Sharp PC-7100 - Lo-Tech XT IDE CF card upgrade

[Elladur]

Hi all,



I've had this PC-7100 in my collection from before I knew I was collecting, I found it on the way home from school in around 1998-99. Sadly a few years back I discovered that the built-in 20MB hard drive was starting to fail. I thought being an MFM/RLL style drive that it would be very tricky or near impossible to find a replacement and figured it was now a display piece rather than a running machine.

A couple of months ago I stumbled onto the XT-IDE series of boards, after a bit of research and the disassembly of the PC-7100 I discovered that the hard drive is actually a Plus brand "HardCard 20". The HardCard plugs into what looks like a standard ISA slot that is plugged into the mainboard of the PC-7100 via a ribbon cable. I purchased a Lo-Tech CF ISA adapter kit and proceeded to build it but until now haven't had a chance to try it out.

Here are a few pictures of the HardCard 20 and the Lo-Tech CF ISA card plugged into the "ISA" riser card (for lack of a better term?).




Finally I had some time that I could sit down and work on installing the card, I spent a few hours yesterday fiddling around with the machine with no luck. I got in touch with James at Lo-Tech and described my problem and after a few emails he recommended that starting a post on here would be the best approach! So here I am

After a few emails it looked like the problem was with the "ISA" slot, while it may contain signals from an ISA bus it doesn't have every signal connected through. After buzzing out which pins were connected I had the list below:

01 - GND       02 - GND
03 - D7        04 - RESET
05 - D6        06 - A01
07 - D5        08 - D4
09 - D3        10 - D2
11 - D1        12 - D0
13 - +5V       14 - +5V
15 - AEN       16 - A0
17 - IOR       18 - IOW
19 - DRQ3      20 - DACK3
21 - Note#1    22 - Note#2
23 - Note#2    24 - Note#2
25 - +12V      26 - +12V

Note#1 - A5, A8 and A9 are tied together on this pin

Note#2 - A2, A3, A4, A6, A7, and A10 through A19 are tied together on this pin and connected to GND

I can only imagine that there is some decoding logic on the mainboard coupled with the address lines configured the way they are that allow this card to behave normally.

This is the part where I throw my hands in the air and ask for help! I need to get stuck in and read up on how the addressing works for this card in a normal system and wrap my head around how this works in the PC-7100 configured the way that it is. Any help or pointers here would be great, I'm hopeful that with some jumpers to patch around the messy addressing I'll be able to get my PC-7100 back up and running


Ross

 

[Krille]

I'm definitely not an expert on the Plus Hardcard 20 but they seem to be pretty bog standard 8-bit ISA cards. Have you verified that the XT-CF board works in another computer? It seems far more likely (to me) to be a problem with the build of the board than some kind of special ISA bus.

TL;DR: You're overthinking this. Make sure the card works first!

EDIT: BTW, Welcome to the forum!

 

[pearce_jj]

Pin 21 - connected to A5/A8/A9 - is effectively a card-select signal; the card will interpret this as 320h with the other lines being low.

However, the BIOS for the card must have been included in the PC-7100 BIOS itself, since there is no way it can decode C800h nor respond to any memory access (no MEMR signal). Hence, as it stands we're out of luck.

Does the machine have an option ROM socket on the system board? We need A2 and A3 as a minimum to make the IDE logic work, and somewhere to put the ROM code as well.

 

[Elladur]

Krille: Thanks for the welcome! - Good idea on testing that the card works before going crazy with diagnostics/hacking away at things, this could be a little tricky however as I don't have many systems that I can easily test it in. I may have to dust off the old Acer XT clone that my dad has stored away.

perace_jj: Sounds like I could potentially find the required signals, isolate the pins on the ISA riser board and hook the correct signals up to get the IDE portion working and sort the ROM out some other way.

Alternatively I could make my own riser card and run all of the signals required, so long as I can locate them.

There are a few options for pinching the required signals, what looks like a co-processor socket (a quick Google may have ruled this out, it has some but not all of the required signals on it...) and also three expansion connectors (1x 30pin for an optional modem, 1x 30pin for an option video card and 1x 44pin marked as a generic expansion port). I'll have a bit of a poke around on the board this weekend and see what I can find out RE: pinouts. Worst case I could make a pass-thru socket for the CPU and pinch the signals directly from there?

There are two ROM sockets but both are populated, 27128A (128kbit/16kB) devices. If the old drive controller firmware is living on the mainboard I'd imagine I would need to devise a way to disable the onboard ROM to allow the XT IDE ROM to be accessed?

If the old drive controller ROM is living on the mainboard, would they have lifted it directly from the ROM on the drive controller? I guess I could try reading the ROM's in and looking for common data in the location that hopefully corresponds to where C800h is? If there is enough room I might be able to duplicate the contents of the mainboard ROM and jam the XT IDE ROM into the correct location and cross my fingers...

Thanks for your help so far guys


Ross

 

[pearce_jj]

The expansion interface is the best bet - it's referenced here has having supported three slots (presumable ISA).

The ROM is a puzzle. I'd suggest having a look at C800 area with DEBUG (dump it to a file then open with a hex editor, for example) to see if the option ROM is even at the standard address. Possibly one of those ROM chips is configured there, or the code could conceivably have been included in the main BIOS. What happens if the controller is physically removed, does the machine produce some HDD related errors?

 

[Elladur]

Nice find! Thanks for that, the expansion slot does indeed sound promising. Funnily enough I had one of those Diconix inkjet printers shown in one of the ads, sadly I threw it out many years ago

With the original drive removed the HDD activity light blinks for a second and then the computer tries to boot from the floppy drive (which also has a few issues at the moment...). No errors relating to the drive/controller.

 

[FuZX]

Years later...
Also have a 7100 with a dead +card
Was there a final conclusion to this quest for the magic CF card ?? ;-)

 

[4xtx]

The expansion interface is the best bet - it's referenced here has having supported three slots (presumable ISA).

The ROM is a puzzle. I'd suggest having a look at C800 area with DEBUG (dump it to a file then open with a hex editor, for example) to see if the option ROM is even at the standard address. Possibly one of those ROM chips is configured there, or the code could conceivably have been included in the main BIOS. What happens if the controller is physically removed, does the machine produce some HDD related errors?

8 years later I can answer this

No - it does not.
I have a PC-7100 which is supposed to have a hard-card and interface, it was removed by the previous owner and is simply ignored.