Help fixing Vector-06c

[cyb2600]

Thank you! I think I see in the schematic where to grab those signals, I'll test.

- - - Добавлено - - -

Are you sure that the bootloader has a built-in RAM test? Never heard of this before...

Still, the bootloader uses the upper RAM for its stack and variables (0xe000-0xffff). Coincidentally, this is also the upper bank of the video frame buffer - the only one used by the bootloader in order to reserve all RAM below 0xe000 for loading programs (hence the pixel "garbage" in the upper right corner of the bootloader grid screen)

I attached 3 data probes to a13-a15 and one to wr to see if the computer was trying to access the upper RAM (0xe000-0xffff).


At around +1000ms I hit the wwod and blk keys. It looks does two bursts of reads and writes to the 0xe000-0xffff region at around +1400 and +1700. Then a14 goes low and will stay low indefinitely.

At around +2700ms I hit blk and sbr, and it looks like it's trying to read that upper block again.

[Advertiser]

[cyb2600]

I also have these manuals and tapes. They appear to be BASIC and Assembly manuals.

[KTSerg]

Are you sure that the bootloader has a built-in RAM test? Never heard of this before...

Это самое начало штатного заводского загрузчика.

Скрытый текст

...
        LXI     SP,0DCF0H
        LXI     D,55AAH
M000EH: PUSH    D
        POP     H
        MOV     A,L
        CMA
        CMP     H
        JNZ     M000EH
...

[свернуть]

- - - Добавлено - - -

I cleaned all three areas up with rubbing alcohol until they were shiny, and I verified none of the traces were broken from corrosion. Unfortunately it didn't make any difference.

I didn't talk about corrosion. The conductors run next to the mounting holes. They are often crushed, and neighboring ones are connected to each other.

[cyb2600]

Ah, I misunderstood. I checked the pins, none of them were bent over or touching traces.

[x-code]

I checked the pins, none of them were bent over or touching traces.

What KTSerg rather means is that due to excessive mechanical load on the motherboard's mounting holes, the traces themselves could have been deformed up to a point where they short-circuit somewhere near a hole. So I'd suggest you carefully inspected traces near each of the mounting holes and made sure they have not been deformed / damaged.

[cyb2600]

I see. No, I don't see any deformed traces. I looked the board over on both sides, the only unusual thing I saw was a cut trace on the top side. It looks like it was deliberately cut, though, not an accident.

I disassembled the Vector BIOS that came with a Vector emulator. It looks like after it initializes some ports, the next thing it does is write zero's to all the RAM. I hooked up the logic analyzer to CAS, RAS, and WE as well as 5 of the data lines. It looks like it is indeed writing 0's to the RAM, and the RAS and CAS timing looks good.



After it's finished zeroing out the RAM, it pushes 55 and AA on the stack. Then it pops them back off, does a 1's complement of 55 (which is AA) and compares the two. If they don't match it loops and tries again. I guess this is a memory test. If the memory is bad, at least where the stack pointer is pointing, it never proceeds.

[KTSerg]

..
After it's finished zeroing out the RAM, it pushes 55 and AA on the stack. Then it pops them back off, does a 1's complement of 55 (which is AA) and compares the two. If they don't match it loops and tries again. I guess this is a memory test. If the memory is bad, at least where the stack pointer is pointing, it never proceeds.

This is the memory test I already mentioned.
Only usually it goes before clearing memory, and not after.

Do you have enough knowledge and experience to replace the ROM chip containing the bootloader?
Do you have the ability to write a new ROM chip with the desired program?

To begin with, you can write a memory test to the new ROM.

Instructions for replacing the ROM. But it may not match a little, since it is for replacing the К556РТ5 microcircuit, and you most likely have К556ТР7.
http://www.sensi.org/scalar/ware/543/

[KTSerg]

There is another way.
You can try swapping the memory chips.
To do this, you need to cut two wires. The sections are marked in yellow.
And solder the two wires as marked in red.
At the same time, the microcircuits will swap and if the replacement ones are in good order, then the computer can work.
And then you will need to run a memory test, find and replace defective microcircuits.

https://s8.hostingkartinok.com/uploa...c86deb5365.jpg

[cyb2600]

With the logic analyzer I could see it writing 01010101 and 10101010 to bank 3 of the RAM, but it was reading back 01011101 10101010 - bit 5 is stuck high, that RAM chip must be bad. At least this confirms the read/write circuit works otherwise. I'm going to try to replace the fault RAM with a 4164 with A7 tied to ground.

Given that this is the third bad chip I've found in this computer I'm expecting there to be more problems. I have a Chinese USB EPROM programmer but I don't think it can burn this type of PROM. I have a very old Stag PP39 programmer that might be able to do it. I'd have to find a blank chip, too. Otherwise I'm sure some sort of adapter could be made.

[KTSerg]

With the logic analyzer I could see it writing 01010101 and 10101010 to bank 3 of the RAM, but it was reading back 01011101 10101010 - bit 5 is stuck high, that RAM chip must be bad. ...

According to the D67 scheme?

... I have a Chinese USB EPROM programmer but I don't think it can burn this type of PROM. I have a very old Stag PP39 programmer that might be able to do it. I'd have to find a blank chip, too. Otherwise I'm sure some sort of adapter could be made.

If we are talking about replacing and programming ROM, then any eeprom will do. I bet w27c512.