3

Over the years during which Commodore sold tape drives, there have been a few different versions of the drive with different analog electronics. I'm interested in experimenting with "turbo" tape formats, but want to avoid writing anything which wouldn't work reliably on a wide range of drives.

Are there any documents which would describe things like the relationship between write-pulse widths and read-pulse timing uncertainty, or other such factors that would need to be tweaked to yield optimal performance?

7
  • 2
    Not an answer but over the many years in different platforms (must say I have virtually no CBM experience though) three of the largest issues were poor quality media, azimuth problems and mechanics like wear such as stretched or perishing drivebelts resulting in speed variation. The latter if minor and the baud rate is low you may get away with but when doing various high speed tests on Spectrums say up to 4000 baud you needed everything to be 100%. I think the answer provided would likely apply to other marques generally speaking too so I shall follow this question with raised interest. – AndyF Feb 8 '20 at 22:54
  • 3
    @AndyF: Looking at documentation for turbo loaders, it seems like most of them use a certain length cycle for a 0 and a cycle roughly twice that long for a 1. If a tape can resolve cycles of length 1 and 1.5x, such an encoding would reduce the average time per byte from 12 fast cycles to 10. If a tape can resolve lengths of 1.0, 1.5, 2.0, and 2.5, then the average time per byte could be reduced from 12 fast cycles down to 7 (encoding two bits per pulse). If it can resolve eight lengths from 1.0 to 4.5, that would improve to 6.6 (three pulses per byte). – supercat Feb 9 '20 at 18:41
  • 2
    Incidentally, I would not be at all surprised if making tape pulses yield predictable rising-edge timing would require adjusting the high pulse lengths based upon preceding data so as to ensure consistent DC bias. This may make it impossible to reliably duplicate tapes using equipment designed for audio, but back in the day I suspect many companies would have viewed that as a benefit rather than a drawback. – supercat Feb 9 '20 at 18:50
  • 2
    To comment on the very last sentence (more later) yes it was an advantage to make life difficult for home tape2tape enthusiasts :) ... Other platforms (Sinclair and Amstrad at least) with early Speedlock loaders had a clicking pulse in parts of the leader tones for example which and I half suspect was quite intended to play havoc with home recorders AutoGainControl for instance. I can honestly tell you we only owned a single tape player back then, so any ahem "backups" were dependent on me working out how to dismantle the protection system. To be fair this was a brilliant learning aid for me – AndyF Feb 9 '20 at 18:57
  • 3
    @AndyF: My big concern with developing such approaches would be ascertaining to what extent DC bias would affect different cassette drives differently as a consequence of different playback circuitry. I wonder how much extra, if any, it would have cost to make the playback circuitry detect phase transitions with symmetrical hysteresis, as is done on floppy drives? That may have required more explicitly working to avoid DC bias (as with e.g. Manchester codes) but would also have allowed for better reliability and faster data rates. – supercat Feb 9 '20 at 18:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.