Maker Pro
Maker Pro

Choice of DC blocking capacitors in Microstrip Design

A

Alec

Jan 1, 1970
0
Hi,

I was wondering as to how to choose the correct value of a DC
blocking
capacitor with least possible reactance with atleast 5V and 20mA.
Center Frequency 910MHz. The value of capacitor that i need is above
200pF..


I know the thumb rule that SRF(self resonant frequency) > 2 Operating
frequency. but i enquired both Farnell and Rscomponents distributors
in UK as am here .., Only low value capacitors satify the above
criteria of SRF. How do i proceed with the same...


Thanks a lot
Alec
 
A

Alec

Jan 1, 1970
0
There's no reason for that rule. Most of the time a reasonable, like
0.33 uF, 0603 maybe, surface-mount cap has a very low RF impedance,
essentially the same ESL as most any other 0603 cap. SRF doesn't
matter, impedance does.

A 0.33 uF 0603 cap is a very low z well into the GHz. Ideally, match
the cap body width and the trace width.

John

Hi,

Is it the same for 0805 case or is there a difference.

Thanks
 
A

Alec

Jan 1, 1970
0
On Mon, 11 Aug 2008 08:48:46 -0700 (PDT), Alec
Hi,
I was wondering as to how to choose the correct value of a DC
blocking
capacitor with least possible reactance with atleast 5V and 20mA.
Center Frequency 910MHz.
[snip]



There's no reason for that rule. Most of the time a reasonable, like
0.33 uF, 0603 maybe, surface-mount cap has a very low RF impedance,
essentially the same ESL as most any other 0603 cap. SRF doesn't
matter, impedance does.
A 0.33 uF 0603 cap is a very low z well into the GHz. Ideally, match
the cap body width and the trace width.

Whose 0.33uF do you use that is not well above its resonance at 1GHz?

Bob

AVX have capacitors whose SRF =950MHz at 47pF for 0805. 33pF --> srf =
880MHz for 1210
 
D

Don Klipstein

Jan 1, 1970
0
John Larkin said:
I was wondering as to how to choose the correct value of a DC blocking
capacitor with least possible reactance with atleast 5V and 20mA.
Center Frequency 910MHz.
[snip]

There's no reason for that rule. Most of the time a reasonable, like
0.33 uF, 0603 maybe, surface-mount cap has a very low RF impedance,
essentially the same ESL as most any other 0603 cap. SRF doesn't
matter, impedance does.

A 0.33 uF 0603 cap is a very low z well into the GHz. Ideally, match
the cap body width and the trace width.

Whose 0.33uF do you use that is not well above its resonance at 1GHz?

Let's look at Alec's figures (a different post in this thread) for an
0805 47pF having self resonance at 950 MHz.

I work that out to .6 nanohenry.

Impedance of .6 nanohenry at a gigahertz is a little less than 4 ohms.

Suppose that .33 uF 0603 also has .6 nanohenry of inductance. Self
resonance would be a mere 11.3 MHz. Does that sound like what you expect?
Impedance at a gigahertz is still a little under 4 ohms.

Just to explore further - characteristic impedance of a microstrip .03
inch wide, .015 inch thick, .063 inch over a ground plane.

http://www.eetchina.com/ARTICLES/2006AUG/PDF/layout0829.pdf?SOURCES=DOWNLOAD

Page 6

I figure 82 ohms for unity dielectric factor. That works out to 273 nH
per meter (If I did that right for unity permeability and dielectric
factor), .416 nanohenry for a piece of such conductor .06 inch long.
Does this sound close to the inductance of an 0603 capacitor? If the
capacitor's ESL is close to that of a same-length same-width trace, then
the capacitor is just a capacitor or effectively just a piece of trace at
frequencies even well past the self-resonance frequency figure.

One more thing - such a capacitor will have a litle more inductance and
a little lower self-resonant frequency in free air than it will close to a
ground plane.

- Don Klipstein ([email protected])
 
T

Tom Bruhns

Jan 1, 1970
0
There's no reason for that rule. Most of the time a reasonable, like
0.33 uF, 0603 maybe, surface-mount cap has a very low RF impedance,
essentially the same ESL as most any other 0603 cap. SRF doesn't
matter, impedance does.

A 0.33 uF 0603 cap is a very low z well into the GHz. Ideally, match
the cap body width and the trace width.

John

I'll second that, with a bit more explanation: smt caps have
inductance that's determined by their physical size. A piece of wire
has inductance, after all; a length even .06 inches or .08 inches of
microstrip has inductance. What it also has, though, is capacitance
to a ground plane. The distributed inductance and the distributed
capacitance cause it to have a particular characteristic impedance,
sqrt(L/C) (neglecting the reactive part of the impedance caused by
resistance, which is generally quite small in the GHz region). If you
model the MLCC smt part as a conductor the length, width and thickness
of the part, you'll be really close to the way it actually performs in
the circuit. That is, make its inductance part of the transmission
line you mount it in. If the capacitor is slightly narrower than the
microstrip trace, the impedance should remain very close to constant
through the capacitor. If you want to get fancy about it, use the
freeware ATLC program to calculate impedance for that cross-section;
or use Agilent's ADS to accurately predict performance to well beyond
1GHz, if you put in an accurate model of your system. -- The reason
I say that ideally the capacitor will have a width slightly less than
the microstrip trace is that the added height of the part adds
capacitance to the ground plane and slightly decreases the series
inductance compared with a thin trace the same width. But...the
propagation velocity of the line is such that .06 inches of line
represents only about 3 electrical degrees at 1GHz using an FR4
substrate, less for lower dielectric constant substrates. One way to
think about what that means is that a reflection off the "leading"
edge is very nearly canceled by the corresponding reflection off the
"trailing" edge, where you have an impedance discontinuity only that
long. If you try to think about this stuff in terms of self-resonant
frequencies, you'll get yourself all needlessly worried. The model I
suggest above is far closer to what's really going on. (This assumes
the series reactance of your capacitor is very small compared with the
transmission line impedance, of course; otherwise, just include that
as a series impedance at that point along the line.)

Cheers,
Tom
 
A

Alec

Jan 1, 1970
0
Nice explanation!  Thanks, Tom!  Comforting enough for me to re-pursue
my thoughts of putting a remote repeater right down the same coax as
the cable TV ;-)

                                        ...Jim Thompson
--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon athttp://www.analog-innovations.com|    1962     |

     It's what you learn, after you know it all, that counts.- Hidequoted text -

- Show quoted text -

I am designing a phase shifter hybrids(branchline) reflection type
using pin diodes. I presumed according to the datasheet in
http://uk.farnell.com/499160/passives/product.us0?sku=avx-08051a470jat2a&_requestid=314218
having a graph with series resonance and freq plot . for my center
freq 910MHz this capacitor has SRF of 940Mhz... But theoretically, xc
= 1/cw ... will give a reactance of 3 ohm but this is where where the
parasitic inductance would be best compensated is it not...
 
A

Alec

Jan 1, 1970
0
Narrowband, there is some cap value where Xc and Xl pretty much
cancel, so that may be better than a bigger cap. But you'd probably
have to experiment, as Xl is layout and geometry and solder dependant
to some extent.

The pin diode wire bonds are maybe a bigger concern.

John- Hide quoted text -

- Show quoted text -

wat is the minimum length of a transmission line that would not hinder
soldering of capacitors(0805 case).., less than 0805 would be
difficult to solder for me.. :)

for the input port to the hybrid is an extra section of line of small
length necessary to feed in thru coaxial.... cause even a small
addition of a section line will result in phase change...?

Thanks
 
T

Tom Bruhns

Jan 1, 1970
0
The row of 0402's spanning the trace width seems to work pretty well.
A wide trace bridged by one narrow cap tends to make an inductive
blip. The expensive DiLabs parts don't TDR any better than cheap
surfmount caps.

One thing I've done is make a z-shaped slit in a fairly wide trace

__________________ ________________________________________________
||
||
||
||
||
=======================
||
||
||
||
_________________________________________||___________________________

And bridge it with a bunch of caps, two physically small ones on the
vertical parts of the slits and several bigger ones, like tantalums
even, on the long horizontal part. Seems to work well for a really
wideband DC block. If the caps aren't too tall, the
lumped-C-to-the-universe effect seems small.

John

Yes, since 50 ohm microstrip on 1.6mm (1/16 inch) FR4 with no internal
planes is over 2.5mm wide (over 0.1 inch wide), you'd want to do
something like what John suggests for that, if you really want to keep
things constant RF impedance. I tend to deal with microstrip on
multilayer boards where the trace width is very similar to the width
of an 0603 part. It becomes one of those things you think about in
system design: the board stackup is not independent of the passive
part selection. But again, until you get to pretty high frequencies
(and especially if you are trying to keep things very wideband) it
tends to not be a problem. If you are dealing with narrow-band, like
I suspect the OP was ("center frequency 910MHz..."), you can pretty
easily tune out the small effect of an 0.08" long capacitor which
doesn't exactly maintain the microstrip's impedance. If I model an
0.08" long 80 ohm section (typical microstrip velocity factor)
inserted into a 50 ohm line, at 1GHz, I see a return loss degradation
to -28dB, and a whopping 6 millidegree insertion loss due to
mismatch: both probably not worth worrying about outside of precision
measurement systems. I expect that a conductor of the same cross-
section as an 0805 cap, as a microstrip trace on a 1/16" board with
ground on the other side, would be a lower impedance than 80 ohms,
probably about 70-75 ohms.

Cheers,
Tom
 
S

Simon S Aysdie

Jan 1, 1970
0
I had completely forgotten the old semi-rigid coax trick.  With a milled
slot for the coax body I actually got away with that trick at 13.3 with very
decent results.

I use it all the time. In fact, I do layouts where the mask is
already cleared off the ground where I know I'll probably want to tap
in. (And making sure there are plenty of gnd vias there.)

I've been using 47 mil od with sma, up to 8G, for now.

-----------
Regarding the DC blocking, I have to echo what JL sez. I've tried to
measure 0402 L, and estimate it to be around 0.2 nH. If the
application is wideband, then just put the biggest cap possible, as
0.2 nH is ignorable to fairly high frequencies. If narrowband, then
just pick for the SRF, as it will be very slightly less series Z.

Also, in some cases and at higher frequencies the dissipation R can
become as significant as the inductance X. So a jelly-bean X7R might
not be so good in that respect. I think ATC has some better Q
wideband caps.
 
S

Simon S Aysdie

Jan 1, 1970
0
Having never done stripline myself, what's the best way to transition
from 75ohm coax to the stripline?

Whether it is 50 or 75 doesn't really matter. If you are at a high
frequency that it matters, then the launch onto the board will
probably be an edge mount 3.5 (or similar) to either microstrip or
grounded co-planer waveguide. Then transitioning to stripline is yet
another discontinuity.

There should be no annular rings on the inner layer pads for the "via"
down to the stripline layer. (That layer has the annular ring, of
course.) Also, back-drilling the via is not too expensive from some
manufacturers. (Cheaper than blind vias.) Instead of, say 30 mil of
"hanging via," you can get it down to maybe 2-7 mil with the back-
drill tolerance. To state the obvious, what you do depends upon
you're requirements.

To really get an idea of the transition's performance before you build
it, you'd need a full 3D simulator. HFSS or CST, that is.

THis was amusing at a quick scan:
http://aoltean.web.cern.ch/aoltean/Work/SIanalysis_aoltean_4092006.ppt

from googling:
http://www.google.com/search?hl=en&q=hfss+cst+3d&btnG=Search
 
Top