Passive Intermodulation

Measurement Techniques

PIM
D

Tx

D
PIM

DUT

Tx
PIM
D

D
PIM

Reflected

Transmitted

| INTRODUCTION

Through interaction with engineers and technicians in many countries throughout the
world,Kaelushashadtheopportunitytodiscussandevaluatemeasurementtechniquesas
they apply to a wide variety of items. We have discussed the problem of IM with
component manufacturers, infrastructure providers, site managers and service providers.
This perspective combined with the knowledge gained through the development and
manufactureofourPassiveIntermodulationAnalyzers,formsthebasisofourapproachto
the measurement of passive intermodulation, and the test capabilities built into our
analyzers.
Ourfindingssuggestthefollowing:

Consistently manufacturing low IM items is extremely difficult to achieve in mass

quantitieswheretheemphasisisoncostreduction

DynamicmeasurementsshouldbeperformedonallPIMcriticaldevices
Fixedfrequencymeasurementsmaynotbeadequateduetothefrequencydependent
characteristicsexhibitedbymanydevicesandsubsystems

| WHY TO WE CARE ABOUT PASSIVE INTERMODULATION (PIM)?

PassiveIMisanincreasingconcerninthewirelesscommunity,becauseofthedebilitating
affect it can have on the performance of the telecommunications network. Passive IM is
generatedwhenevermore thanone frequencyencounters a nonlinearelectrical junction
or material. The resulting creation of undesired signals, mathematically related to the
originalfrequencies,canresultindecreasedsystemcapacityand/ordegradedcallquality.
Reduced capacity and call quality at a cell site leads to reduced income for the wireless
serviceproviders.Theeconomiclossiscompoundedwhentheaffectedcustomerbecomes
frustratedandswitchesovertoacompetingprovider.

2000 Kaelus Inc. All Rights Reserved.

www.kaelus.com

| Q&A

Question: IsPassiveIMADesignIssue,AManufacturingIssue,orAMaintenanceIssue?
Answer: AlloftheAbove.
Agooddesignisanecessary,butnotasufficientconditionforsuccess
Atonetime,manycompaniesconsideredIMsomethingthatcouldbecontrolledsimplyby
followingafewwelldocumenteddesignrules.

Avoidtheuseofferrousmetals

Solderorcoldweldalljunctionswherepossible

Minimizethenumberofcontactjunctions
Designallcontactjunctionssuchthattheyarepreciseandundersufficientpressure
tomaintaingoodcontact

Avoiddissimilarmetalsindirectcontact
Plateallsurfacestopreventoxidation
Makecertainplatingisuniformlyappliedandofsufficientthickness

Thebiggerchallengeisbuildingit
Although, the rules seem straightforward, manufacturing excellence isthe key to success.
MinordeviationsfromtheidealprocesscancreateunacceptablelevelsofPassiveIM.
Someoftherealworlddeficienciesthatoccur:

Pooralignmentofparts
Inadequatelytorquedscrewsandfasteners
Badsolderjoints
Insufficientorincompletecleaningofpartspriortoplating
Contaminatedplatingtanks
Platingmaterialbuildup
Usingwrongmaterials
Poorplatingadhesion

Thenyouhavetointegrateandinstallit
Once you have good designs and wellmanufactured components, the next challenge is
integrationandsysteminstallation.RecallthatalljunctionsarepotentialPIMgenerators.
2000 Kaelus Inc. All Rights Reserved.

www.kaelus.com


Therefore,everyinterconnectisapotentialsourceoftrouble.Jointsthattestoutfineina
benign environment can become problematic when mechanically stressed (bending of
cables,torsionalloadsonconnectorinterfaces,over/undertorquedconnectors).
In summary, it is not a given that interconnecting components that meet their IM
specificationassuresasubsystemthatalsomeetstheIMspecification.
AndthenMotherNaturetakesover.
Thenaturalelementsaretheenemyofthoseportionsofthenetworkthatareexposedto
them.

Windinducedvibration
dailytemperaturevariations
moistureinitsvariousforms
thermalloadingbythesun
airbornedirt

Each of these works to break down and erode the quality of the components in the
network,andultimatelythecommunicationchannel.Theweakeningofjoints,separationof
junctions,invasionofmoisture,oxidationofmaterials,andcontaminationbydirtanddust
allgiverisetopassiveintermodulationanddesensitizationofthereceiver.
Theresultisacellsiteperformingbelowitsdesignpotential
| TEST METHODS
QuantifyingpassiveIMperformance
Good communication quality requires maintaining an acceptable CarriertoInterference
(C/I)ratio.Therefore,thegoalistokeepIaslowaspossible.Intheidealcase,Iwould
always be below the receiver noise floor. One source of undesired interference is passive
intermodulation.
A typical specification requires a passive IM level no greater than 110dBm when two
+43dBm carriers are injected into the device under test, i.e., 153dBc. To put this into
perspective, this is a ratio of 1:2,000,000,000,000,000, or the equivalent of trying to
measurethedistancetothesuntoanaccuracyofonetenthofamillimeter.

2000 Kaelus Inc. All Rights Reserved.

www.kaelus.com

Theoldway
A commonly used approach to measuring passive IM is to inject two fixed frequency CW
tones of 20W per tone into the device under test and measure the IM power levels
generated. This test scenario is consistent with that proposed by the international
standardscommitteetaskedwithdefiningtestmethodsforthemeasurementofpassiveIM
(IECTC46WG6).
To realize the necessary test setup, the classic approach has been to rack and stack two
synthesizersfeedingtwohighpoweramplifiersthatconnecttoanarrayofRFcomponents,
whichcombine,filterandduplexthesignals,andthenroutethesignalofinterestthrougha
lownoiseamplifiertoaspectrumanalyzerfordetectionanddisplay.Apowermeterisused
tosetthepropertransmitpowersandthenreadjustonaperiodicbasistocompensatefor
drift.
TherackandstackapproachtobuildingapassiveIMtestbenchisclearlyadifficultprocess.
Furthermore, because there are so many discrete instruments, components and
interconnectingcables,themeasurementresultsaresometimesdifficulttorepeat,andthe
setups are often unstable and susceptible to damage. As if that werent enough to
discourage all but the stouthearted, the measurement is timeconsuming, and as we are
beginningtorealize,theresultscanbeinconclusiveandmisleading.
TheKaelusWay
WithanunderstandingofhowpassiveIMtestingwasperformed(andtheassociated
deficiencies),Kaelussetforththefollowingdesigngoals:

Buildatestinstrumentthatproduceshighquality,repeatablemeasurements

Staycurrentwithemergingtechnologytrendsandteststandards

Maketheinstrumenteasytosetupanduse
Advancethestateoftheartbyprovidingnewandusefultesttools
Bringstandardizationtothisdifficultmeasurementinordertoestablishaclearbasis
ofperformancecomparison

Listentothecustomerandincorporatetheirrecommendationsasappropriateto
providethemostcomprehensiveandeasytousetestinstrumentpossible

This led to the three significant design features incorporated into Kaelus Passive IM
Analyzers:
1. Ahighlyintegrateddesign:Minimizethenumberofexternalcomponentsbypackaging
all of the RF components in a single enclosure. This results in improved reliability,
stabilityandrepeatability.

2000 Kaelus Inc. All Rights Reserved.

www.kaelus.com

2. Highspeeddigitalreceivertechnology.ThespeedandthefrequencyagilityoftheKaelus
analyzersmakeitpossibleto:

MeasuremultipleIMproductssimultaneously
MeasureandrecordtransienteventssuchasPIMburstandchangingIMdueto
mechanicalandenvironmentalstress
Measuresweptfrequencyintermodulationcharacteristicstoverifyperformanceover
theentirespecifiedfrequencyrange

3. Innovativenewtestcapabilities:Theresultinginstrumentiseasytouse,timeefficientand
thorough, so that it is practical to make the measurements necessary to assure product
qualityandnetworkintegrity.
| REVEALING REALITY

Most measurements are made with a single pair of carrier frequencies, usually set at the
transmitbandedgesforthewirelessstandardyouweretesting.
For example, consider the measurement of a device used in a PCS1900 Band network.
Typically,thistestisperformedbysettingCarrier1to1930MHzandCarrier2to1990MHz.
(Note:TherearesomePIMtestsetupsthat,duetotheirdesign,mustuseonecarriersetto
afrequencyinthereceivebandwhiletheothercarrierissettoafrequencyinthetransmit
band.Thevalidityofthisapproachopentodebate.)
After the transmit frequencies and powers are established and the spectrum analyzer
configuredtomeasuretheIMresponse,thetestisinitiated.Duringthemeasurement,care
is often taken not to disturb the device under test or the test setup in case instabilities
mightproduceadataaspikeattheprecisemomentthespectrumanalyzerhappenstobe
sweepingpasttheIMfrequency.
ExperiencesuggeststhatthisapproachtomeasuringPIMismuchtoolimitedandleavestoo
manypotentialproblemshidden.
DynamicMeasurements
Despitethebesteffortstodesignandmanufactureproductsthatarestableregardlessof
their environment, the passive IM response of components and subsystems can vary
dramatically when subjected to stress. In a dynamic measurement, IM performance is
monitoredduringtheapplicationofanappropriatestimulus.
Theissueofdynamicmeasurementshasbeengivenagreatdealofattentionwithregardto
cable assemblies. The concern has to do with the vulnerability of the connector/cable
interfaceaswellasIMcreatedinthecable(bymicrocracksinsolidconductorcablesand
2000 Kaelus Inc. All Rights Reserved.

www.kaelus.com


discontinuouscontactinbraidedcables).TestinginvolvesmeasuringtheIMasthecableis
flexedand/orabendingmomentisappliedattheconnector/cableinterface.
In our view, all components on the base station transmit side of the
wireless network should be subjected to this scrutiny. Two
measurements routinely performed by Kaelus on all PIM critical
componentsusedinouranalyzersarethetaptestandthebending
momenttest.Thetaptestissimplyamatteroftappingthedeviceand
watchingtheIMresponse.Forexample,tappingthetuningscrewson
filtersfrequentlygenerateshighlevelsofpassiveIM.Afterthetapping
is stopped, the IM sometimes returns to its low IM condition and
sometimesitremainshigh.Thetaptesthasbeendemonstratedtobe
highly useful in screening devices and cables that will fail at some
future(andinconvenient)time.

Twomeasurements
routinelyperformed
byKaelusonallPIM
criticalcomponents
usedinouranalyzers
arethetaptest
and,thebending
momenttest.

Thisisillustratedinthefollowingexample.Inthiscase,weusedtheStripChartModeofthe
Kaelus analyzer to record the IM response over time. The device under test is a PCS1900
bandpassfilter.WhatweareobservingisthechangingIMduetotapping.Latertestingalso
showedthisdevicetohavemuchdegradedIMperformancewithtemperaturechanges.

TM

Measured IM3 During "Tap Test"

-60

IM3, dBm at 2 x 20W Tones

-70

-80

-90

-100

-110

-120

-130
0

10

2000 Kaelus Inc. All Rights Reserved.

www.kaelus.com

15

20

25

30

Time, Seconds
Figure 1: Dynamic IM Measurement

35

40

45


Thebendingmomenttestisperformedbyapplyingamodestsideforcetotheconnectoron
the device under test. If the connector is not adequately attached to the body of the
component, or if the launching mechanism within the device is not solid, IM will be
generated.
SweptFrequencyMeasurements
Until now, the accepted method for measuring passive IM was to set the carrier to two
fixed frequencies, usually, but not always in the transmit band, and measure the IM
generated.Whathasbeendiscoveredisthatonmanydevices,thisisnotadequate,because
thesedevicesexhibitIMcharacteristicsthatvaryasafunctionoffrequency.KaelusPassive
IMAnalyzersmakethismeasurementatmultiplefrequenciespracticalandeasy.
An example clearly demonstrates the virtue of the swept frequency IM measurement. In
thiscase,aPCS1900bandduplexerwithabadconnectorismeasured.Itshouldbenoted
thattheconnectordamagewascausedbyovertorquingtheconnectorandisundetected
by visual inspection and swept return loss measurements. The comparison between the
goodandthebadduplexerreturnlossisshowninthefollowingfigure.

TM

Duplexer ANT Port Return Loss

0.00
Bad Duplexer
Good Duplexer

-5.00

Return Loss, dB

-10.00

Uplink Band 1850 to 1910 MHz

Downlink Band 1930 to 1990 MHz

-15.00

-20.00

-25.00

-30.00

-35.00

-40.00

1850

1860

1870

1880

1890

1900

1910

1920

1930

Frequency, MHz

2000 Kaelus Inc. All Rights Reserved.

www.kaelus.com

1940

1950

1960

1970

1980

1990


The good duplexer meets the 115dBm specification at all frequencies, but the bad
duplexerhasaseverelydegradedIMasthetranmsitcarriersaremovedfromthebandedges.
BothduplexerswouldpassQAscreeningifonlybandedgecarrierswereusedforthetesting.
Tofullycharacterizethedeviceundertest,thesweptfrequencycapabilityoftheKaelusPassiveIM
AnalyzerrecordstheIMperformanceaseachofthetwocarriersissweptinfrequencyacrossthe
transmit(downlink)band.Onlyonecarrieratatimeisswept,whiletheremainingcarrierisheld
fixedatthebandedge.Intheplotbelow,Carrier1isfixedat1930MHzwhileCarrier2issweptfrom
1950to1990MHzproducingthirdorderintermodulationproducts(IM3)thatoccuratfrequencies
from1870to1910MHz.
Themeasureddatacanbeloggedtoafile,copiedtothecomputerclipboardwhereitcanbepasted
intoanotherapplication,oritcanbedirectlyhardcopiedtoaprinter.

| SUMMARY

Withthewirelessserviceprovidersworkinghardtowinandmaintaincustomers,controllingthe
generationofpassiveintermodulationdistortioncanimpactthefinancialsuccessoftheproviders.
ThepresenceofpassiveIMatacellsiteimpactsbasestationcapacityandservicequality.Therefore,
itmustbeconsidered,measuredandcontrolledfromdesignthroughmanufacturing,andatthebase
stationinordertorealizeoptimumperformance.
TM

3rd Order IM Response

-90
Bad Duplexer

Good Duplexer

-95

IM3 Magnitude, dBm

-100

-105

-110

-115

-120

-125

-130
1870

1875

1880

1885

1890

1895

IM3 Frequency, MHz

2000 Kaelus Inc. All Rights Reserved.

www.kaelus.com

1900

1905

1910