FDC Gap Length?

[Allison]

>
>Subject: Re: FDC Gap Length?
>   From: shoppa_classiccmp at trailing-edge.com (Tim Shoppa)
>   Date: Sun, 17 Jul 2005 08:25:30 -0400
>     To: cctalk at classiccmp.org
>
>> the gap length you arrive at is different from the one in the NEC table
>> [for different sector sizes]
>
>Part of the calculation is the time it takes the data separator to re-sync
>after each gap.  Now, a "true" 765 doesn't have a data separator so maybe
>they leave this out.  It's also true that this number ought to be nearly
>constant depending on sector sizes.

The base 765 expected there would be some form of external data seperator
either digital or PLL.  In 1981 none of the FDC chips had data speration
as it was pushing the amount of variable hardware on the chip. In later
years that was integrated into the chip.

The gap will not be the same for all sector sizes as it's it serves the 
function of filling the space between sectors. The gaps are those bytes
that are accounted for in the raw and formated disk space.

>Do you have the Intel 8272 data sheet?  (This was Intel's part number
>for their 765 clone.  It may not just be a clone but the exact same
>mask...).  IIRC there are more details/examples/math in the 8272 data
>sheet, including some recommendations for data separators and the good/bad
>of different types.

Intel is a licensed mask.  Their data sheet is not any more informative 
than NEC full data sheet and user manual, it may be easier to find 20+ 
years later.  The IBM softsector floppy disk specification is the origin 
point for the whole show.  I may add that the later chips are 765B or
7265 and Intel never did those.  The differnce on those are coupled to 
gaps 4a and 4b and the index gaps.  Those came into being with the 
3.5" floppy.


Allison