Re: [SI-LIST] : Other ways of transmitting differential signalling besides edgecoupled traces or broad side coupled traces

About this list Date view Thread view Subject view Author view

From: Bob Lewandowski (blewandowski@Vixel.com)
Date: Thu Feb 17 2000 - 18:56:47 PST


Salvador Aguinaga wrote:

> ---------------------- Forwarded by Salvador Aguinaga/MW/US/3Com on 02/17/2000
> 06:31 PM ---------------------------
>
> From: Salvador Aguinaga/MW/US/3Com on 02/17/2000 01:09 PM
>
> Sent by: Salvador Aguinaga - eNetworks HW Engineer, Carrier Systems Group
>
> To: "'si-list @silab.eng.sun.com'" <si-list@silab.eng.sun.com>
> cc:
> Subject: Other ways of transmitting differential signalling besides edge
> coupled traces or broad side coupled traces
>
> All,
>
> I would like to know you thoughts on 2 routing structures/geometries for
> differential signalling (LVPECL)
>
> 1. Routing differential pairs not coupled to each other. Can differential
> signalling be successfully transmitted over two independent 50 ohm traces?
>

You can run 100 ohm differential as 2 - 50 ohm uncoupled transmission lines. However,
if the lines are not coupled then the integrity of the grounds on both lines is
critical because the signals are really being sent single endedly. Also, the delays
of both signal paths must be matched exactly. This is NOT trivial on a board layout,
particularly if you are dealing with fast edge signals (<< 1nS). If lines are very
tightly coupled then a common ground is not necessary, witnessed by unshielded twisted
pair cables. The real problems occur when you have significant coupling between the
lines and a ground present, as is the case in most circuit boards. Then there is
significant interaction between single ended and differential transmission. This is
due to differing amounts of the return current flowing in the ground structure or in
the opposite phase conductor.

>
> 2. Routing differential pairs broadside
>
> _______________ plane
> core
> ______ signal + (50 ohm referenced to the top plane)
> prepreg
> _______________ plane
> core
> ______ signal - (50 ohm referenced to the top plane)
> prepreg
> _______________ plane
>
> Can this be done for differential signaling ("broadside") ? Are there other
> methods of transmitting differential signals over traces that are not
> Edge-coupled or Broadside coupled?
>

Any method that brings the two conductors into close enough proximity that there is
both electrical and magnetic coupling between the conductors will work for
differential transmission. Then the impedance to the difference signal is defined and
controlled propagation can take place. Broadside coupling works better for tightly
coupled lines, because the geometry for close coupling is easier to achieve, i.e.,
closer proximity yields higher capacitance as well as higher mutual inductance per
length. Edge coupling does not allow for very tightly coupled lines, because the
capacitance and mutual inductance between lines is lower. For the same impedance
level ground planes want to be farther away from broadside coupled lines than edge
coupled lines.

Eric Bogatin has some good presentations on differential transmission on his web site
at:

 http://www.bogatinenterprises.com/

> --
>
> --------------------------------------------------------------------------------
>
> žal

---Bob Lewandowski
    Sr. Principal Engr.
    Vixel Corp.

**** 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 Apr 20 2000 - 11:35:06 PDT