Date   : Sun, 08 Feb 2004 20:38:13 +0000
From   : Mike Tomlinson <mike@...>
Subject: Re: Tube ULAs

In article <C7F1DE6F-5A6C-11D8-8EFA-000A95C626D4@...>, Richard
Kilpatrick <oldcomputers@...> writes

>It booted once - only once, seemingly - when I first powered up the 
>machine. Initially, when powered up the 310 would come up like so:
>
>Acorn business computer
>
>Acorn ADFS
>
>Language?

Suggests it's lost it's configuration, if it has one (like a Master 128
with a flat battery) or there is no language ROM fitted.

>And didn't accept any keyboard input. Switching to Tube resulted in a 
>cursor. Stripping the machine of 1MHz bus devices, and doing 
>/something/ got the Tube up, but I can't remember what.

If it's based on the Master, it might want the equivalent of *co.
extube.  Alternatively, try ctrl-shift-break (it might be the ADFS
waiting for a disk drive to respond.)  Also give ctrl-space-break a go
to boot the machine with ADFS disabled.

>I desperately need proper 
>ABC/ACW DNFS, ADFS, Language (standard BBC B+) and B.Ed ROMs.

Chris at 8bs may be able to blow those into eprom for you.  Were they
missing from your example?

>I'm not used to things just breaking. There must be either a logical 
>step missing, or a failed component. My gut feeling would suggest that 
>the Tube ULA, if anything, is the most likely candidate for flaking 
>out.

It can cause the "just a cursor" symptom, yes, but then any hardware
failure preventing the 2nd processor from initialising can do that too
:(

> Is the Tube ULA a standard component regardless of CPU,

It's the same on the Acorn 6502 and Z80 co-pros at any rate.  There's
two versions:  the earlier one is a Ferranti ULA (called the 9C ULA
because the part number begins with those letters.)  This wasn't too
reliable.  The later version was made, I think, by VLSI.

> or is it 
>dedicated to the second processor concerned? I ask because the one on 
>the ACW is different.

In what way?  Chip numbers?

> I haven't checked the Z80. The ROMs on the 286 
>board - at least one of them - have labels dating from 1986 and stating 
>M512, which implies Master 512 code, but the board is assuredly 286 
>based and contains a 286 CPU (not PLCC as I mistakenly stated 
>elsewhere, but a surface-resting type held down by the heatsink).

Are you *sure* it's a 286?  The reason I ask is that the M512 used an
81086 in the same flat, pin-less chip held in place in the socket by a
metal plate and clip (the chip format is called LCC, Leadless Chip
Carrier) - though I do know the 286 came in the same form.

>What options are there for testing the Tube?

Very few, without diagnostic software - and this will need to have been
written with each processor in mind.  Apart from that, there's only a
basic "is a 2nd proc there or not?" OS call.

Mike