Re: [SI-LIST] : LVDS vs. CML

About this list Date view Thread view Subject view Author view

From: D. C. Sessions (si-list@lumbercartel.com)
Date: Fri Mar 09 2001 - 07:41:33 PST


On Monday 05 March 2001 21:32, John Lipsius wrote:
# Hello,
#
# I'm interested in how one might judge one superior to the other due to
# something inherent. Or, is it board layout, path discontinuities, and
# environmental factors that are the only determinant.
#
# Assume identical 50 ohm and pcb environments, etcetera. And, the
# same rcvr type as the driver (no level shifting).
#
# For example: "assuring no skew between wires of the pair is more
# important than any factor inherent to CML vs. LVDS" might be one
# answer.
#
# 1. An initial question is: why is CML offered when there's LVDS ?
# (Just a switching capability difference?)

For a few reasons:
1) CML is trivial in bipolar, LVDS is much less so since it requires both
    sourcing and sinking current along with crossover management.
2) LVDS is tricky at best to implement in advanced CMOS processes,
   since the 2.4 V common-mode range is lethal to small-geometry
   (read high-performance) transistors.
3) LVDS has problems with crossover between receiver sections. The
    common-mode range is wide enough that both P- and N-channel
    input differential receivers are required, with the two outputs mixed.
    Obviously, this has an impact on receiver performance.
4) LVDS floats the receiver termination, which makes for high CM
    reflection coefficients.
5) CML is easily implemented in minimum geometry NMOS transistors at
    core supply potentials, which means no added power supply just to run
    the I/O.

All in all, the only advantages that LVDS has over CML are its superior
tolerance for ground shift and independence of power-supply potential.

# 3. Lastly, ANYONE KNOW OF SOME CLEAR REF. SOURCES ON CML regarding
# both ac/dc specs and these issues for real links? Apparently, there's
# no 'standard' version of CML, correct. The web appears completely dry.

There's been some interest shown in creating a JEDEC standard CML specification.

Personally, I prefer SLVS (http://www.primenet.com/~sessions/SLVS-400-pdf)

-- 
| The race is not always to the swift, nor the battle to the strong. |
| Because the slow, feeble old codgers like me cheat.                |
+--------------- D. C. Sessions <dcs@lumbercartel.com> --------------+

**** To unsubscribe from si-list or si-list-digest: send e-mail to majordomo@silab.eng.sun.com. In the BODY of message put: UNSUBSCRIBE si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP. si-list archives are accessible at http://www.qsl.net/wb6tpu ****


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Thu Jun 21 2001 - 10:11:10 PDT