Microstip Vs Stripline

www.basebandhub.
com

STRIPLINE OR MICROSTRIP:
Which is better for EMC?
Karthik Raj Guruchandran
www.basebandhub.com

Abstract:
This article talks about the two most commonly used routing configurations on PCBs Microstrip and Stripline
configurations.Thearticlecoversthebasicstructuraldifferences,impedancecalculationformulasandanalysisof
emissionperformanceina3Dsimulationenvironmentforboththeconfigurations.
Background:
Ifeltprovidingabitofbackgroundonthemicrostripandthestriplineconfigurationswoulddefinitelyhelpafew
readers.Ifyouarealreadyfamiliarwiththeseconfigurations,thenfeelfreetomoveontothenextsection.
In simple terms, the Microstrip is a layout configuration where a signal trace is routed on top of a ground
reference.Thetraceisgenerallypresentonthetoporthebottomlayersonamultilayeredboard.Anexampleof
amicrostripconfigurationisshownbelow.
Figure:MicrostripConfiguration
SointhisconfigurationwehavetypicallyaFR4dielectricsurroundingthetraceononesidewhileairsurrounding
itontheother.
Ontheotherhand,inastriplineconfiguration,thetraceissandwichedbetweentwogroundlayersprovidingthe
trace an excellent reference and shielding on both sides. The downside being that it generally requires more
layersanditisnotpossibletoachievethisconfigurationwitha2layerboard.
Anexampleofstriplineconfigurationcanbeseenbelow.

www.basebandhub.com

Figure:StriplineConfiguration
There are afew more configurations that are commonly used which tend tobe derivatives ofeither microstrip
called the embedded microstrip or the stripline called the dual stripline. Please note that we will not be
discussingabouttheseconfigurationsinthisarticle.
Introduction:
As electronics engineers, we tend to use both

Stripline and Microstrip configurations quite a lot
andmostofus(includingme)preferaStriplineand
try and avoid microstrip wherever possible. Though
many favour the stripline configuration, I think we
generally are not aware of the performance
difference between the two configurations.
Striplines have the same relative dielectric constant
on both sides of the trace and have better
grounding/referencewhencomparedtoamicrostrip
configuration where the relative dielectric constant
(r) of typically around 4.5(for FR4) is on one side
while a r of 1(for air) is on the other side. The
stripline trace is also slower than a microstrip
because the effective r of microstrip is lesser than
the strip lines. So, what do these mean in terms of
emissions? Does it mean that the one configuration
is better for EMC than the other even if designed
properly?Ifso,whatwouldbedifferenceinemission
levelsbetweentheconfigurations?Oneoftheways
to answer these questions was to simulate these
traces on a 3D simulation environment and plot the
near fields associated with each trace to derive a
direct comparison between the two systems. For
this I used Microwave studio from CST a fully
integrated 3D simulation tool that can be used to
computeSparametersaswellasE&Hfields.
ModelConstruction:
Before actually constructing my model, I used the

followingformulaetocomputethetraceimpedance
of both the microstrip and stripline traces.
Alternatively, one could use calculators available in
theinternettodothesamejob.

www.basebandhub.com

Microstrip
(1)
:
+
+
=
T W
H
Ln
r
Z
8 .
98 . 5
41 . 1
87
0
Where,
WTracewidth
HDielectricthickness
TTraceThickness
rRelativedielectricconstant
Stripline
(1)
:
( )
+
+
=
T W
T H
Ln
r
Z
8 .
2 9 . 1 60
0
Where,
WTracewidth
HDielectricthickness
TTraceThickness
rRelativedielectricconstant
Ididthistomakesurethesourceimpedance(input
port in the model) the transmission line and the
termination impedance (output port in the model)
are of the same value in order to have maximum
energytransfer.
I then selected arbitrary board dimensions of

50.8mm x 25.4mm and considered just two layers
formymicrostripmodeleachwith35umofcopper
with a dielectric of 1mm thickness between them.
The dielectric I considered was lossy FR4 with a
relative dielectric constant of 4.9. Ifyousubstitute
these values in the formula for Microstrip traces
above,thenyouwouldseethattogetanimpedance
of50ohms,thetracewidthhastobe68mils(thou).
Now,withthetracedimensions,copperthicknesses
andtheFR4thicknessinhand,allIhadtodowasto
construct a board model in CST. It is fairly straight
forward job and if you know CST, then you would
probablyknowhoweasyitistoconstructaboardof
thissort.
I did a similar job for the stripline configuration and

using the equations shown, I calculated the trace
width to be 28 mils in to achieve fifty ohm
impedance. I then constructed a stripline
configuration model in a similar way as the
microstripmodel.
Simulationsetup:
OnceIhadthetracesconstructed,Ithenterminated
the two ends of both the microstrip trace and the
striplinetracewitha50ohmporttomatchwiththe
calculated trace impedance. I added a few field
monitors to monitor the 3D field, which I later used
to calculate near field that would contribute to
crosstalk. I then used a transient solver to compute
the S11 and S21 of the traces over 05GHzfrequency
range.
SimulationResults:
Now to the most interesting bit the results. As

mentioned before I simulated both Sparameters
and 3D fields on both the configurations. The S2,1
forboththeconfigurationsareshownbelow.
At 2.5GHz, the Microstrip was 0.15dB better than a

Stripline, which gradually increases to 0.27dB at
5GHz.
This clearly shows that both the traces have been

designedformaximumtransmission which makes it
easier to compare the emission performance side
byside.Sothereisverylessdifferencebetweenthe
stripline and microstrip configuration when the
traces are designed with proper impedance
matching.The3GHzwouldbetherealisticmaximum
frequency of interest for many high speed
applications and both the configurations perform
almost equally and achieve similar levels of
transmissionandlosswhendesignedwiththeexact
characteristicimpedanceineachcase.
On the 3D results, I calculated the field at 1mm

perpendicularly above the traces assuming an input
powerof1mWintothetrace.Wecouldseethatthe
Hfield is stronger in the Microstrip configuration

www.basebandhub.com

when compared to the stripline configuration. So,
when an input power of 1mW was assumed the
value of the simulated peak Hfield at 1GHz from a
stripline trace was found to be 0.141 A/m. With this
data the field in dBmW can be calculated
(considering the impedance to be 50 ohms) to be
8.01dBm,whilethesametraceat200MHzhadanH
fieldcontentof8.51dBm.
Themicrostriptraceat1GHzhadanearfieldcontent
of was around +10.36dBm which was about 18dB
morethanthestriplinetracewhileat200MHzithad
a Hfield of +9.66dBm, which means that the
stripline has definitely better emission performance
when compared with the microstrip. In general
terms, we could say that the microstrip trace is
around1520dBworsethanitsstriplinecounterpart.
Conclusion:
The Stripline configuration is arguably the best

configuration out of the two that we discussed in
terms of emission performance winning the
microstrip by over 15dB. One point to remember
though: this is an ideal scenario with perfectly
matchedimpedancesfromthedrivertothereceiver.
InreallifewedonothaveasingletraceonthePCB
but hundreds of them. Although matching
impedance is one major area of focus for a good
layout design there are many other considerations
one should bear in mind like crosstalk, ground
loops, etc. In the next article, I shall discuss about
right angle bends in PCBs and their effect on
emissionperformance.SoWatchout!
Figure:StriplineEmissionsat200MHz Figure:MicrostripEmissionsat200MHz
References:
1)http://www.ultracad.com/articles/formula.pdf

Microstip Vs Stripline

Uploaded by

Copyright:

Available Formats

Microstip Vs Stripline

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Microstip Vs Stripline

Uploaded by

Copyright:

Available Formats

www.basebandhub.

As electronics engineers, we tend to use both

Before actually constructing my model, I used the

I then selected arbitrary board dimensions of

I did a similar job for the stripline configuration and

Now to the most interesting bit the results. As

At 2.5GHz, the Microstrip was 0.15dB better than a

This clearly shows that both the traces have been

On the 3D results, I calculated the field at 1mm

The Stripline configuration is arguably the best

You might also like