Date   : Mon, 03 Mar 2008 23:48:09 +0100 (GMT)
From   : johan@... (Johan Heuseveldt)
Subject: TUBE chip, accessing 'Parasite' side

Hi,

On Mon 03 Mar, Pete Turnbull wrote:
> On 03/03/2008 17:14, Johan Heuseveldt wrote:

> > I refound the schematics for the internal CoPro 80186, but couldn't find
> > anything else.
> 
> Try http://www.bbcdocs.com/ and follow the link for Hardware.  The 
> diagrams for three coprocessors including the external 6502 are near the 
> bottom of the page.

I've never found that before! And I looked so much there.
I have great difficulties in accessing pages, as incoming pages freezes
very easily, although there is a significant improvement after Demon
went to XS4ALL (without changing the email address)

Direct links to somewhere deep into a site, mostly gives a direct result.
Weird!

>   With my own 6502 external 2ndProcessor, I've investigated a
> > few things. It appears that the power connections to the TUBE chip are more
> > advanced than I first assumed, which makes sense as this chip - besides the
> > logical model for the TUBE software protocols - must also handle two
> > independant power connections, and some safety precautions are apperently
> > necessary!
> 
> I expect you're looking at the circuitry around LK1 and LK3.  It also 
> has provision for power to be taken from the cheese wedge PSU or 
> alternatively from the Tube.

Actually, when powered from the wedge PSU, there's still a connection to the
TUBE chip - pin 4 - powered from the flat cable, from LK3 indeed, which is
hard wired as such (a-b). Normal VCC power - the external wedge PSU - is
wired to pin 2. GND is at pins 1 and 5.

The a-c connection would connect to the main PSU, which is the external wedge
PSU. With LK1, the +5V from the flatcable (coming from the Beeb of course)
can be connected to the boards main VCC line.

A third connection at the TUBE chip - pin 3 - is a special case, having a 12
ohms 1 Watt resistor to the main VCC (from the wedge PSU). The Z80 board has
even an extra coil in serial with that. I suppose this is vital when used
with two PSUs, as either one can be switched off!

The schematics of the internal 80186, shows pins 2 and 4 connected and wired
to the single internal VCC, leaving the special one at pin 3 open.

> > The '004' document for the TUBE doesn't say anything how fast the parasite
> > side can/may be accessed. Even an pin description is without pin numbers.
> 
> No, sadly it only says how quickly the parasite needs to respond to keep 
> up, ie the lower limit -- not what the upper limit is.  The Arm Eval or 
> 32016 circuits might give a clue; I've not looked.  You won't have a 
> problem with a 68B09, though (I know because (a) it's been done and (b) 
> that's about the same bus speed as the corresponding 2MHz 6502A or a 
> 4MHz Z80A).

Indeed my assumption is I'll get away with no slow down used. 
The notREAD and notWRITE signals are created by synchronizing the RnW with
the E clock, which can be driven constantly, just as in the Beeb, as the TUBE
chip has a separate notCS input. See 8271 and 7002 (B, B+)

Some first observations from the 3 MHz 6502 external board shows the nRD and
nWR coming from a 74 FlipFlop, so I was suspecting a slow down there. The
internal 65C102 CoPro in the Master is even at 4 MHz.

But perhaps I would be able to access the schematics mentioned above. 
If so, it could also give me the value for the timing C for the RESET
signal from the 74LS122N, which is unmarked, but looks the same as all the
decoupler C's.

Thanks for the info. After finishing this, I'll go there...


Greetings,
Johan

-- 
Johan Heuseveldt <johan@...              >
  aka  waarland

  The best place is a Riscy place
 
Acid consumes 47 times its weight in excess reality.