Date   : Sat, 29 Mar 2008 22:30:13 -0000
From   : ken.lowe@... (Ken Lowe)
Subject: Reviving a dead BBC+

Mark McDougall wrote:

> You need to see if the CRTC CS# line pulses *at all* - if not then that 
> would explain why it's not being programmed and the H/V sync lines don't 
> pulse. I'm not familiar with the circuit so I have no idea what's driving 
> the CRTC CS# pin..

I think you might be onto something here. I believe the &FExx decoding 
circuit that drives CRTC is not functioning correcty. With a lot of track 
tracing, reference to the beeb (not B+) wiring schematic and reference to 
the B+ service manual, it looks like the circuit that drives CRTC notCS on 
the B+ is functionally identical to the circuit on the original B; albeit 
slightly different pins are being used on the chips.

It appears that address lines A9 and A10 are being held permenantly low, and 
this in turn keeps the CRTC line (and a number of other lines like vidproc) 
held high. I've carried out some checks to see what could be pulling A9 & 
A10 low. From what I can tell the only components that use A9 & A10 are the 
ROMS, memory RAS / CAS controllers, the &FExx decoder (which is used to 
drive the CRTC circuit) & the 6512 CPU.

I don't think the fault lies with the &FExx address decoding circuit, 
because A9 & A10 are actually split across two ICs (IC22 and IC21 on the B+ 
or IC22 and IC24 on the original Beeb). For both address lines to be low, I 
would need to have a fault with two separate ICs.

I've removed all the ROMS (including OS) to see what happens to these 
address lines. They still remain low, even though some of the other address 
lines show some signs of activity (not too sure how this can happen if the 
OS has been removed).

Finally, I snipped the A9 & A10 lines to the memory RAS / CAS contoller IC51 
but that made no difference (now I need to get a replacement 74LS257AN).

As far as I can tell now, the only component left is the 6512 CPU. I need to 
try and find a replacement from somewhere.

I'm also a bit puzzled about the arrangement of the CRTC PCB track on the 
B+. I believe pin 4 of IC39 (&FExx decode circuitry) should be connected 
directly to pin 25 of IC78 (6845). This isn't the case. I've checked the 
equivalent circuit on the original BBC and there is a direct connection 
between IC26 pin 12 (&FExx decode circuitry) and pin 25 of IC IC2 (6845). I 
suspect there may be a broken track or bad via on the B+ board. Unforunately 
it must be underneath one of the soldered ICs making it very difficult to 
track down. In the interim, I've temporarily installed a wire jumper but 
again this has made no difference.

Regards
Ken.