2 Nov
2011
2 Nov
'11
12:45 p.m.
On 2 Nov 2011 at 14:42, David Riley wrote:
It's ridiculously easy to do A/D and D/A on modern
micros, and even
the cheap, easy-to-drive ARMs have USB interfaces. Should be a cheap
board to build and stick a JTAG interface on to let people muck around
with their own firmware as much as they want.
Given that almost all disk drives are purely digital devices, there's
no need for A/D conversion. Basically you need a microcontroller
with "capture" (almost all have it) and PWM (again, almost all have
it) and a bunch of memory (128K should handle everything up to about
a "1.44MB" floppy; 256K if you want "2.88MB" capacity. Even an AVR
running a 20MHz is more than sufficient for most stuff. Perhaps
using an ARM with a faster timer improves resolution, but there are
decreasing returns as the clock speed goes up. Floppy drives are
pretty "wiggly" devices in the real world.
And naturally, USB capabilities.
It isn't hard--I've done it. The work is in the decoding./encoding
software, not in the hardware.
There are other competing devices--one called "Deviceside" that seem
to have almost disappeared. My guess is that you wouldn't get wild
interest.
Far more interesting is a floppy emulator. I did a proof-of-concept
one using a 16MHz AVR reading/writing to SDHC to emulate a 360K
floppy drive. It takes raw sector-by-sector data and encodes it.
Again, very simple (There are competing products, but, AFAIK, no open-
source ones.).
It really isn't rocket science.
Phil's gizmo is a bit different because it can handle datarates into
the ST506 drive range. Still, I wonder if a fast-enough uC might not
be adequate for the job.
I might be happy to participate in either, but I don't want to wind
up fielding the PCB design and all of the software support. I've got
better things to do with my time.
--Chuck