[SI-LIST] Re: Reference plane for differential pairs
- From: steve weir <weirsp@xxxxxxxxxx>
- To: Alexandre.Gourari@xxxxxxxxxxxxxx, si-list@xxxxxxxxxxxxx
- Date: Fri, 17 Sep 2004 12:31:20 -0700
Alex,
There are additional potential problems. Without evaluating the
decoupling, it is quite possible that the +12V used for bulk power
distribution will have CM noise that stresses if not outright exceeding the
CM input range of the receivers. This could be a bit insidious because the
spectrum on the +12V that causes trouble might be rather low, and caused by
infrequent events, like board insertions and removal again depending on how
graceful the hot-swap is. Correction might require substantial changes to
the +12V filtering.
Additionally, the 12V planes without some form of termination will look a
lot like just a floating chunk of metal. That will lead to cavity
excitation by any even mode current and potentially serious EMC issues on
its own. It may or may not be OK depending on what the enclosure metal
work looks like.
Depending on how many SATA lines David has, I would be more inclined to
saw-up the +12V and provide some tightly stitched logic ground to act as
the return reference than to apply all the band-aids necessary to use the
+12V. If there are only a few SATA lines, moving the SATA lines or sawing
up the +12V shouldn't be that difficult a task. OTOH, if there are many
lines, then IMO the stack-up has been poorly selected and should be
thoroughly reviewed for better alternatives. Poor stack-ups are the sorts
of nightmares that put people in the position of calling in Mike as the
cavalry at $400. / hour to implement expensive fixes like BC.
Steve.
At 07:31 PM 9/17/2004 +0100, Gourari, Alexandre wrote:
>IMHO unless your major concern is crosstalk between whatever your
>differential pairs carry and whatever your 12V rail powers you be better off
>using 12V plane to close your diff pair common mode noise return path. It
>will always be better for EMI and pair-pair xtalk.
>The trick is you don't have much control over that for high frequency due to
>inherent interplane capacitance, but if you want to improve on that toward
>lower frequency you need to throw in some decoupling caps like Steve
>suggested. I assume your backplane stack-up is defined and you can't play
>with it.
>
>Best regards,
>
>Alex Gourari,
>
>
>-----Original Message-----
>From: steve weir [mailto:weirsp@xxxxxxxxxx]
>Sent: Friday, September 17, 2004 1:12 PM
>To: David.sanderson@xxxxxxxxxx; si-list@xxxxxxxxxxxxx
>Subject: [SI-LIST] Re: Reference plane for differential pairs
>
>
>David, since the 12V is unlikely to be tightly coupled to the logic ground
>near either the source or the load, you should avoid using it as a return
>reference. If you are really hard-up, then you might get away with adding
>LOTS of decoupling capacitors or very low inductance capacitors such as X2Y
>or IDC especially near the connectors, and switching the return reference
>from logic ground on the connector side to 12V plane on the
>backplane. But, I would still discourage this.
>
>Steve
>At 10:49 AM 9/17/2004 +0100, David Sanderson wrote:
> >I am a little confused as regards reference planes and differential pairs
> >(SATA in this case).
> >I am routing out differential signals from a controller board to a
> >backplane. The backplane has a couple of 12V planes. Can I use these as a
> >reference plane for my differential pairs? There are no (significant) high
> >frequency decoupling capacitors between 12V and 0v in the system. My
> >feeling is that, yes the 12V plane will give me the 100 ohm differential
> >impedance (assuming 50 ohm single ended to each trace), but what about the
> >return path for any current imbalance? Can it be fixed by putting down
> >some high frequency decoupling caps somewhere on the backplane, probably
> >close to the connectors at each end of the link?
> >
> >Or would it be best to avoid using the 12V plane as a reference?
> >
> >It is somewhat related to a question a day or two ago asking whether a
> >copper plane beneath differential pairs actually needs to be connected to
> >anything. I think the consensus was, theoretically no, because net current
> >should be zero, however, imbalances will always occur, so there will
> >probably be a small current flowing from/to the drivers and hence the
> >plane should be connected to ground to complete the path.
> >
> >Thanks in advance for any words of wisdom.
> >
> >David Sanderson
> >
> >
> >------------------------------------------------------------------
> >To unsubscribe from si-list:
> >si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
> >
> >or to administer your membership from a web page, go to:
> >//www.freelists.org/webpage/si-list
> >
> >For help:
> >si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
> >
> >List FAQ wiki page is located at:
> > http://si-list.org/wiki/wiki.pl?Si-List_FAQ
> >
> >List technical documents are available at:
> > http://www.si-list.org
> >
> >List archives are viewable at:
> > //www.freelists.org/archives/si-list
> >or at our remote archives:
> > http://groups.yahoo.com/group/si-list/messages
> >Old (prior to June 6, 2001) list archives are viewable at:
> > http://www.qsl.net/wb6tpu
> >
>
>
>------------------------------------------------------------------
>To unsubscribe from si-list:
>si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
>or to administer your membership from a web page, go to:
>//www.freelists.org/webpage/si-list
>
>For help:
>si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>List FAQ wiki page is located at:
> http://si-list.org/wiki/wiki.pl?Si-List_FAQ
>
>List technical documents are available at:
> http://www.si-list.org
>
>List archives are viewable at:
> //www.freelists.org/archives/si-list
>or at our remote archives:
> http://groups.yahoo.com/group/si-list/messages
>Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
>
>------------------------------------------------------------------
>To unsubscribe from si-list:
>si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
>or to administer your membership from a web page, go to:
>//www.freelists.org/webpage/si-list
>
>For help:
>si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>List FAQ wiki page is located at:
> http://si-list.org/wiki/wiki.pl?Si-List_FAQ
>
>List technical documents are available at:
> http://www.si-list.org
>
>List archives are viewable at:
> //www.freelists.org/archives/si-list
>or at our remote archives:
> http://groups.yahoo.com/group/si-list/messages
>Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
>
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List FAQ wiki page is located at:
http://si-list.org/wiki/wiki.pl?Si-List_FAQ
List technical documents are available at:
http://www.si-list.org
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
