FreeLists / si-list / [SI-LIST] Re: pcb board capacitors

[steve weir <weirsi@xxxxxxxxxx>]

Lee, I guess different people have different perspectives.  Personally, I 
consider the charge bucket notion a low frequency concept and decoupling an 
important high frequency concept.

At high frequencies, the problem with charge buckets doesn't seem to be 
making them big enough but getting the charge in and out fast enough.  The 
nemesis of bypass, induction is fundamentally limiting here.  I think Larry 
discussed the mythical 1uF femptohenry bypass device several months 
ago.  Or take the 5GHz at the start of the thread.  At 5GHz, a mere 30nF 
presents just a milliohm capacitive reactance.  At 3.3 V that is just 
100nanocoulombs charge store.  That's not a very big Coulomb bucket if you 
ask me.  But the equivalent inductive reactance of 31fH in a one square 
section takes a mighty thin .025 micron plate separation.  Clearly, the 
notion of just shunting our way to ultralow impedances does not scale 
easily with frequency.

I think that understanding decoupling by its original definition of 
combined series and shunt becomes critical at high frequency.  In radio 
land, this has been a well-taught concept.  Much higher series loss for a 
given loss tangent is a strong benefit of thin and high Er 
dielectrics.  Efforts have been made by researchers to develop different 
conductor materials for power planes for similar reasons. Yet another 
example is the powerful benefits of decoupling good implementations of 
in-chip bypass afford.

Regards,

Steve.


At 05:14 PM 12/5/2005 -0800, Lee Ritchey wrote:
>These things we call  bypass capacitors are actually coulomb buckets that
>supply local charge to support switching events.  When they are not large
>enough to supply the charge their terminal voltage goes down and we get
>ripple.  Fix the ripple by putting large enough coulomb buckets nearby that
>will operate at the frequencies of interest.
>
>Wish we could erase the name decoupling capacitor from our vocabulary, it
>misleads rather than enlightens as to the purpose of the capacitors.
>
>Lee W. Ritchey
>Speeding Edge
>P. O. Box 2194
>Glen Ellen, CA 95442
>Phone- 707-568-3983
>FAX-    707-568-3504
>
>I just used the energy it took to be angry to write some blues.
>Count Basie
>
>
> > [Original Message]
> > From: Fred Townsend <fred@xxxxxxxxxxxxx>
> > To: <dan1_st@xxxxxxxxx>
> > Cc: <si-list@xxxxxxxxxxxxx>
> > Date: 12/4/2005 10:48:14 PM
> > Subject: [SI-LIST] Re: pcb board capacitors
> >
> > Wouldn't it be great if there was a chart you could just look up 3.3 V @
> > 5 GHz and get the proper value. Then all the SI engineers could turn in
> > their badges and go home.
> >
> > I think you have a mistaken idea what a bypass cap does. Can you look up
> > from this text and answer the question? I give you a clue. It doesn't
> > have much to do with filtering, at least in the frequency domain sense.
> > That's done back in the power supply.
> >
> >
> >
> >
> > Answer: At 5 GHz there are really two functions. Let's consider the more
> > common. For every trace or wire, particularly power and ground, there is
> > a certain amount of inductance. For every di/dt in the chip there is a
> > corresponding -v= L(di/dt). This voltage opposes Vs causing the voltage
> > at the chip to increase or decrease according to sign of di.  If you are
> > dealing with an analog chip it will introduce a common mode signal to
> > the chip. If you are dealing with digital (time domain) it  will extend
> > rise and fall times.
> >
> > Let's look at the variables again.  -v=L(di/dt) Three variables you
> > haven't given us a clue as to their value. We would need a wegee board
> > to give you a WAG.
> >
> > Now let me shock you. Capacitors have inductance as do leads to and from
> > capacitors. Therefore capacitors have a self resonate frequency at which
> > they switch from being capacitors to inductors. This means the capacitor
> > package and capacitor leads are two more variables you interject into
> > the equation.
> >
> > What happens if you can't get enough capacitance, below the self
> > resonate frequency, you need. Inductance can be reduced by paralleling
> > capacitors. For critical applications it is common to see a capacitor
> > network of different packages and values. Yet another variable.
> >
> > At 5 GHz it is possible to see another problem. The pins of an IC form a
> > matrix of capacitive values. If the pins happen to furnish an unwanted
> > positive feedback path oscillation may occur. When this happens it is
> > necessary to null out or swamp out the unwanted capacitance. Such
> > capacitors are known as neutralizing caps.
> >
> > I hope I have explained why you will not be able to use a chart or rule
> > of thumb to determine what capacitor value to use. At 5 GHz the PCB
> > layout person controls lot more of of the design than the circuit design
> > engineer.
> >
> > Fred Townsend
> >
> >
> >
> >
> > david stern wrote:
> >
> > >hi
> > >
> > >I am designing a board working at 5GHz. The input
> > >voltage is 3.3v.
> > >What values of capacitors I have to use in order to
> > >filter the input voltage how I calculate the capacitor
> > >values where can i find material about it?
> > >
> > >THANKS
> > >
> > >DAVID
> > >
> > >
> > >
> > >
> > >
> > >
> > >
> > >__________________________________
> > >Start your day with Yahoo! - Make it your home page!
> > >http://www.yahoo.com/r/hs
> > >------------------------------------------------------------------
> > >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:
> > >http://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:
> > >             http://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
> > >
> > >
> > >
> >
> >---------------------------------------------------------------------------
>------------------------
> > >Text inserted by Panda Platinum 2006 Internet Security:
> > >
> > > This message has NOT been classified as spam. If it is unsolicited mail
>(spam), click on the following link to reclassify it:
>http://127.0.0.1:6083/Panda?ID=pav_7481&SPAM=true
> >
> >---------------------------------------------------------------------------
>------------------------
> > >
> > >
> > >
> > >
> >
> >
> >
> > ------------------------------------------------------------------
> > 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:
> > http://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:
> >               http://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:
>http://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:
>                 http://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:
http://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:     
                http://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