[Info-vax] The (now lost) future of Alpha.

[Scott Dorsey]

>> I did come across a tape controller and drive combination which
>> supported gapless tapes, circa 1995, the intention being to digitise
>> them in-house.  It's been a long time, but the names Pertec (possibly
>> the controller) and Tandberg (the drive) are lurking in my memory banks.
>> 
>> > Were such gapless tapes still around much in the 1990s? No
>> > idea - but preserving the original lossless recordings was 
>> > probably better value (for the exploration companies)
>> > than sending a survey team out to repeat the survey because
>> > the tapes had become unreadable over time, and data had been
>> > lost.

For the most part, gapless tapes were made by incremental tape drives which
wrote one word at a time.  The Kennedy drives we used wrote at 800bpi and 
had a simple serial interface.  You put eight bits on it, pulled the strobe
down, and it wrote a byte.  When you were ready to write another, you did 
it again.  It was mindblowingly slow compared with a streaming tape drive,
which was a good thing because the data being put onto it was coming slowly.

Some incremental drives could write interrecord gaps, but because they were
connected up to a data recording system that was operating at a constant
rate with no computer and no buffering (in our case it was just a rack of 
TTL and A/D chips), writing the gap would interrupt the data.  Later on
with the coming of the FIFO chips from Monolithic and other similar
hardware we were able to buffer 16 words of data at a time and shoot off
an interrecord gap every 512 words.

If you have to get data off gapless tape today, the best (but not very fast)
solution is to use the same type of incremental drive that was used to create
the tapes.  For the most part these were lower bit density (300, 556, or 800)
than computer tapes.

One of the wind tunnels at work still has a couple incremental drives in the
control room just for the purpose of reading old tapes from the sixties and
seventies.  By 1980 or so they had gone to computer-based data acquisition.
--scott
-- 
"C'est un Nagra. C'est suisse, et tres, tres precis."