RE: [SI-LIST] : AGND & DGND stitching

About this list Date view Thread view Subject view Author view

From: Charles Grasso ([email protected])
Date: Wed May 16 2001 - 09:45:06 PDT


Ken et al,

Most everyone involved in the control of those insidious
residual flux ( or...emi!) has used AC to DC shorts as
part of a properly integrated high speed design.
I have found that the burden in cost and time generally
prohibts 100 "flux control" in a pwb so a common-mode
reference and appropriate grounding is usually a good idea.

Charles Grasso
Ansoft Corp

-----Original Message-----
From: [email protected]
[mailto:[email protected]]On Behalf Of Ken Cantrell
Sent: Wednesday, May 16, 2001 9:52 AM
To: Patel, Bhavesh; SI_LIST (E-mail)
Subject: RE: [SI-LIST] : AGND & DGND stitching

Bhavesh,
Not sure if this will solve your problem, but I always use a stitch
separation distance of lambda/20 of the fastest edge rate. This is from
either one of Montrose's books, and of course, Ott's seminal work before
that. The "bigger" picture is whether or not this is required. Opposing
views state that if you design your lines/boards correctly, you won't have
the residual coupling flux to begin with.
I don't use single point grounds either. I guess the idea was to prevent
low-frequency AC V&I from leaking into the DC frame and producing skew in
the oscillator and noise on the DC power frame. The further away you are
the more inductance, the more inductance the more the noise current time
rate of change is slowed down thereby reducing Vcm. I don't subscribe to
that theory. I think that might have been feasible at very low frequencies,
but doesn't apply to today's technology. Multi-point provides the lowest
inductance path to ground, and has worked best for me so far. On board
physical separation, and even split ground planes are sometimes helpful
especially in high current applications.
Ken

-----Original Message-----
From: [email protected]
[mailto:[email protected]]On Behalf Of Patel, Bhavesh
Sent: Tuesday, May 15, 2001 7:41 PM
To: SI_LIST (E-mail)
Subject: [SI-LIST] : AGND & DGND stitching

Hi! I needed some feedback regarding the concept of stitching AGND and DGND
planes.
>From some reference material that I have read:
1) AGND & DGND should be stitched at one point, this way you ensure the
return point and do not create ground loops.
Now, on a board that I have the designer stitched agnd and dgnd with
shorting bars all over the place with no constraint i.e. number and location
on top layer and bottom layer. Later, when the board was being debugged some
of the chips that had both agnd and dgnd pins were seeing a huge potential
difference which made some sections of the IC not to function properly.
My take on this is that the multiple stitching points are giving rise to
weird ground loops which create this potential difference.
What is the best approach for connecting AGND and DGND and should the
highest operating frequency on the board should play a role in the number of
stitches .
Thanks in advance
Bhavesh Patel

**** To unsubscribe from si-list or si-list-digest: send e-mail to
[email protected]. 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
****

**** To unsubscribe from si-list or si-list-digest: send e-mail to
[email protected]. 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
****

**** To unsubscribe from si-list or si-list-digest: send e-mail to
[email protected]. 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 - 10:11:59 PDT