
I love this project. Fwiw, the original datasettes were about 300 baud. Because of the whole serial bus screwup, the 1541 disk drives weren’t much faster.
I’m pretty sure the donor mechanism here is a step down, yet this outperforms the datasette handily. Well done!
`load “*”,1,1
What also would have been cool as a datasette medium is VHS-C tape.
Just like regular VHS, the audio quality was on par with CD quality.
In fact, there had been audiphile people who used the VHS to record musuc from digital satellite radio way back in the 1980s.
HiFi capable VCRs would also record in stereo on a vertical helical track (mono track was linear recorded).
In my opinion, VHS (any form factor) would have been a good, mass-produced medium for audio recording.
The bandwith of such HQ datasettes would have been manifold! What a missed opportunity in history.
Even with a “primitive” VHS mono audio cassette deck, the quality was mich higher than the MC.
Sorry for the typos!
VHS was a good, mass-produced medium for audio … CD mastering :) https://en.wikipedia.org/wiki/PCM_adaptor
Amiga missed big market by only including partial genlock hardware (needed pulse external generator), with full genlock and one dedicated input pin they could have had native support for reading PCM data from VHS tapes.