20 Oct
2024
20 Oct
'24
7:03 a.m.
Hi Mike!! Oh I was so nervous taking it apart - I was thinking of you,
and worrying about how I'm probably going to end up having to sadly explain
why this system doesn't work anymore. But 'lo and behold, I did manage to
get it back together! (notes are in my PC-5000 page linked earlier, which
is now a bit long winded for a single page - I'll spit it out eventually).
I was trying to go after the CMOS battery, or its equivalent - but wow,
Sharp made that very difficult. As you'll see in the photos, the
component side of the mainboard is actually pointed down. Since this is
still very much a working system, I didn't feel comfortable completely
disassembling it - you have to get the entire mainboard out. But I got to
the vicinity of the area at least. On the positive, my thinking is that
since the battery is "inverted" from normal and pointing towards the ground
- if it does leak, it'll just leak into the plastic base of the system.
With the disk drives, I can now get new software onto the system more
easily. But, looks like I'll have to give up on making a boot disk floppy
for the system. I've added a ton of notes to my PC-5000 page about it.
@Fred Cisin - I did manage to digest and follow your DEBUG.COM advise, and
it all did work (in getting past "incorrect DOS version"). But when it
came to the business of actually executing a format, they still did not
work. As others have suspect, we're just going to need to find that
original Sharp MS-DOS 2.00 boot disk someday. In poking around the MS-DOS
2.00 source code on github, it actually doesn't have a pre-built FORMAT.COM
- instead it has a FORMAT.DOC file that describes notes on what is expected
for an OEM vendor to implement to support doing a format. While we did
finally find a format that allows SYS.COM to work, it's still mysterious on
why the (bootable) bubble memory report 6 hidden files, but a SYS'd disk is
only reporting 2 hidden files (I think someone else here did cover that, in
suggesting some vendors did need extra files to fully implementation their
DOS?) So it's been a valiant effort, but I'm content enough just being
able to move files on/off the system via disks - meanwhile we'll just hunt
for that original Sharp MS-DOS 2.00 image, it surely is somewhere "out
there" eventually.
(again, notes on this are on my PC-5000 page - but the short of it is:
- MS-DOS 1.25 FORMAT.COM didn't care about version but didn't even try to
actually format (locked up)
- (Sharp) MS-DOS 2.11 FORMAT.COM was patched, and would show help on
command line arguments, but wouldn't actually run when given /2 /8 or /S,
etc.
- MS-DOS 3.30 FORMAT.COM was patched, but declared the target drive as
ASSIGNed or SUBSTed and refused to proceed
The disk controller interface on this system just isn't "PC Compatible"
enough.
-SL
On Sat, Oct 19, 2024 at 12:43 PM Mike Stein <mhs.stein(a)gmail.com> wrote:
Great to see my old PC5000 receiving so much loving
attention; thanks
Steve!
FWIW, related to the discussion elsewhere about BBSs, that PC5000 was
originally owned by Canada Remote Systems, a smaller Canadian version of
Compuserve based here in Toronto, It was a fairly large (by Canadian
standards) commercial BBS system of the 80s and 90s, noted primarily for
its extensive file collection; I think I still have some of their
collection disks on 8" media somewhere.
On Sat, Oct 19, 2024 at 10:05 AM Steve Lewis via cctalk <
cctalk(a)classiccmp.org> wrote:
Some interesting things (on PC-5000):
- I copied over QMATH, and old command line "parsing-calculator" I did in
Turbo Pascal decades ago (probably about 1992, so post MSDOS5 at least),
and it runs on the PC-5000! It's packaged in my "VUC" tools here (along
with CDIR) voidstar78/VUC4DOS: voidstar Utility Collection for (MS/DR/PC)
DOS (github.com) <https://github.com/voidstar78/VUC4DOS>
That just speaks well to the PC-5000 really being MS-DOS compatible (and
that it ran an executable from a much later generation of MS-DOS). Note
that CDIR itself doesn't run on that MS-DOS 2.00 system (I suspect
anything
that "touches color" won't run, based on trying to run a few other
similar
type things)
- VER is saying MS-DOS 2.00, same as the startup/bootup note. Though it
does say the "Command v2.02" shortly after (I suspect as it is loading the
command.com, or in any case just prior to invoking autoexec.bat)
- I don't have a "native" DEBUG.COM for 2.X yet (and ended up in a bad
time
that archive.org is majorly down again). And trying to run DEBUG.COM
from
3.30 disks on the PC-5000 just says "Incorrect DOS version" (was we've
discussed, which as mentioned I'll have to debug the debug.com on another
system first to patch it)
I'll dig into the .COM patching later, have some errands this weekend
first. Plus, it turns out I "blew up" my parallel port *again*. Modern
day, we take it for granted about USB being hot-swappable. Well,
parallel-ports apparently aren't that forgiving - and I keep forgetting
that. I've zapped two parallel cards now while swapping between parallel
devices. (which I'm making good progress figuring out the "retro
printer"
that will emulator old printers and let us print from old software and go
straight to a PDF, but still working on it) The LPT devices themselves
are fine, just I really heard the electrical pop and just the parallel
port
is absolutely dead (everything else seems fine though)
More to report later, but was excited that "something from the future"
(qmath.exe built almost a decade after the PC-5000 was sold) worked.
-SteveL
On Fri, Oct 18, 2024 at 6:35 PM Fred Cisin via cctalk <
cctalk(a)classiccmp.org>
wrote:
the
stored
On Thu, 17 Oct 2024, Steve Lewis wrote:
tomorrow.
I follow all that (on the DEBUG.COM notes) and
appreciate the notes -
that
> will save some time, I look forward to trying a few things out
>
> I forgot to do VER explicitly, but on boot up it is saying MS-DOS
2.00.
And just now, I recalled that on github there is MS-DOS source (and
bins) -
I think Dave's Garage, he recently did a
video on building and booting
MS-DOS 4.0 from that source. Maybe I should use this as an excuse to
try
> a 2.0 build? Or least, reading through the FORMAT.ASM, I see all
> DOSVER checking stuff - helps confirm patch
addresses, or maybe try
just
recompiling that one utility without this check.
Reading through the CONFIG.txt in the MS-DOS 2.0 github repo, it's
interesting near the end: (the use of forward slash instead of
backslash,
ha! and just above this, the comments mention
/dev/<dev>)
"A typical configuration file might look like this:
BUFFERS = 10
FILES = 10
DEVICE = /bin/network.sys
BREAK = ON
SWITCHAR = -
SHELL = a:/bin/command.com a:/bin -p"
GOOD
So, you should be able to patch FORMAT 2.11 ti work on the DOS version
that is running.
BUT, whether Format /S or SYS will work remains to be seen.
I have seen cases where the opening banner does not quite match the
version number, such as 4.01 V 4.00
and a conditional jmp needs an exact match.
So, definitely run VER
and/or
in debug A(Assemble)
MOV AH,30
INT 21
INT 3 ; ends program and displays registers
and see what it shows in AX
(running that in CMD of my Windows 7 gives 0005 (5.00)!)
One of the early homework assignments when I taught PC Assembly was to
go
into DEBUG and patch LINK.EXE and EXE2BIN.EXE to
eliminate DOD version
checking.
--
Grumpy Ol' Fred cisin(a)xenosoft.com