LVDS CM Termination

bbrown!hpx400!HAYNES_DENNIS/error_bbc////////HPMEXT1/DENNIS#b#HAYNES#o#TUC#o#06%u2 at globe.indirect.com bbrown!hpx400!HAYNES_DENNIS/error_bbc////////HPMEXT1/DENNIS#b#HAYNES#o#TUC#o#06%u2 at globe.indirect.com
Wed Oct 18 18:29:00 PDT 1995


O.K.  How about helping out a silicon
designer with a bus transmission line
question?

In the last Low Voltage Differential
SCSI technical working group meeting
we discussed the issue of a COMMON MODE
termination requirement for the LVDS bus.

My simplistic, first order, semi-intuitive,
first-order-arm-waving thought process
led me to reason that it would be best to
have the COMMON MODE termination impedance
of the differential bus equalto zero.

However, what came out of the meeting was
that the bus should be terminated with
about 150 ohms (150 ohms +/- 20% was the
proposed spec) of additional common mode
resistance.

Here is the picture that is worth 1K words.

(Note: for simplicity, the "bias
generators" were intentionally left out. )


                                60 ohms
                        |---------/\/\---o
            Rcm         |
       |-----/\/\-------|
       |     150 ohms   |       60 ohms
       |                |---------/\/\---o
       |
    ------
      --    Vcm
    ------  1.25V
      --
       |
       |
     ----
    //////


So....how important is Rcm. Why does it
need to be there?  What happens if it goes to
zero?  Could you use a terminator without it?
Would you buy a terminator from a man who
draws circuit diagrams with ASCII
characters?

Any comments would be appreciated!

Dennis Haynes
Burr-Brown Corp.
Email: haynes_dennis at bbrown.com
Phone: 520 746 7262






More information about the T10 mailing list