Bake It To ReMake It: Cooking Old Magnetic Tape To Recover Data

DEC microVAX with tape drive

Those of us old enough may remember the heyday of the text adventure game genre from the first time around. London-based Magnetic Scrolls was an early pioneering company producing titles for the first Amiga and Atari ST platforms. Fast-forward to 2017 and [Hugh Steers], the original co-founder and core developer for Magnetic Scrolls has formed an initiative to revive and re-release the original games on modern platforms. Since the 1980s-era DEC MicroVAX used originally for development is not particularly rare in retro computing circles, and media containing source code was found in someone’s loft space, reviving the games was not a tall order.

First, he needed to recover a copy of the original source code from the backup tapes. But there was a problem, it turns out that the decaying tapes used a unstable polyurethane-based binder to stick the oxide material (which is what stores the data) to the backing tape, and this binder can absorb water over the years.

Not much happens until you try to read the tape, then you trip over the so-called sticky-shed syndrome. Secondly you may find that a small amount of the oxide layer sheds from the tape, coating the read head, rollers and guides inside the complicated tape mechanism. This quickly results in it gumming up, and jamming, potentially chewing up the tape and destroying it permanently.

This was further exacerbated by the behaviour of the DEC TK50Z tape drive, which needed to shuttle the whole length of the tape as part of its normal operation.

A temporary solution was to bake the tape in an oven to drive out the moisture and reduce the stickiness enough to run it through the drive safely. Then only the oxide-shedding problem remained. The TK50Z drive was swapped for a TZ30 which shuttles the tape less, but also critically with a simple hack, would allow the heads to be cleaned with IPA between read passes. This was enough to keep the gumming up at bay and allow enough data to be read from the tapes to recover several games worth of code, ready for the re-releasing process.

The video after the break shows [Rob Jarratt] working through the process of the data recovery.

12 thoughts on “Bake It To ReMake It: Cooking Old Magnetic Tape To Recover Data

  1. Been there with all my old 1/4 inch analog tapes.
    My earliest 1973, latest 1999.
    Baked them 60C for 3 hrs per batch of 2 or 3.
    It was a long boring process but very glad to have it done.
    I wasnt very good at separating data tapes from my song recordings, so interspersed with my turgid songwriting efforts, are FSK program recordings from my old Acorn Atom.
    Not sure which set of recordings are more musical. :-)

  2. For a while, audio tape manufacturers provided special tape for language labs that had very high wear resistance. This was achieved by putting a thin plastic layer on top of the oxide layer. A disadvantage was that it increased the distance from the head to the oxide layer, reducing high frequency response.

    That suggests a possible technique of coating the baked tape with a thin plastic layer to reduce or prevent shedding. It’s likely that a lot of experimentation would be needed to make it work before trying it on a rare sample.

  3. Genuine question here: I understand that the friction between the tape and moving parts (rollers, reading heads, etc.) may shred some material and/or turn the oxide layer into some nasty gunk…

    …however, as tapes are spooled against its own “back”, would the unspooling pose a “sticky problem”? Could the process of “baking” to dry it offer the risk of make the front face oxide adhere to the back face of the tape spooled over it?

    Really curious!

    1. Wasn’t there an article on here a while back about old film tape basically becoming a single solid lump because the media would glue itself together as it aged/ degraded? I think I remember they used x-ray or MRI to somehow recover a surprising amount of data off of it.

  4. For newcommers, this is the DEC tape technology that was later adquired by Compac and then adquired by HP, the so called DLT tapes. Although LTO is not endorsed with DLT they are mostly the same.

    I still have a couple of DLT drive units just because this.

  5. Oxyde shedding and sticky-shed syndrome, but also dust and molds are common problems with audio and video reels and cassettes.

    That’s why in serious audio/video digitization businesses, before trying to read any tape, we first clean them in specialized cleaning machines providing 3 levels of cleaning:
    – both sides of tape are rubbed against non woven tissue to remove loose durt/oxyde, sometimes helped with a air pump to suck air through tissue
    – front side is going against a sapphire blade that removes slightly stuck particles
    – front side also goes against burnishers (basically metal posts) to burnish the surface

    I do have RTI Tapechek (yes: chek, not check!) cleaners for 1/2” Betacam family tapes, 1/2″ VHS family tapes, and 3/4” U-Matic/BVU family tapes. But they exists for other formats: 1” audio/video reels, 1/4” DV/DVCam/DVCpro…

    This cleaning is mandatory to at least optimize the reading quality (less noise and drop outs), but also to preserve the player VTRs. But even with that, we must manually clean players’ heads/drums and other parts in tape path very often.

    Since TK50, DLT and LTO formats all use also 1/2” tape, it could be feasible to temporarily transfer tape from its original cartridge to a Betacam or VHS cassette in order to clean them the same way. Or to use some spare parts from these machines to build one.

    And as being said in the article, for the sticky-sched syndrome, tapes must be baked, which can temporarily restore they playability, long enough to digitize them.

    Some brands are known to be much more impacted by sticky-sched and/or oxyde schedding: Ampex is one of the worst, but also 3M or Fuji for example.

    And this is not only true for 40 years old analog tapes, but also for much newer formats like Digital Betacam for example. One of my customer here in France had big stock of Fuji’s brand Digital Betacam tapes that had a surprising high level of loose oxyde and were difficult to read.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.