ation of

　　On-Wafer Vector

　　Network Analyzers

　　J.Randy Fenton

　　Cascade Microtech,Inc.

　　AFTFG 68th Microwave Measurement Conference

　　Nov.28–Dec.1,2006

　　See it.Touch it.Measure it.®

　　ƒInvestigate a proposed VNA comparison technique to validate on-wafer VNA systems

　　ƒOutline a detailed procedure and define terms

　　ƒExplore limitations

　　ƒShow sample results

　　ƒProvide conclusions

　　See it.Touch it.Measure it.®

　　1

　　ƒChoose an on-wafer calibration technique

　　ƒ

　　ƒ

　　See it.Touch it.Measure it.®

　　ƒWinCal XE Calibration and Measurement Software Calibration Validation

　　Application Tool Kit

　　ƒA convenient calibration validation check

　　ƒPreferably automated

　　•Quick and easy

　　•Generate insightful reports

　　img3

　　See it.Touch it.Measure it.®

　　2

　　by comparing it to an on-wafer VNA system of trusted accuracy

　　img4

　　Validate an on-wafer VNA system of unknown accuracy…

　　img5

　　See it.Touch it.Measure it.®

　　Apply comparison technique described in

　　“A Method for Comparing Vector Network

　　Analyzers”

　　D.C.DeGroot,R.B.Marks and J.A.Jargon

　　50th ARFTG Conference Digest,pp.107-114

　　Portland,OR,December 1997

　　to an On-Wafer Environment

　　See it.Touch it.Measure it.®

　　3

　　ƒMeasure a family of validation structures on both VNA systems

　　ƒCalculate the Error Vector Magnitude(EVM)difference between the two systems for each structure

　　ƒFind maxEVM for entire family of structures

　　ƒDetermine Repeatability bounds for each system

　　ƒCompare the EVMs and maxEVM to the sum of the Repeatability bounds

　　See it.Touch it.Measure it.®

　　img6

　　img7

　　ƒCriteria:Are the measurement differences between the VNAs bounded by the overall repeatability limits?

　　ƒYes–the Test system is validated for these devices.

　　ƒNo–the Test system has residual errors

　　unaccounted for during calibration that are significant relative to the Reference system.

　　See it.Touch it.Measure it.®

　　4

　　img8

　　img9

　　ƒThis comparison technique is only valid for the

　　specific set of validation devices used during the experiment.

　　ƒAssumes the Reference system has trusted accuracy.

　　ƒHowever,for many users this quick and convenient

　　validation check provides enough important feedback to either gain or lose confidence in their measurement system.

　　See it.Touch it.Measure it.®

　　EVMij=|Srefij–Stestij|

　　Srefij EVM

　　Stestij

　　ƒERROR VECTOR MAGNITUDE(EVM)difference between two measurements is

　　defined as:

　　img10

　　Where:

　　•Srefij are S-Parameter data files from Reference system

　　•Stestij are S-Parameter data files from Test system

　　See it.Touch it.Measure it.®

　　5

　　ƒACCURACY is the level of agreement of a measured or calculated quantity to its actual true value.

　　ƒMEASUREMENT ERROR is closely related to accuracy and is defined as the difference between a measured value of quantity and its true value.

　　See it.Touch it.Measure it.®

　　ƒMEASUREMENT REPEATABILITY is the variation in multiple measurement results taken by the same instrument on the same item and under the same conditions.

　　•A measurement may be said to be repeatable when this

　　variation is smaller than some agreed limit.

　　•Repeatability conditions1 include:

　　•the same measurement procedure

　　•the same observer

　　•the same measuring instrument,used under the same conditions

　　•the same location

　　•repetition over a short period of time.

　　1Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results

　　See it.Touch it.Measure it.®

　　6

　　ƒMEASUREMENT REPRODUCABILITY is the variation in multiple measurement results taken by different persons or instruments on the same item and under the same conditions.

　　•A measurement may be said to be reproducible when this variation is smaller than or equal to some agreed limit.

　　See it.Touch it.Measure it.®

　　img11

　　img12

　　img13

　　ƒRepeatability and reproducibility do not necessarily imply accuracy.

　　ƒMeasurements may be repeatable and reproducible yet inaccurate.

　　ƒTo use reproducability as a validation test,the Reference system must be trusted.

　　See it.Touch it.Measure it.®

　　7

　　img14

　　WinCal XE calculates the estimated repeatability bounds for VNAs using the method described in

　　“Calibration Comparison

　　Method for Vector

　　Network Analyzers”

　　R.B.Marks,J.A.Jargon,

　　and J.R.Juroshek

　　48th ARFTG Conference Digest,pp.38-45 Clearwater,FL,December 1996

　　See it.Touch it.Measure it.®

　　ƒTwo types of repeatability bounds:

　　•Device Dependent Bounds:Apply the error terms to the S-Parameter measurement data for each particular device

　　•Worst Case Bounds:Apply the error terms to an S-Parameter data set with all terms equal to one

　　See it.Touch it.Measure it.®

　　8

　　ƒThe overall repeatability bound was determined by adding Reference system bound(∆refR)and Test system bound(∆testR)using device dependent and worst case methods.

　　∆refR+∆testR=Overall Repeatability Bound

　　See it.Touch it.Measure it.®

　　img15

　　Reference System

　　Test System

　　S300 probe station

　　Agilent E8361A PNA

　　Summit 12K probe station

　　Agilent E8364A PNA

　　Infinity Probe(50GHz,GSG,150µm pitch)Impedance Standard Substrate(101-190)50GHz Gore RF test cables

　　WinCal XE calibration software

　　Infinity Probe(50GHz,GSG,150µm pitch)Impedance Standard Substrate(101-190)50GHz Gore RF test cables

　　WinCal XE calibration software

　　img16

　　See it.Touch it.Measure it.®

　　9

　　img17

　　25Ω,100Ω,150Ω,300Ω

　　ƒ40ps Transmission Line

　　ƒ40ps Open Stub

　　ƒAttenuators:

　　10dB,20dB

　　ƒResistors:

　　12.5Ω,25Ω,50Ω

　　ƒOffset Short

　　ƒInductor

　　ƒCapacitor

　　ƒOffset Resistors:

　　ƒLong Offset 25ΩResistor

　　The majority of these structures are

　　available on the Cascade Microtech

　　101-190 and 005-016 Impedance

　　Standard Substrates

　　See it.Touch it.Measure it.®

　　10dB attenuator S21 EVM

　　vs.10dB attenuator

　　device-dependent sum

　　of repeatability bounds

　　img18

　　img19

　　10dB attenuator S11 EVM

　　vs.10dB attenuator

　　device-dependent sum of

　　repeatability bounds

　　See it.Touch it.Measure it.®

　　10

　　img20

　　Inductor S11 EVM

　　vs.inductor

　　device-dependent sum of repeatability bounds

　　See it.Touch it.Measure it.®

　　img21

　　Capacitor S11 EVM

　　vs.capacitor

　　device-dependent sum of repeatability bounds

　　See it.Touch it.Measure it.®

　　11

　　Maximum S21 EVM

　　vs.worst case

　　sum of S21 repeatability bounds

　　img22

　　img23

　　Maximum S11 EVM

　　vs.worst case

　　sum of S11 repeatability bounds

　　See it.Touch it.Measure it.®

　　img24

　　10dB Reference system data

　　with 10dB Reference system

　　device-dependent repeatability bounds

　　See it.Touch it.Measure it.®

　　12

　　ƒTest system reproduced Reference system results within repeatability bounds

　　ƒTest system can be trusted for measurements of these types

　　ƒVNA comparison technique proved to be useful and

　　insightful as a validation and troubleshooting check

　　See it.Touch it.Measure it.®

　　img25

　　ƒValidation technique only valid for devices measured

　　ƒWorst case repeatability bounds much more conservative compared to device-dependent repeatability bounds

　　ƒVNA residual errors are device dependent

　　See it.Touch it.Measure it.®

　　13

　　ƒIf possible,choose validation structures that

　　resemble your device under test

　　•Choose validation structures that cover different regions of the Smith chart

　　•Choose validation structures with varying levels of attenuation

　　ƒAutomate using WinCal XE to make the procedure quick and insightful

　　See it.Touch it.Measure it.®

　　ƒThis technique would also be useful for on-wafer

　　round-robin inter-laboratory measurement system comparisons

　　ƒFuture work will include comparing on-wafer VNAs employing different calibration algorithms

　　See it.Touch it.Measure it.®

　　14

　　ƒTest system reproduced Reference system results within repeatability bounds

　　ƒTest system can be trusted for measurements of these types

　　ƒValidation technique only valid for devices measured

　　ƒWorst case repeatability bounds much more conservative compared to device-dependent repeatability bounds

　　See it.Touch it.Measure it.®

　　ƒVNA residual errors are device dependent

　　ƒIf possible,choose validation structures that resemble your device under test

　　•Choose validation structures that cover different regions of the Smith chart

　　•Choose validation structures with varying levels of attenuation

　　ƒAutomate using WinCal XE to make the procedure quick and insightful

　　See it.Touch it.Measure it.®

　　15

　　ƒVNA comparison technique proved to be useful and insightful as a validation and troubleshooting check

　　ƒThis technique would also be useful for on-wafer round-robin inter-laboratory measurement system

　　comparisons

　　ƒFuture work will include comparing on-wafer VNAs employing different calibration algorithms

　　See it.Touch it.Measure it.®