New finds: enough Torch stuff to sink a battleship

From: Jules Richardson <julesrichardsonuk_at_yahoo.co.uk>
Date: Sat Jun 28 17:08:00 2003

> My guess is that you have normal XXX boards mounted in a non-standard
> case. Possibly the case from some other Torch product.

Yep, from some of the Torch docs I have, the case seems to be from a Torch
300-series workstation. These appear to have been the getting-more-familiar (!)
BBC core with either a Z80 or 68k Atlas board attached and optional seperate
disk cabinet. Seems like there was a stage in production where the disk ran via
a controller in the expansion cabinet from a SCSI bus whilst the floppy used a
seperate cable going to the controller on the BBC - not the 'SCSI floppy' type
setup that my XXX has.

>> wired to the back of the case though). There's also markings on the board
>> for a VME bus connection, but no socket or associated circuitry.
>
> The VME interdace is a DIN41612 plug and a few buffers. AFAIK all
> production machines had it fitted.

fair call. Don't know why mine wouldn't, even if it were a prototype - if the
circuitry's there on the card why not install the chips and sockets? Strange.

> Yes, it's analoge RGB at TV rates + 2 syncs + drive for a piezo speaker.

Ok. The Torch monitors I have all need seperate syncs - explains the RGB board
in the Torch 725 machine, which essentially seems to be a sync splitter from
the BBC's composite output.

>> labelled 'video' - but as the case is a hack who knows...
> It sounds like this case once took a BBC micro board...

yep, that fits with my 300-series theory.

> There's an 8 pin power connector on the stickleback. Is is connected to
> anything other than the PSU and the battery? The touch-switch circuitry
> is on the PSU board, and sends an interrupt over one pin of that
> connector (I can dig out schematics/pinouts if you need them).

 *wanders off to check...

Just PSU and battery. Connections as follows:

          ^- front of board

   GND o o +5V
   GND o o +12V
  BATT o o -5V
   +5V o o nc

          v- rear of board

The battery is obviously wired to ground too. 12V from the PSU also runs to the
battery via a lot of black insulation tape; I haven't looked what's under there
yet...

The PSU isn't a standard XXX PSU (which I do have one of; I see the
touch-switch connections you're talking about) - this machine has a 100W Wier
PSU in it, likely stock to the 300 series if that's where this case came from.


[Torch "Hard Disc"]
>> OK, been inside now - it is a BBC, with a Torch Z80 coprocessor. Torch SASI
>> interface hooked up to the BBC's 1MHz bus, with a Xebec interface hanging
>> off
>
> Are you sure it's SASI, not SCSI (not that it makes much difference.

Yep it is - says so on the board :) I have the manual for the Xebec interface
to which it plugs into and that's definitely SASI / ST506. There's not much to
the SASI board as it turns out - see earlier post; most of it turned out to be
comms circuitry for driving the modem and nothing to do with the drive
interface. Far as the SASI side of things is concerned, it's just some octal
buffers, latches and the like; only about ten LSxx ICs.


I had some time to mess around with the machine earlier, armed with spare BBC
for testing (which turned out to have an intermittant RAM fault so wasn't much
use :-)
Far as the Torch machine is concerned, it seems to have a power-on reset fault
such that it needs break hitting after powering up (I swapped the Torch
keyboard for a real BBC one so at least my keypresses make sense). That
shouldn't be hard to sort out.

Now, without the SASI controller plugged into the 1MHz bus, if I unplug the Z80
board I can at least get the MCP to drop into the 'no Z80' condition where it
gives an error message. I can't get that if the SASI controller is present
(whether I have any combination of Xebec board & hard drive present or not) -
just the MCP test on the screen and a flashing cursor. I'm curious why I don't
see any disk activity at all, like it's not even trying to access the hard
drive (ditto replacing that drive with a known-working unit, just with no valid
data on it). Maybe the Xebec interface is faulty...
 
[ Torch 725 ]
> Definitely not a Quad-X

nope :-(
I'm quite amazed Torch could produce so many different boxes all based around
the same BBC / Z80 / 68k choices though :-)

>> Information on what exactly a Torch Unicorn was would be useful, to be
>> honest.
>
> I thought it was a 68000 coprocessor for the Beeb.

Ahh, yet another case with the same guts then :)

>> I do have a spare hard disk labelled as Quad X, those tape drives and
>> controller boards, a surplus Torch-stock PSU, complete ROM set, and the
>> complete schematics for the QX VME card. I'll just go build myself
>> one... ;-)
>
> Good luck in finding that custom chip (oh, what did they call it?) that
> handles video, etc...

the Openchip? True. I do have the programming guide and the spec (which I
believe you said that you had), but unfortunately not the chip. I bet the guy I
got this stuff from has a whole pile of them!

[tape drives]
> Oh, it just hooks up to the SCSI bus... It's getting the software sorted
> out that's the problem.

ahh, fair enough. Not come across anything that'll help you yet though :-(

> Still, it can't be worse than the PERQ3a. On that infernal machine, the
> minor device numbers have no relation to the SCSI addresses at all!

take the ol' tial-and-error approach huh? :)

> The Manta is something I know nothing about, so if you have technical
> docs I'd be interested in probing you for information...

yep, I do - no schematics though :-)

Just skimmed the docs, and it looks like it's a pretty flexible unit supporting
up to 4 drives, which can be just about anything by the looks of it. Although
there are built-in definitions for common drives I get the impression it's all
pretty flexible so just about anything can be defined. Parameters can be set
for a given drive, including data rate (125/250/500KB per sec), number of heads
(i.e SS/DS), bytes per sector (128/256/512/1024), sectors per track (0..244),
cylinders (0..245), write precomp, step rate, head settle delay, motor on/off
delays, starting sector number etc.

There's also a parameter for setting the relationship between logical and
physical steps, for e.g. formatting a 40 track disk in an 80 track drive.

They actually look like quite nice boards... one of those in the PC could come
in handy...

cheers

Jules


________________________________________________________________________
Want to chat instantly with your online friends? Get the FREE Yahoo!
Messenger http://uk.messenger.yahoo.com/
Received on Sat Jun 28 2003 - 17:08:00 BST

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:36:09 BST