[SI-LIST] Re: plane-to-plane decoupling

  • From: steve weir <weirsi@xxxxxxxxxx>
  • To: Richard Jungert <r_jungert@xxxxxxxxxxx>
  • Date: Mon, 15 Sep 2008 23:34:42 -0700

Richard, I'm not advocating anything that is either expensive or can't 
be bought right off the shelf.. 

A typical discrete PLL is both a low power part and has very limited I/O 
so sticking it on a peninsula so that tranverse currents don't cause I*Z 
drops  that upset the VCO is usually a freebie.   I think that arguments 
against splits come mostly from rote practice.  Used properly there is 
nothing wrong with them.

Steve.
Richard Jungert wrote:
> Steve.
>
> I agree with you somewhat except for getting into exotic capacitors 
> that have long lead times and such. I remember well a 45mb/s  DS3 
> design we did years ago where this concept fixed the jitter problem 
> real well.  It actually was real easy to Cluge the regulator in there 
> just to see if it would work. And low and behold.. yes it did.  One 
> has to be carefull thinking that a cheaper capacitor is the solution 
> from a pure parts cost standpoint but watch out . that lead time can 
> bit you especially if they have to build capacitors in huge lot sizes 
> with long lead times.
>
> Splitting the ground/power plane and matching them is also part of the 
> solution under the PLL.  If you do this there is almost no jitter. One 
> of my first projects out of school was to fix a nasty PLL problem 
> where there was too much phase noise. I ended up ( after 3 months of 
> experiments ) isolating the ground and connecting the common ground 
> plane to the small PLL Plane in just one place to fix the problem.  
> Some engineers will argue that split planes is not the solution but I 
> have to disagree.
>
> Just wanted  to share some real life experiences.
>
> Richard
>
>
>
> > Date: Mon, 15 Sep 2008 14:09:05 -0700
> > From: weirsi@xxxxxxxxxx
> > To: r_jungert@xxxxxxxxxxx
> > CC: mrose@xxxxxxxxxxxx; si-list@xxxxxxxxxxxxx
> > Subject: Re: [SI-LIST] Re: plane-to-plane decoupling
> >
> > Richard, in my experience, linear regulators can be effective for
> > cleaning up PLL power on the PCB but they are usually completely
> > unnecessary. The bandwidth of even a good linear regulator by itself is
> > still in the audio range. If your PLL is insensitive to audio frequency
> > disturbances by virtue of a good loop, then what you are really buying
> > with the linear regulator is a very expensive series L/R that works 
> with
> > the bypass caps on the PLL side to filter noise from the dirty digital
> > realm. With a little care you can usually save a lot of: board
> > real-estate, money, and complication by designing a passive filter 
> to do
> > the same or better job.
> >
> > Steve.
> >
> > Richard Jungert wrote:
> > > Mike.
> > >
> > > I would recommend putting the PLL circuit near the edge of the 
> board and also cut a power and ground plane split 270 degrees around 
> or under the PLL circuit. Cut them both the same shape. If you put it 
> near the center of the board there is a very high probablility for 
> much more jitter.
> > >
> > > Also, in my experience its not a great idea to power the PLL with 
> a common voltage. Put a small low power voltage regulator in to just 
> power the PLL circuit and this will isolate him from the rest of the 
> noisy digital logic. Power and ground plane noise will sneak right 
> into the PLL and modulate your clock.
> > >
> > > I have found on three other designs that this approach is very 
> effective in cutting down phase noise on the clock circuits.
> > >
> > > Another trick that cuts noise on ground/power planes is decoupling 
> with caps out to the chassis ground. Decouple out to the chassis 
> ground and watch the noise decrease. This is also a trick used to 
> supress EMC problems.
> > >
> > > Richard Jungert
> > >
> > >
> > >
> > >
> > >> Subject: [SI-LIST] Re: plane-to-plane decoupling
> > >> Date: Mon, 15 Sep 2008 12:14:31 -0500
> > >> From: zabinski.patrick@xxxxxxxx
> > >> To: mrose@xxxxxxxxxxxx; si-list@xxxxxxxxxxxxx
> > >>
> > >> Mike,
> > >>
> > >> In answer to your last question, I believe the lowest-inductance 
> option
> > >> involves capacitors placed directly between the two power planes (and
> > >> not through the ground plane).
> > >>
> > >> That said, the scenario scares me.
> > >>
> > >> For many components that require multiple planes (e.g., FPGAs w/
> > >> SerDes), some of their supplies are very sensitive to noise. For
> > >> example, nearly all SerDes cores have an analog supply for 
> phase-locked
> > >> loops (PLLs), and noise injected into them can severely degrade 
> jitter
> > >> performance. Similarly, noise on the I/O supply of a parallel bus can
> > >> degrade SI of the output waveform. In many instances, the chip 
> supplier
> > >> provides guidelines for isolating such voltages, often recommending
> > >> specific isolation circuits. Adding decoupling between the various
> > >> voltages will/can defeat the isolation circuits and inject noise from
> > >> one plane onto the next.
> > >>
> > >> Although adding decoupling between the power planes will help with
> > >> discontinuities of the signals traversing the split, you could be
> > >> inadvertently creating other problems. As such, I suggest you look
> > >> closely at the power planes that you're about to inject noise 
> into and
> > >> ensure they are immune to outside influence.
> > >>
> > >> Good luck,
> > >> Pat Zabinski
> > >> Mayo Clinic
> > >>
> > >>
> > >>
> > >>> Some diff pairs on L4 will cross power plane splits (actual 
> different
> > >>> power sources and loads) and I wanted to provide an effective AC 
> path
> > >>> for any common-mode return currents. I was thinking about 
> placing some
> > >>> nearby decoupling caps from plane-to-plane across the split. Do you
> > >>> think it would be better to decouple from plane-to-ground on
> > >>> both sides
> > >>> to steer the current through the L6 ground layer? L5 and L6
> > >>> are already
> > >>> coupled through the inter-plane capacitance (they're about
> > >>> 4mils apart).
> > >>> Which will provide a lower inductance path?
> > >>>
> > >>
> > >> ------------------------------------------------------------------
> > >> 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 technical documents are available at:
> > >> http://www.si-list.net
> > >>
> > >> 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
> > >>
> > >>
> > >>
> > >
> > > _________________________________________________________________
> > > Get more out of the Web. Learn 10 hidden secrets of Windows Live.
> > > 
> http://windowslive.com/connect/post/jamiethomson.spaces.live.com-Blog-cns!550F681DAD532637!5295.entry?ocid=TXT_TAGLM_WL_domore_092008
> > > ------------------------------------------------------------------
> > > 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 technical documents are available at:
> > > http://www.si-list.net
> > >
> > > 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
> > >
> > >
> > >
> > >
> >
> >
> > --
> > Steve Weir
> > Teraspeed Consulting Group LLC
> > 121 North River Drive
> > Narragansett, RI 02882
> >
> > California office
> > (866) 675-4630 Business
> > (707) 780-1951 Fax
> >
> > Main office
> > (401) 284-1827 Business
> > (401) 284-1840 Fax
> >
> > Oregon office
> > (503) 430-1065 Business
> > (503) 430-1285 Fax
> >
> > http://www.teraspeed.com
> > This e-mail contains proprietary and confidential intellectual 
> property of Teraspeed Consulting Group LLC
> > 
> ------------------------------------------------------------------------------------------------------
> > Teraspeed(R) is the registered service mark of Teraspeed Consulting 
> Group LLC
> >
>
> ------------------------------------------------------------------------
> See how Windows connects the people, information, and fun that are 
> part of your life. See Now 
> <http://clk.atdmt.com/MRT/go/msnnkwxp1020093175mrt/direct/01/>


-- 
Steve Weir
Teraspeed Consulting Group LLC 
121 North River Drive 
Narragansett, RI 02882 

California office
(866) 675-4630 Business
(707) 780-1951 Fax

Main office
(401) 284-1827 Business 
(401) 284-1840 Fax 

Oregon office
(503) 430-1065 Business
(503) 430-1285 Fax

http://www.teraspeed.com
This e-mail contains proprietary and confidential intellectual property of 
Teraspeed Consulting Group LLC
------------------------------------------------------------------------------------------------------
Teraspeed(R) is the registered service mark of Teraspeed Consulting Group LLC

------------------------------------------------------------------
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 technical documents are available at:
                http://www.si-list.net

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: