Date : Mon, 27 Jul 2009 23:33:56 +0200 (BST)
From : johan@... (Johan Heuseveldt)
Subject: econet bridge protocol
Hi,
Update...
On Wed 15 Jul, Johan Heuseveldt wrote:
> On Mon 06 Jul, Phil Blundell wrote:
> > On Mon, 2009-07-06 at 00:45 +0200, Johan Heuseveldt wrote:
>
> [snip]
> > In particular, if you connect the three bridges in a chain, something
> > like:
> >
> > net 1
> > |
> > Acorn bridge
> > |
> > net 2
> > |
> > SJ bridge
> > |
> > net 3
> > |
> > Acorn bridge
> > |
> > net 4
> >
> > and then capture a packet trace on net #3 while power-cycling the bridge
> > between 1 and 2, that might be useful. More or less any network monitor
> > is probably good enough for this; I mostly use either NETMON on a BBC,
> > or my own RISC OS monitor on an A5000. Bridge packets tend just to be
> > broadcasts of a couple of bytes so you don't need any sophisticated
> > monitoring technology :-)
>
> I had started to create a list for the stuff needed. Good greeve, that's
> quite a lot! :-)
And still being inaccurate, as I've found when setting it all up!
> This week I realise I forgot I also need two extra clock boxes;
This was the first error; 4 nets mean three bridges, but indeed
also 4 clocks; not three! For the sake of the test it could be
important that every net is clocked. Otherwise, the bridge is
probably not acting as we would like!
The SJ bridge has its own, so still needing 3 seperate clock boxes.
One Acorn Iss2 was already in use. Another one needed some digging.
It's a set with clock box, 5 connection boxes with two 5-pin DIN
connectors - two of them being terminators with a single 5-pin DIN
connector. A mains adaptor and a wire insertion tool completes the
set. Most likely very familiar to you all.
A while back, Econet termination was discussed. In case there is
still a need for a schematic, here it is for issue 2; passive
termination - without power supply, which is quite different as
the one in the Advanced User Guide.
<http://www.waarland.demon.nl/Econet/EcoTerm.jpg>
> [...] was a SJ Research clock box. But that seems to be dead, and
> so far I haven't been able to fix it.
Problem is at least the power supply. The primary side of the
transformer is broken. The lead between the PSU and clock box is
fixed; no standard connector here. With some luck the actual clock
box is ok!
> Another problem a space, well, the lack of it.
For the moment I manage! :-)
> I do hope next week I can give it a go, with the setup you gave.
> Hopefully I'll get somewhere.
Another clock box was present at the table. It's a metal case,
and this is the one which in the Adv. User Guide. There is
only a single 5-pin DIN connector here! Parts for termination
are not present.
A good PSU was recently bought, so I checked it out. The selector
is set at 7.5 volts, although 9 and 12 also did the job, though
with a brighter red LED.
I think for long nets, you need 12 or 15 volts?
More concern was one of the Acorn Bridges; It was dead. It turns
out the line drivers weren't functioning. I tried the drivers of
the good bridge, and then I have a good bridge! Phew!
I noticed that the good bridge had two differnet drivers; the
AM26LS30 and an AM26LS34. When I asked around, someone responded,
having the same two different drivers in a bridge as well. It's
too far to collect for borrowing.
I know there's another person with a bridge, but he's living even
further away.
I though I had some spares, but I couldn't found them, so most
likely I'm wrong here.
Checking the Farnell catalogue, I couldn't find it. Checking a
paper based datasheet - down loaded from www.datasheetcatalog.com -
it was stated the AM26LS30 is also a direct replacement for the
National Semiconductor DS3691, and that one was in the Farnell
catalogue! It's however in a SOIC package. I will order a couple,
assuming it must be possible to solder these on DIP sockets.
I realised I was short on Y-cables and terminators. Two days of
searching in town, resulted in disappointment only. And the small
local shop where I got a few of them with ease, no longer exists.
So I'll need to order several parts as well to make my own Y-cables
and terminators.
I have more than enough drop leads. Already tested and ok.
Last week I did some checking on station connections. Sometimes
termination can be a problem, and a real terminator should be
used. Sometimes a Beeb is sufficient to fulfill this need. I have
the two SJ terminator plugs normally in use. Digging out the 2nd
Acorn clock box gives me two boxed versions as well. To be on the
safe side, I'll need a few more; more parts to order.
Net 1 is terminated on one side with a RiscPC. If not terminated
properly on the other side, it has always complained. With the older
- issue 1 - Acorn clock box on this net, is it now quite happy with
a Beeb as terminator on that other end! So, no dedicated terminators
on net 1!
Net 2 is the normal net, normally in use. Currently still the
SJ Research MDFS file server attached. For the actual test, this
can be removed, as well as the Master on this net. Only the Master
has termination on this net end, via an Y-cable. The other net end is
directly put in th MDFS server. I'm quite surprised.
Net 3 is also fully operational, although very small currently!
It has an A400 hooked up, and no terminators in use. One end ends
in the A400, the other ends in the SJ bridge.
Three nets operational, and 1 single terminator in use.
So, I'm almost there!
<to-be-continued>
Greetings,
Johan
--
Johan Heuseveldt <johan@... >
aka waarland
The best place is a Riscy place
In a fight between you and the world, back the world.
Franz Kafka