Date   : Thu, 19 May 2005 16:35:17 +0000
From   : Jules Richardson <julesrichardsonuk@...>
Subject: Re: Fwd: Re: RAM chips for Model B

On Thu, 2005-05-19 at 17:06 +1200, Michael Foot wrote:
> This is what I get when I shift jumper S25.
> 
> When in the North position (CAS 0 + CAS 1) I get 32K with the screen messed up.
> When in the South position (CAS 1 only) it will not boot.
> When removed (CAS 0) I get 16K and the screen is clear.

That's interesting. Having the jumper removed is invalid I think; it
should always be present in one of the two positions.

The /CAS decoding is done by IC45, an LS139 decoder IC. Jumper S25 in
the north position allows both address inputs to the decoder to
function, which results in /CAS0 and /CAS1 being driven. When S25 is
south, the most-significant address input on the decoder is grounded,
which results in only /CAS1 being driven. 

The fact that your machine works with *no* S25 jumper suggests that the
most significant address input to IC45 just happens to be floating high,
resulting in only /CAS0 being driven and a 16KB BBC - albeit one using
the 'wrong' bank of memory for a 16K machine.

So, it seems sensible that you have a memory fault in one of the eight
ICs 53 - 60, as driven by /CAS1 (the fact that I recall you saying that
the machine worked in 16K mode was what was confusing me - but actually
it's only working with the 'incorrect' 16K bank). The fact that the
machine seems to work OK when driving /CAS0, but not even booting when
driving /CAS1 does seem to indicate a memory fault rather than something
in the video chain (or associated decoding). Could still be a damaged
track / solder joint on the damaged memory bank of course.

cheers

Jules