20 Mar
2011
20 Mar
'11
6:18 p.m.
On 20/03/11 03:23, Chuck Guzis wrote:
Occasionally, I get a request for the 3.5"
version of the 8" NEC
drive that you've got for some PC98-based CNC controller. They're
pretty rare on the ground and have all of the weird signals. You
can't just take a Teac FD235HG and swap one in.
They're not cheap when you find them.
I'll bet!
It seems the "track 0 sensing quirk" is a little more complex than I
originally thought. The old DiscFerret microcode (anything pre-rev
0x0021) would only check the state of TRACK0 after making STEP active.
You were guaranteed at least one STEP pulse, even if TRACK0 was
active... "Oops."
This seems to upset the 8048 controller on the NEC drive. While it won't
block the step operation, it will totally screw up the track 0 sensing
until the head is stepped out a few tracks, then back to TK0. There also
seems to be some significant delay between STEP going active and TK0
becoming valid -- far more than any other drive I have...
The workaround, of course, is to check the state of TK0 both while STEP
is active, and while it's inactive. If it goes active when STEP is
inactive, then the drive is just slow to respond... :)
I've added the Step Termination Reason bit too -- this is a bit in the
STATUS1 register which is set if the last seek was aborted due to
reaching track 0.
Now to convert my scribbly paper notes on the FD1165S into Wiki markup,
and hunt down a Famicom Disk System drive and discs.....
--
Phil.
classiccmp at philpem.me.uk
http://www.philpem.me.uk/