James.Reilly@tel.vtt.fi (Jim Reilly) (11/16/90)
I have a question about our Sparc stations (150 Mb) tape drive, well
actually about the tapes. The tapes (Sony QD-6150) are 620 ft, 12500
ftpi.
What does ftpi stand for ? I've assumed it means bpi, and that the
calculation for how much can be stored on it is:
(600 ft)(12500 bits/in)(12 in/ft)(18 tracks)(1 byte/8 bits) =
202500000 bytes
which should roughly hold 150 Mbytes of real data. Am I correct, or
hopelessly wrong ?
(please respond to reilly@tel.vtt.fi, as I don't read Usenet too often)
Internet Mailbox: James.Reilly@tel.vtt.fi
UUCP Mailbox: ...!{mcsun, uunet}!tel.vtt.fi!reillywolfgang@uunet.uu.net (Wolfgang S. Rupprecht) (11/29/90)
James.Reilly@tel.vtt.fi (Jim Reilly) writes: >I have a question about our Sparc stations (150 Mb) tape drive, well >actually about the tapes. The tapes (Sony QD-6150) are 620 ft, 12500 >ftpi. >What does ftpi stand for ? Flux Transitions Per Inch > I've assumed it means bpi, and that the calculation for how much can > be stored on it is: Well, the asymptotic maximum is one bit per flux transition. Normal encodings of data range anywhere from 1 bit per 2 transitions (MFM) to over 0.9 bits / flux transition for some of the RLL encodings. The reason you can't get exactly 1 bit / flux reversal is that you need to guarantee some transitions so that you can keep your recovery clock sync-ed to the data stream. If you get a string of thousands of 1's or 0's recorded with a direct recording scheme, its hard to determine at recovery time where each bit starts and ends. Your recovery clock would slowly wander and you'd end up getting a few more or less bits than were really recorded. On the other end of the scale, its often advantageous to not try to record a very high frequency 0-1-0-1-0-1-0-1 pattern. High frequency patterns require very closely spaced bits on the tape. Tape heads will tend to average adjacent bits, so its best to avoid this pattern also. Something like RLL 2,7 will make sure that there are no "runs" less that 2 bits wide, or more than 7 bits wide. This neatly solves both of the above problems. Wolfgang Rupprecht wolfgang@wsrcc.com (or) uunet!wsrcc!wolfgang Snail Mail Address: Box 6524, Alexandria, VA 22306-0524