On 29 Jul 2008 at 20:55, Philip Pemberton wrote:
At the moment it's driven permanently high;
this seems to work OK insofar
as the sync detector (which is why the datasep is there -- to extract a clock
signal for the MFM data stream) will pick up SYNC-A1 signals and the data
seems to be valid. What I don't know is if this is how things are supposed to
be done...
ISTR (and it's been a long time since I read the document) that the
later versions of the 765 (765A?) imposes something like a 500 usec.
VCO inhibit after the leading edge of INDEX/. Earlier versions (765,
8272) imposed something like a 1000 usec. VCO inhibit. This can be a
real problem if the original disk was not formatted with an IAM,
leading to failure to recognize the IDAM for the first sector on the
track.
Even so, the 500 usec. "blind spot" exhibited by the 8272A/765A can
be a problem for diskettes formatted on other systems where a too-
short gap occurs before the first sector header. This leads to all
sorts of dodges in reading them on PCs, such as taping over the index
hole or tweaking the drive spindle speed down a bit.
WD controllers of the 17xx family did not do this; you could start a
sector very close to the index pulse and still be fine.
My advice would be to fuggedaboudit, particularly if you want to read
diskettes produced on other systems than PCs.
Cheers,
Chuck
The uPD765 and 827x datasheets are predictably rather sketchy on this
front... All they really say is that the VCO line inhibits the VCO in the PLL,
which would have the effect of allowing the PLL's loop filter to discharge,
and reset it to a predetermined state. What they don't say is under what
conditions the FDC will do that...
So I guess the million dollar question is what I should do with said VCO line.
Wire it to /INDEX via an inverter to reset the PLL on every rotation? Or just
wire it to VCC (VCO enabled) and leave it?
Thanks,
--
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/