New QL user & owner

[M68008]

Great find MetroZodiac!! And thanks for the dump Chr$!

It's a pity that the high ROM is missing, but still quite interesting! The ROMs contain SuperBASIC, the arithmetic routines and an OS skeleton that typically ends calling the high ROM, where much of the TRAP code and device drivers seem to reside.
There are a number of SB keywords still missing compared to later ROMs (Turtle graphics, draw relative, date, SB editing, MODE, SB arrays, VER$ and a few others), and two extra ones (TUTOR and BANG).
TRAP#1 D0=0 is there and it would return version '1.00' if called.

[M68008]

Unless it's FB, for which the last 3rd is available.

Where is it available?

[Chr$]

The last third of FB is in this zip file, originally share by Urs.

[Chr$]

An interesting discovery with this ZX-83 Issue 2 that I have just investigated.

The 'cockroach' upside-down chip stuck on IC17 has nothing to do with the ROM/EPROM side of things. The chip is a 74LS04. It obviously has +5v and GND. The 2 other connections to the PCB are pin 1 goes to CPU pin 34 (and ZX8301 pin 7) and pin 8 goes to ZX8302 pin 25.

The output pin 2 of the 74LS04 is connected to pin 3, another inverter. It's output is connected to pin 5, another inverter. And it's output is connected to pin 9, another inverter, the output of which (pin 8) is connected to the ZX8302 pin 25.

So what's that all about? Signal goes in, get inverted 4 times and then comes out the same way it went in. Some kind of delay bodge?

[tofro]

So what's that all about? Signal goes in, get inverted 4 times and then comes out the same way it went in. Some kind of delay bodge?

Yep, that's why such stuff is typically done (or, signal conditioning when a digital signal is smeared ll over the place). The clock for the ZX8302 apparently is delayed by a bit compared to the CPU clock.

[Chr$]

Yep, that's why such stuff is typically done (or, signal conditioning when a digital signal is smeared ll over the place). The clock for the ZX8302 apparently is delayed by a bit compared to the CPU clock.

Thanks tofro. Chuffed I guessed pretty much correctly.

In other news: After some poking around just now before we go down for Christmas (eve) dinner I discovered that pin 27 on both the EPROM sockets is tied to 5v, so that rules out the use of a 27256 at least as a 32k chip.... so I used a couple anyway, but as 27128 by only writing half of them. For testing I used JM ROM, split it into 3 parts, 2 fitted internally and I utilised my original dongle for the EPROM containing the 3rd part.

At first all I got was coloured dashes zipping across the screen so I dug out my stash of spare chips and went straight for the ZX8301. With that swapped, I got the tweed, then the normal F1/F2 screen prompt, it works! That's a relief. What a survivor, I could possibly claim it to be the oldest working QL.

Just need to find the original ROM 3rd now, which is going to be tough.

[M68008]

Was the empty one perhaps being used for debugging code at $C000?

Yes, it would make sense for the ROM there to be for debugging: The TUTOR keyword disables interrupts and jumps to $C000, and unused CPU exception vectors also jump to the $Cxxx area.
Btw, the BANG keyword seems to be for testing the address exception.

Code: Select all

TUTOR
00003382 4E40          trap #0
00003384 46FC2700      move #$2700,sr
00003388 4EF90000C000  jmp $c000.L
BANG
0000338E 327C0001      movea.w #$0001,a1
00003392 4A51          tst.w (a1)
00003394 7000          moveq #0,d0
00003396 4E75          rts

[Chr$]

Thanks Daniele,

Here's another picture from the ebay listing for the dongles. It shows a 'TUTOR' ROM marked C000, so that makes sense, whatever Tutor was/is.

So the question is, can this be made to work to any degree without the original 8000 ROM? Lacking some features, but some kind of stable environment?

[QLvsJAGUAR]

Chris, can you tell us the story of the purchase? Did you travel from Germany to England (Wolverhampton?) to meet the seller and collect the machine. Any anecdotes?

Well as you asked! ...

That's what I call a story!

[jobdone]

With the upper 16K missing the RAM test should still work as this lies within the first 16K ROM. Given the fact that the test pattern doesn't appear, it looks like there is another hardware failure.

I can confirm this does produce a pattern in an emulator.