Date   : Thu, 02 Jun 2011 19:23:33 +0100 (WET-DST)
From   : BBCMICRO@... (Peter Coghlan)
Subject: Model B with video corruption

>
>My theory is there is a fault around IC 23, 24, 26, which is causing data bus
>conflicts between chips. Either that or a faulty chip somewhere.
>

According to my circuit diagram, IC23 is an eight input NAND gate which
(I think) is responsible for slowing the clock to 1MHz when particular
peripherals are addressed and IC24 and IC26 are address decoders. I doubt if
IC23 is responsible. I suppose something could be selecting more than one i/o
device at the same time. However I think it would be easier to try to
eliminate the actual devices themselves which could have faults of their own
as well as causing problems if the get selected when they shouldn't.

I would suggest pulling all non-essential i/o chips such as the ADC, the user
VIA. the serial ULA (assuming not soldered) and any extra paged ROMs. As you've
noticed, removing an 8271 or 68B54 requires further steps to be taken to avoid
hanging the machine.

If the corruption is still present, I would suggest trying to figure out what
could cause the particular corruption being experienced. For example, if
incorrect characters are appearing in MODE 7, compare the ascii values of the
characters present with what is supposed to be there. It may be that the
problem can be narrowed down to a specific bit or bits which may in turn point 
to a problem with particular memory chips or bus tranceivers.

Does the machine respond to commands? Is the contents of the screen memory
wrong or is it only what is on the screen that is wrong? If the screen memory
is corrupt, is non-screen memory corrupt too? Is there corruption in all
screen modes? Any random noises from the speaker?

Regards,
Peter Coghlan.