Zorro Bus Termination...

[chiark]

So, I'm trying to use both a picasso II and a Buddha on my 1000 on a simple device (BC-520CV) that plugs into an 86 pin connector and gives 3 Zorro II slots.

The devices are recognised in the early startup menu, which is nice . The buddha works on its own, which is nice.

When the picasso is bought into the mix, the machine is temperamental. Sometimes it'll boot and then hang when Picasso drivers are loaded, and sometimes it'll start giving spurious read errors from the buddha.

My thoughts are that I'm doing something that the 1000 was never really designed to do, but also that the bus isn't properly terminated.

There's quite a few threads on usenet around this form the early 90s, and bus termination seems to recommend a pull up resistor to +5v through 4k7ohm, and a connection to ground through 1k resistor and a 1000pF capacitor.

Looking at the B2000 schematics, all the lines are terminated with the resistor and cap to ground, but not pulled up to +5.

It strikes me as odd that the +5v connection is needed. It's not there in the 2000, however the signal is buffered before it gets to any cards...

Anyone ever built a zorro terminator, and anyone got any advice?

Code:
4.7K
Signal<----*--------/\/\/\/\/\-------------->5V+
|
| 1000 pf.
| 1.0K
| | |
*--------/\/\/\/\/\----| |------->Ground
| |

*NOTE: must be ceramic
And the thread... http://groups.google.com/group/comp....11fe1eabbb5c1e

[chiark]

Well, I dropped Dave Haynie a mail last night and got a reply waiting for me this morning - what a guy!

Originally Posted by Dave Haynie

There have been a couple of approaches. And it's not just on the control
signals, but at least on my designs, I had termination on the address
and data buses as well.

On the A2000... damn, it's been awhile... something like a 1K resistor
to a 220pF or 300pF capacitor to ground. The original "Zorro" expansion
board (the original bus design done at Amiga in California) used a 220
pullup/330 pulldown ("Thevenin termination".. same idea as orginally used
on SCSI) on nearly every bus signal -- not the interrupts, of course. I
tried that, but on the 2-layer A2000, I got too much crosstalk, driving
that kind of current.

So first off I'll try the Haynie termination on all lines. Shouldn't take long to make that.

As an aside, it's great that there's such an obvious enthusiasm from Dave to carry on thinking and corresponding about these things, even though we're now a generation (biological, not technical) away from them... 25 years!

[pandy71]

instead 220/330 maybe better is use active termination (also used in SCSI)

http://cds.linear.com/docs/Design%20Note/dn34.pdf

even specialized IC exist which can make this easier and less complex

http://focus.ti.com/paramsearch/docs...ODE_STRY_PGE_T

[chiark]

Thanks Pandy! I searched out the SCSI specs last night as I had some handy SCSI terminators around and was wondering if I could reuse them...

I'm going to try the approach in the B2000 first given that Dave has warned about the crosstalk, especially as the 1000 has grounding issues...

Might be looking at those ICs in the future