[SI-LIST] Re: Capacitors arrays. worth it or not?
- From: <istvan.novak@xxxxxxxxxxx>
- To: "Hill, John" <jhill@xxxxxxxxxx>, Istvan.Novak@xxxxxxx
- Date: Wed, 24 May 2006 21:07:06 -0500 (CDT)
John,
Yes, you are correct; the series resonance frequency of the part is determined
by the capacitance and the total loop inductance. The capacitor body is part
of the loop.
There are the following clarifications. I would not use the length of the 0508
part to calculate the delay, which determines the quarter-wave resonance. As
Larry pointed out, the height of the part has to be used, 30-50 mils for a
typical 0508 capacitor. In addition, the tpd=c/sqrt(Dk) does not apply in this
case. the resonator is formed of a heavily loaded transmission line. A crude
approximation of this propagation delay is tpd=sqrt(LC), where L is the
inductance of the vertical terminal structure and C is the full capacitance of
the part.
Regards,
Istvan Novak
SUN Microsystems
From: "Hill, John" <jhill@xxxxxxxxxx>
Date: Wed May 24 16:58:36 CDT 2006
To: Istvan.Novak@xxxxxxx
Cc: steve weir <weirsi@xxxxxxxxxx>, Larry Smith <LSMITH@xxxxxxxxxx>,
Mark.Randol@xxxxxxxxxx, si-list <si-list@xxxxxxxxxxxxx>
Subject: [SI-LIST] Re: Capacitors arrays. worth it or not?
Istvan,
Yes, I ran the numbers:
Electrical length of AVX 0508 LICC is = physical length * square root of
Dielectric Constant.
The Dielectric Constant is somewhere above 10K. Let's use 10K.
The wavelength is 4 time the quarter wavelength = 4 * the Electrical
length = 4 * 1.27 mm * square root of 10K.
Frequency = C / wavelength = C / 4 * 1.27 mm * square root of 10K = 590
MHz.
But the resonance on the data sheet is 10 MHz.
So I guess we are back to the lead inductance and the capacitance of the
MLCC causing the series resonance.
And this lead inductance can be reduced with a reduction in the loop
area of the leads to the MLCC and by reducing the length of the leads. I
am guessing that is why the IDC and X2Y capacitors work. They reduce the
area of the loop of current feeding the capacitor. Is this correct?
Best regards,
John
________________________________
From: Istvan.Novak@xxxxxxx [mailto:Istvan.Novak@xxxxxxx]
Sent: Wednesday, May 24, 2006 4:49 PM
To: Hill, John
Cc: steve weir; Larry Smith; Mark.Randol@xxxxxxxxxx; si-list
Subject: Re: [SI-LIST] Re: Capacitors arrays. worth it or not?
John,
The series resonance frequency of the mounted capacitor does
not directly relate to the open-ended transmission line of
the capacitor body. The vertical transmission line that
Larry referred to has high losses, partly because of the metal
and dielectric losses, but also due to the electrical loading
of the plates. You can get this kind of model described in:
"Slow Wave Causal Model for Multi Layer Ceramic Capacitors"
on http://home.att.net/~istvan.novak/papers.html
Regards,
Istvan
Hill, John wrote:
Steve,
I am not so sure of this. In the MLCC there is a transmission line with
an end that is open. It is the open 1/4 wavelength transmission line
that provides the low impedance on the other side of the part.
Specifically, a 1/4 wavelength away from the open is a short.=20
In the IDC part the end is shorted, not open. There is a difference and
I am uncertain how to predict the resonance. That is why Larry's
viewpoint is important. It provided a better understand of the fields in
the parts and why the resonance occurs in the first place. I would like
to understand the IDC parts as well as we now understand the MLCC parts.
Can anyone explain to me the physics behind the IDC resonance as
compared to the MLCC resonance?=20
I would also like to know the Dielectric constant of the MLCC ceramic to
run the numbers and confirm the electrical length of the MLCC part. It
would be interesting to confirm the measured series resonance is truly
predicted by the 1/4 wavelength of the MLCC parts.
John
-----Original Message-----
From: steve weir [mailto:weirsi@xxxxxxxxxx]=20
Sent: Wednesday, May 24, 2006 4:12 PM
To: Hill, John; Larry Smith; Mark.Randol@xxxxxxxxxx; si-list
Subject: RE: [SI-LIST] Re: Capacitors arrays. worth it or not?
John, the physics is the same for an IDC. The only difference is=20
that you multiple coupled lines in each part of Larry's model. Until=20
you get to really high frequency, you can simply replace those with=20
smaller equivalent inductance of Ls - Lm.
Regards,
Steve.
At 10:59 AM 5/24/2006, Hill, John wrote:
Larry,
Your analysis of a capacitor as a transmission line is very
insightful.
It explains the physics behind why a MLCC has a series
resonance. I
then
tried to apply the same technique to an IDC capacitor with both
terminals of the capacitor on both sides of the part.
I was not able to determine where the first resonance would be.
How
would you apply this technique to an IDC capacitor? And for that
matter
does anyone know where the first resonance would be for an IDC
capacitor?
Best regards,
John
-----Original Message-----
From: Larry Smith [mailto:LSMITH@xxxxxxxxxx]
Sent: Tuesday, May 23, 2006 12:47 PM
To: weirsi@xxxxxxxxxx; Hill, John; Mark.Randol@xxxxxxxxxx;
si-list
Subject: RE: [SI-LIST] Re: Capacitors arrays. worth it or not?
Steve, John - here is another interesting way to look at it.
When a
ceramic capacitor is mounted such that the capacitor plates are
parallel
to the PCB plates, the capacitor forms a vertical transmission
line
with
inductance per unit length and capacitance per unit length. The
far
end
of the transmission line is open circuit (air) and the near end
is
nearly shorted (PCB power planes present a very low impedance to
the
mounted capacitor). This makes a quarter wavelength resonator
similar
to a 50 Ohm transmission line that is open on one end. At the
quarter
wavelength frequency, the open circuit at the far end becomes a
low
impedance node at the near end. This is the series resonant
frequency
of the capacitor. The peaks and dips that you might see beyond
the
series resonance of the capacitor are associated with the half,
3/4,
full, 1 1/4, etc, wavelengths of the capacitor transmission
line.
To see this effect, you must mount the capacitor on vias and
pads that
have less inductance than the capacitor itself otherwise the
resonance
is completely dominated by the mounting inductance. Tall
capacitors
with very low ESR show this effect the best. I was evaluating a
bunch
of capacitors one time and the 33nF NPO 2220 size capacitor
strongly
exhibited these properties. It is very tall and has low ESR
because of
the many, many plates. I had it mounted on a fixture estimated
at 83pH
but the inductance associated with the capacitor itself was
about 1 nH.
There were perhaps a half a dozen dips and peaks beyond series
resonance
associated with the transmission line properties of the
capacitor.
David Hockanson and I did a couple of papers on this at 2002 and
2003
ECTC conference and another one at 2005 Design Con. You can
actually
use the transmission line properties of a capacitor to develop a
ladder
SPICE model and extract the element values. The model
accurately
predicts the reduction of inductance and the increase in ESR of
a
capacitor mounted on low inductance pads. This is important to
simulate
the parallel resonance that may occur between two capacitors or
between
a capacitor and power planes. Ceramic capacitors have a lot
more ESR
and less ESL than might be expected from a simple RLC model
beyond
series resonance.
Mounting capacitors with the plates perpendicular to the PCB
planes
(rather than parallel) eliminates or at least greatly changes
this
mechanism. As Istvan mentioned, the resonances are very much
reduced.
I believe that there is still an increase in ESR and a reduction
of ESL
beyond series resonance as the current does not want to get very
far
away from the PCB power planes (big inductive loop). Once
again, you
would have to mount the capacitors on very low inductance mounts
in
order to see this.
Regards,
Larry Smith
Altera Corporation
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx
[mailto:si-list-bounce@xxxxxxxxxxxxx]
On Behalf Of steve weir
Sent: Monday, May 22, 2006 11:27 PM
To: Hill, John; Mark.Randol@xxxxxxxxxx; si-list
Subject: [SI-LIST] Re: Capacitors arrays. worth it or not?
John, you've got it. The vertical orientation makes for N very
small
cavities all acting in parallell. Hence, the secondary
resonance
occurs at a much higher frequency.
Regards,
Steve.
At 01:50 PM 5/22/2006, Hill, John wrote:
Steve,
I'm sorry to be a little thick on this issue, but I may
be getting
the
two orientations confused. If I understand you
correctly, the
following
is true:
It is Edie currents that keep the current concentrated
in the lower
plates of a capacitor when the capacitor is mounted
horizontally,
which
we are defining as having the plates parallel to the
board. This
creates
a resonate cavity under the part.
When the capacitor is mounted with the plates
perpendicular to the
board
the current flows through all the plates.
Is this correct?
John
---------------------------------------
The information in this email and attachments hereto may
contain
legally privileged, proprietary or confidential
information that is
intended for a particular recipient. If you are not the
intended
recipient(s), or the employee or agent responsible for
delivery of
this message to the intended recipient(s), you are
hereby notified
that any disclosure, copying, distribution, retention or
use of the
contents of this e-mail information is prohibited and
may be
unlawful. When addressed to Takata customers or vendors,
any
information contained in this e-mail is subject to the
terms and
conditions in the governing contract, if applicable. If
you have
received this communication in error, please immediately
notify us
by return e-mail, permanently delete any electronic
copies of this
communication and destroy any paper copies.
---------------------------------------
-----Original Message-----
From: steve weir [mailto:weirsi@xxxxxxxxxx]
Sent: Monday, May 22, 2006 2:41 PM
To: Hill, John; Mark.Randol@xxxxxxxxxx; si-list
Subject: Re: [SI-LIST] Re: Capacitors arrays. worth it
or not?
John, when the plates of a cap are parallel to the
planes, eddy
currents block field penetration to the upper plates
through the
cavity. The field still goes around the terminal
metalization. This
makes a resonant cavity.
Steve.
At 10:50 AM 5/22/2006, Hill, John wrote:
Mark,=3D20
We have uploaded an application note from
American Technical
Ceramics
concerning placing capacitors vertical and
horizontal. The URL is:
http://si-list.org/files/tech_files/ATC%20select_cap_wireless.pdf
I have also sent an e-mail to the application
engineer about the
issue
of eddy currents limiting the field. The data in
the application
note
does not look like it supports the idea and I do
not understand the
physics.=3D20
John
=3D20
---------------------------------------
The information in this email and attachments
hereto may contain
legally =3D
privileged, proprietary or confidential
information that is
intended
for =3D
a particular recipient. If you are not the
intended recipient(s),
or
the =3D
employee or agent responsible for delivery of
this message to the =
=3D
intended recipient(s), you are hereby notified
that any disclosure,
=3D
copying, distribution, retention or use of the
contents of this
e-mail
=3D
information is prohibited and may be unlawful.
When addressed to
Takata
=3D
customers or vendors, any information contained
in this e-mail is =
=3D
subject to the terms and conditions in the
governing contract, if =
=3D
applicable. If you have received this
communication in error,
please
=3D
immediately notify us by return e-mail,
permanently delete any =3D
electronic copies of this communication and
destroy any paper
copies.
---------------------------------------
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx
[mailto:si-list-bounce@xxxxxxxxxxxxx]
On Behalf Of Mark Randol
Sent: Monday, May 22, 2006 10:33 AM
To: si-list
Subject: [SI-LIST] Re: Capacitors arrays. worth
it or not?
-----Original Message-----
From:
si-list-bounce@xxxxxxxxxxxxx=3D3D20
[mailto:si-list-bounce@xxxxxxxxxxxxx] On
Behalf Of Jerry
Martinson
=3D3D20
I've always wondered how discrete cap
performance is
affected=3D3D20
if the caps plates are parallel with the
underlying plane
or=3D3D20
if they are perpendicular. I'd think
that having them
rolled=3D3D20
90 degrees
(perpendicular) might make them perform
better in some=3D3D20
regions. I wonder how this would extend
to arrays and=3D3D20
whether arrays are configured rolled or
not. Does anyone=3D3D20
know? =3D3D20
American Technical Ceramics (ATC) used to
recommend 'vertical'
placement
of their porcelain caps for just this reason.
I've seen it make
several
100MHz's of difference in the measured resonance
frequency. That
was
on
a relatively thick 2 layer PCB, so on a board
with a thinner
component
to ground layer spacing this could be more
significant <guess>. I
didn't find it on their web site, but here is
their link.
http://www.atceramics.com/
Now how much of this was due to plate coupling
to the substrate, or
reduced effective capacitance and inductance
because of current
crowding
towards the new 'bottom' of all the plates,
beats me. It seems to
me
in
the horizontal orientation, the upper plates
would have slightly
more
inductance due to the greater loop area. Which
effect dominates,
capacitance or inductance? =3D3D20
The problem at the time was fixed, so we didn't
investigate
further.
--
Mark Randol, RF Evaluation & Application
Engineer
Not speaking for my company, etc
------------------------------------------------------------------
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
Other related posts: