🚀 go-pugleaf

About a week ago I sent a message asking for help, namely...

>I am trying to build a transputer link analyser that passively
>monitors the communications link between two transputers, and produces
>a high level (hopefully) form of debugging/performance monitoring.
>I am currently trying to find a suitable differential driver/receiver
>pair that I can use that will not introduce more than 3nS skew.

I have received a number of replies, some of which were very helpful.
I obviously did not make it clear what I was intending to do.   I am
going to use a C011 link adaptor to read in the link data, it is the
electrical interface that I am worried about.   I have had a number of
suggestions about suitable chips:-

Erik Johansson of Lawrence Livermore National Laboratory replies:-

>Our links are buffered between boards using 74LS240s (or 74LS540s).
>You could equally well use 74LS244s or 74LS541s.  We use a 100 ohm
>series resistor to provide impedance matching on the sending side, and
>the signal is carried by a twisted pair wire with one side tied to
>ground.  On the receiving end, we use a 100k ohm pull-up resistor to
>keep unconnected links idle (it is a pull-up because our logic is
>inverted between boards).  Although this scheme is not differential,
>we have not had any problems with noise, and have driven the links at
>both 10 and 20 MHz.

David Boreham of inMoS replies:-

>Try National Semiconductor parts DS8921A, DS8922A, DS8923A.
>The best parts known are made by AT&T but are not generally
>available. We have in-house link monitors and will be 
>bringing out a commrecial one in the next few months.

And finally from Lyle Bingham of Computer Systems Architects,

>You might investigate use of the National Semiconductor 26F31 and 26F32
>devices.  AMD makes a workable set of differential buffers that we
>have used for years.  These are the 26LS31 and 26LS32.
>If you can handle the speed, the optimum for skew is probably the
>26C31/32 pair from National.  Also of note is the National DS 8921 AN
>differential driver/receiver pair.

Thanks for all your help, and any further suggestions can be mailed to
me at my address below.

Dave.
    David I. Laurenson                |
    Department of Computer Science    |
    Edinburgh University King's Bldgs |  JANET   dil@uk.ac.ed.lfcs
    Mayfield Road, EDINBURGH  U.K.    |