On Wed, May 31, 2000 at 05:36:39PM +0100, Jacob Dahl Pind wrote:
Oh, i might try some of this info I have a catweasel,
sitting in my amiga
thoug.
FYI, I read through my code and found that that spiel I posted earlier had
the sense of the STEP bit backwards, it is of course 0 to assert it, since
the control register is simply a latch which drives the FDC bus directly
(through NON inverting drivers). Dunno about the Amiga Catweasel but I'll
bet it's close to the same, you can't get much simpler than this.
I'm having a lot of problems with what look like crazy numbers coming out of
the Catweasel/ISA RAM every so often. In one case I caught a situation where,
on two different reads of the same sector header, every interval (between
transitions in the FDD head data) is identical, except for one which is 25%
lower and results in those transitions being decoded as 001 instead of 0001,
so the whole header comes out wrong. I sent Jens a note but he's too busy
getting ready for a show to deal with it right now, but I really don't see
how that can be fixed in software. It would be one thing if a long interval
appeared as several short intervals due to extra apparent transitions caused
by noise (software could filter that out, up to a point), but a single short
interval seems more like a hiccup in the interval timer (which measures the
time between transitions in the RDATA pin, in 14 or 28 MHz ticks).
John Wilson
D Bit