Re: GCR Encoding

[Eric Smith <eric-no-spam-for-me@brouhaha.com>]

"Bryan Parkoff" <bryan.nospam.parkoff@nospam.com> writes:
>     Can Disk II Family read three or four (zero) bits that it will always
> show invalid?  Is it the truth that Disk II's hardware is designed to read
> only one or two (zero) bits?

Yes, that's correct.

> Can you explain why?

The Disk II controller (and IWM) can't reliably read "nibbles" containing
more than two consecutive zero bits.  In fact, the older 13-sector state
machine PROM (P6) can't reliably read more than one consecutive zero bits.

The one bits in the nibble are stored as flux transitions on the media.
Zero bits are only indicated as the lack of a flux transistion.  On
a write, every time there is a pulse on the write data line, the drive
"analog board" electronics and head cause a flux transistions.  On read,
the head and analog board detect the flux transitions and produce pulses.

The controller transfers data using a shift register running at a
crystal-controlled rate, but the drive motor speed is not sufficiently
well regulated to make it possible to accurately time long streams
without flux transistions.

Also, there is speed variation from one drive to another.

The bit time is nominally approximately 4 microseconds.  Suppose
you have a disk that was written on a drive that was 8% fast.  Now
you read that disk on a drive that is 8% slow.  The bit timing will
be 4.7 microseconds.  On the other hand, you might write a disk on a
drive that is 8% slow, then read it on a drive that is 8% fast.  This
would make the timing 3.4 microseconds.

Here are some example timings:


           interval between pulses

                    fast wr    slow wr
data     nominal    slow rd    fast rd
-----    -------    -------    -------
11        4 us       4.7 us     3.4 us
101       8 us       9.4 us     6.8 us
1001      12 us     14.1 us    10.2 us
10001     16 us     18.8 us    13.6 us


Suppose the state machine sees two pulses with an interpulse delay of
13.8 microseconds.  Should it decode that as 1001 or 10001?  It is
ambiguous.

Restrictions on consecutive zero bits are used in all disk channel
codes.  The IDE or SCSI drive in your PC or Macintosh has such
restrictions.  However, they are dealt with entirely by the embedded
control system of the disk drive, so you as an end user never notice it.

With the Disk II controller, this restriction has to be dealt with by
software running on the main processor.

> Why do each byte require between $80 and $FF?

Because as the bits are being shifted in from the disk, the only way the
software can tell that a whole nibble has been read is by the most
significant bit being set.

Suppose that this restriction did not exist.  Then the values $FE and
$7F would both be valid nibbles.  When the software is reading a nibble
and gets a $7E, how would it know that the full nibble had been read,
and that there wasn't another zero bit in the nibble about to be shifted
in, making it $FE?

>  I do understand that between $00 and $7F can't be used
> because it will show more than two (zero) bits.

That explains $00.  The only reason $7F can't be used is that there
would be no way for the software to tell that the full nibble has been
read.

> How did Steve Worniak decide to design logic chip that can only handle
> two (zero) bits?

The logic chip (74LS323) doesn't care whether the bits are 0 or 1.  Only
the software cares.

> I do know that he decides to use software by reading each bit.

The software does NOT read each bit, at least not individually.  It
loops reading the shift register until the most significant bit is set.

> If hardware is chosen to use its own hardware, three or more (zero)
> bits can be accurate because the rotation speed is always fixed.

The rotation rate is never "fixed".  Even on drives that use a tachometer
and servo for motor speed control, there is always some variation.

> I am so curious why Apple II decide NOT to use standard MFM encoding
> that it does exist before 1970s, but all PC like 8080 and/or 8085 use
> MFM encoding (probably FM encoding) before 1970s.

Because in 1977 when Woz designed the Disk II controller, it would have
cost a lot more money, and used a much larger circuit board, to implement
MFM.

The Disk II controller *is* capable of reading and writing FM.

> Please explain every detail if you have great knowledge.

Please pay me my standard consulting fee, and I'll be glad to explain
every detail.

Alternatively, get a book.  There are several that explain this stuff.
Beneath Apple ProDOS is a good start.  (There are some fundamental
errors in the description in Beneath Apple DOS, so I don't recommend
it as a source for information on the low-level operation of the
Disk II, but it is still *very* useful at explaining how the Apple DOS
software worked.)

There is another book that has an even better explanation of how the
Disk II controller works, but I don't recall the title or author at
the moment.