On Thu, Mar 1, 2018 at 12:56 PM, Ethan via cctalk <cctalk at classiccmp.org>
wrote:
That is a very standard IDE drive that you can replace
with just about any
IDE drive you can find, at least to get things up
and running. The
controller won't support the faster transfer speeds of later drives, and
may not support the full capacity of the larger drives, but the newer
drives should be backward compatible. Something in the 500MB range would
probably be a good choice. Also, looking at the information I have on the
motherboard the drive controller can't be disabled. You may be able to add
a secondary drive controller, but booting from the hard disk on that
controller _may_ not be possible.
On the old 386 era PCs you have to specify cyl/head/sector/lz type stuff
in the BIOS usually? It's possible to sub in a CF card on the IDE bus with
a cheap adapter, but I'm not sure how the cyl/head/sector stuff plays out.
Maybe go with something fairly small like 32MB and a CF to IDE adapter
(it's just wires, CF cards are similar to early PCMCIA which is ATA which I
think is just buffered ISA but I could have it wrong?)
I did this recently on a 486 but it had an auto-detect feature for the
hard drive parameters. Maybe they don't really matter when using a CF Card
-- does anyone know?
The CF card has a geometry it returns via the standard IDENTIFY command.
It's just a IDE drive after all. Some of the fields are a little different
than a spinning disk, and some newer BIOSes (newer than 386 era,
mid-486/early-586 era) have heartburn due to that. You can often find out
what it is if you can connect it to an PATA interface. USB adapters,
however, lie.
I've deployed thousands of CF cards over the years, and they all have
different geometries. Our imaging software had to cope (so couldn't just dd
the image on, but instead read the geometry, created appropriate
partitions, newfs'd a FS and mounted it, then copied the files to it), and
it was in the late-486/mid-586 era hardware. I was so happy when LBA
addressing went in...
Warner