17 Jul
2006
17 Jul
'06
3:58 p.m.
Chuck Guzis wrote:
If those are segment addresses, they would be quite normal for a BIOS
extension ROM.
My guess is that this board was designed to go in an XT, and the
extension ROM adds high-density disk support.
Getting back to the UM8390 chip. It's a single-chip PC/AT disk
controller, it does mostly what you'd expect. Personally, I've never got
one to work in FM mode, though. I do have the data sheet somewhere, which
I can find if you're interested.
It has an on-chip address decoder which decodes A0..A9 to put the
765-like part, and the other I/O ports, and the expected addresses. From
what I remmeber, there's a pin you tie high or low to select the primary
or secondary controller addresses. Alternatively you can just invert one
of the high-order address lines (I forget which one, it's trivial to work
out) going into the IC.
There was a similar chip -- UM8389 I think -- which was a PC/XT disk
controller. It didn't do high denisty, it didn't have the clock selection
port, and it used a different crustal (IIRC 8MHz, the 8390 uses 24MHz).
Otherwise it was pin-compatible. Many years ago, I bought a multi-I/O
board for an XT that used this chip. I desoldered it, replaced with with
a 8390, changed the crystal, and set it to the secondary address. Used
that to link 8" drives to my XT...
-tony
To my somewhat jaundiced eye, the UM8398 looks to
be yeat another
integration of the NEC 765 with some support circuitry. I think I've even
got one of these on an ISA board and it looks for all the world like a
plain-jane PC/AT style controller.
Well I've got one here on an 8 bit ISA board - but slightly worrying is that
it has its own PROM on board, and seems to be jumper selectable for an oddball
address (0xC800, 0xCC00, 0xCA00).