QBus termination?

From: Eric Smith <eric_at_brouhaha.com>
Date: Tue Feb 8 11:37:28 2005

Someone wrote:
> The CPU terminates one end, and no terminator is required at the far
> end, unless you connect chassis together. No machine I've ever seen
> has had a QBUS terminator in it (other than the termination the CPU
> provides)

Paul replied:
> That's incorrect.
>
> The PDP11 Architecture Handbook (page E-31) says:
>
> The 120 ohm LSI-11 Bus must be terminated at each end by an
> appropriate terminator. ... This is to be done as a voltage
> divider with its Thevenin equivalent equal to 120 ohms and 3.4
> nominal. This type of termination is provided by an REV11-A
> refresh/boot/terminator, BDV11-AA, KPV11-B, TEV11, or by certain
> backplanes and expansion cards.
>
> That all makes sense. DEC was very much aware of the transmission
> line nature of buses, and carefully specified all the implications.
> It's more visible in the Unibus world, because those tend to be larger
> systems with longer wires, but it applies in both.

DEC's Micronotes, written later than the PDP-11 Architecture Handbook,
explain that if you have fewer than 20 AC loads in a single backplane,
it doesn't need to be doubly terminated. This is presumably because
their analysis indicated that in that restricted case, the bus lines
were not acting sufficiently like transmission lines to require it.

Thus I think you're exaggerating a bit to say that the statement you
were replying to is "wrong". It's wrong under some conditions, and
not under others.

Eric
Received on Tue Feb 08 2005 - 11:37:28 GMT