网络分析仪原理

nullNetwork Analyzer BasicsNetwork Analyzer BasicsNetwork Analysis is NOT.…Network Analysis is NOT.…nullWhat Types of Devices are Tested?Device typeActivePassiveIntegrationHighLowAntennas Switches Multiplexers Mixers Samplers Multipliers DiodesDuplexers Diplexers Filters Couplers Bridges Splitters, dividers Combiners Isolators Circulators Attenuators Adapters Opens, shorts, loads Delay lines Cables Transmission lines Waveguide Resonators Dielectrics R, L, C'sRFICs MMICs T/R modules Transceivers Receivers Tuners Converters VCAs Amplifiers VCOs VTFs Oscillators Modulators VCAtten’s TransistorsnullDevice Test Measurement ModelNFStimulus typeComplexSimpleComplexResponse toolSimpleDC CW Swept Swept Noise 2-tone Multi- Complex Pulsed- Protocol freq power tone modulation RFDet/ScopeParam. An.NF Mtr.Imped. An.Power Mtr.SNAVNASAVSA84000TG/SADed. TestersI-VAbsol. PowerGain/FlatnessLCR/ZHarm. Dist. LO stability Image Rej.Gain/Flat. Phase/GD Isolation Rtn Ls/VSWR Impedance S-parametersCompr'n AM-PMRFIC testFull call sequencePulsed S-parm. Pulse profilingBER EVM ACP Regrowth Constell. EyeIntermodulation DistortionNFMeasurement planenullLightwave Analogy to RF Energy RFIncidentReflectedTransmitted LightwaveDUTWhy Do We Need to Test Components?Why Do We Need to Test Components?Verify specifications of “building blocks” for more complex RF systems Ensure distortionless transmission of communications signals linear: constant amplitude, linear phase / constant group delay nonlinear: harmonics, intermodulation, compression, AM-to-PM conversion Ensure good match when absorbing power (e.g., an antenna)nullThe Need for Both Magnitude and Phase4. Time-domain characterization 5. Vector-error correction2. Complex impedance needed to design matching circuits 3. Complex values needed for device modeling 1. Complete characterization of linear networks AgendaAgendaWhat measurements do we make? Transmission-line basics Reflection and transmission parameters S-parameter definition Network analyzer hardware Signal separation devices Detection types Dynamic range T/R versus S-parameter test sets Error models and calibration Types of measurement error One- and two-port models Error-correction choices Basic uncertainty calculations Example measurements AppendixnullTransmission Line BasicsLow frequencies wavelengths >> wire length current (I) travels down wires easily for efficient power transmission measured voltage and current not dependent on position along wireHigh frequencies wavelength » or << length of transmission medium need transmission lines for efficient power transmission matching to characteristic impedance (Zo) is very important for low reflection and maximum power transfer measured envelope voltage dependent on position along lineI+-Transmission line ZoTransmission line ZoZo determines relationship between voltage and current waves Zo is a function of physical dimensions and r Zo is usually a real impedance (e.g. 50 or 75 ohms)nullPower Transfer EfficiencyFor complex impedances, maximum power transfer occurs when ZL = ZS* (conjugate match)Maximum power is transferred when RL = RSRL / RSnullTransmission Line Terminated with Zo For reflection, a transmission line terminated in Zo behaves like an infinitely long transmission lineZs = ZoZoVrefl = 0! (all the incident power is absorbed in the load)Zo = characteristic impedance of transmission linenullTransmission Line Terminated with Short, Open Zs = ZoVreflFor reflection, a transmission line terminated in a short or open reflects all power back to sourceIn-phase (0o) for open, out-of-phase (180o) for shortnullTransmission Line Terminated with 25 W VreflStanding wave pattern does not go to zero as with short or openZs = ZoZL = 25 WnullHigh-Frequency Device CharacterizationTransmittedIncidentTRANSMISSIONGain / LossS-ParametersS21, S12GroupDelayTransmissionCoefficientInsertion PhaseREFLECTIONSWRS-ParametersS11, S22ReflectionCoefficientImpedance, Admittance R+jX, G+jB ReturnLoss G, rT,tIncidentReflectedTransmittedRBABR=nullReflection Parameters¥ dBNo reflection (ZL = Zo)01Full reflection (ZL = open, short)0 dB1¥Voltage Standing Wave RationullSmith Chart Review ¥ ®Smith Chart maps rectilinear impedance plane onto polar plane..2.4.6.81.0Polar planeZ = ZoL=0GConstant XConstant RSmith chartnullTransmission ParametersVIncidentTransmission Coefficient = T= = tÐfDUTGain (dB) = 20 Log = 20 log tInsertion Loss (dB) = - 20 Log = - 20 log tnullLinear Versus Nonlinear BehaviorLinear behavior: input and output frequencies are the same (no additional frequencies created) output frequency only undergoes magnitude and phase changeFrequencyDUTInputOutputNonlinear behavior: output frequency may undergo frequency shift (e.g. with mixers) additional frequencies created (harmonics, intermodulation)nullCriteria for Distortionless Transmission Linear Networks Constant amplitude over bandwidth of interestMagnitudePhaseFrequencyFrequencyLinear phase over bandwidth of interestnullMagnitude Variation with Frequency F(t) = sin wt + 1/3 sin 3wt + 1/5 sin 5wtTimeLinear NetworkFrequencyFrequencyFrequencyMagnitudeTimenullPhase Variation with Frequency FrequencyMagnitudeLinear NetworkFrequencyFrequencyTime0-180-360°°°TimeF(t) = sin wt + 1 /3 sin 3wt + 1 /5 sin 5wtnullDeviation from Linear Phase Use electrical delay to remove linear portion of phase responseLinear electrical length added+yieldsFrequency(Electrical delay function)FrequencyRF filter responseDeviation from linear phaseFrequencyLow resolutionHigh resolutionnullGroup DelayFrequencyGroup delay rippleAverage delaytoPhasefDfFrequencyDwwgroup-delay ripple indicates phase distortion average delay indicates electrical length of DUT aperture of measurement is very importantnullWhy Measure Group Delay?Same p-p phase ripple can result in different group delaynullCharacterizing Unknown DevicesUsing parameters (H, Y, Z, S) to characterize devices: gives linear behavioral model of our device measure parameters (e.g. voltage and current) versus frequency under various source and load conditions (e.g. short and open circuits) compute device parameters from measured data predict circuit performance under any source and load conditionsH-parameters V1 = h11I1 + h12V2 I2 = h21I1 + h22V2Y-parameters I1 = y11V1 + y12V2 I2 = y21V1 + y22V2Z-parameters V1 = z11I1 + z12I2 V2 = z21I1 + z22I2nullWhy Use S-Parameters?relatively easy to obtain at high frequencies measure voltage traveling waves with a vector network analyzer don't need shorts/opens which can cause active devices to oscillate or self-destruct relate to familiar measurements (gain, loss, reflection coefficient ...) can cascade S-parameters of multiple devices to predict system performance can compute H, Y, or Z parameters from S-parameters if desired can easily import and use S-parameter files in our electronic-simulation toolsnullMeasuring S-Parameters IncidentTransmittedS21S11Reflectedb1a1b2a2=0DUTForwardIncidentTransmittedS12b2a2ba1=0DUTReverse1nullEquating S-Parameters with Common Measurement TermsS11 = forward reflection coefficient (input match) S22 = reverse reflection coefficient (output match) S21 = forward transmission coefficient (gain or loss) S12 = reverse transmission coefficient (isolation)Remember, S-parameters are inherently complex, linear quantities -- however, we often express them in a log-magnitude formatCriteria for Distortionless Transmission Nonlinear NetworksCriteria for Distortionless Transmission Nonlinear Networks Saturation, crossover, intermodulation, and other nonlinear effects can cause signal distortion Effect on system depends on amount and type of distortion and system architecture nullMeasuring Nonlinear BehaviorMost common measurements: using a network analyzer and power sweeps gain compression AM to PM conversion using a spectrum analyzer + source(s) harmonics, particularly second and third intermodulation products resulting from two or more RF carriers nullWhat is the Difference Between Network and Spectrum Analyzers?.Measures known signalMeasures unknown signalsnullAgendaWhat measurements do we make? Network analyzer hardware Error models and calibration Example measurements AppendixnullGeneralized Network Analyzer Block DiagramnullSourceSupplies stimulus for system Swept frequency or power Traditionally NAs used separate source Most Agilent analyzers sold today have integrated, synthesized sourcesnullSignal SeparationTest PortDetectordirectional couplersplitterbridgemeasure incident signal for reference separate incident and reflected signalsnullDirectivityDirectivity is a measure of how well a coupler can separate signals moving in opposite directionsTest port(undesired leakage signal)(desired reflected signal)Directional CouplernullInteraction of Directivity with the DUT (Without Error Correction)Data MaxAdd in-phaseDeviceDirectivityReturn LossFrequency03060DUT RL = 40 dBAdd out-of-phase (cancellation)DeviceDirectivityData = Vector SumData MinnullDetector TypesTuned ReceiverScalar broadband (no phase information)Vector (magnitude and phase)nullBroadband Diode DetectionEasy to make broadband Inexpensive compared to tuned receiver Good for measuring frequency-translating devices Improve dynamic range by increasing power Medium sensitivity / dynamic range10 MHz26.5 GHznullNarrowband Detection - Tuned ReceiverBest sensitivity / dynamic range Provides harmonic / spurious signal rejection Improve dynamic range by increasing power, decreasing IF bandwidth, or averaging Trade off noise floor and measurement speed10 MHz26.5 GHznullComparison of Receiver Techniques< -100 dBm Sensitivity 0 dB-50 dB-100 dB-60 dBm Sensitivity Broadband (diode) detectionNarrowband (tuned-receiver) detectionhigher noise floor false responseshigh dynamic range harmonic immunityDynamic range = maximum receiver power - receiver noise floorDynamic Range and Accuracy Dynamic Range and Accuracy Dynamic range is very important for measurement accuracy!nullT/R Versus S-Parameter Test SetsnullProcessor / Displaymarkers limit lines pass/fail indicators linear/log formats grid/polar/Smith chartsnullInternal Measurement AutomationSimple: recall states More powerful: Test sequencing available on 8753/ 8720 families keystroke recording some advanced functions IBASIC available on 8712 family sophisticated programs custom user interfacesnullAgilent’s Series of HF Vector AnalyzersMicrowaveRF8510C series 110 GHz in coax highest accuracy modular, flexible pulse systems Tx/Rx module test8720ET/ES series 13.5, 20, 40 GHz economical fast, small, integrated test mixers, high-power amps8712ET/ES series 1.3, 3 GHz low cost narrowband and broadband detection IBASIC / LAN8753ET/ES series 3, 6 GHz highest RF accuracy flexible hardware more features Offset and harmonic RF sweepsnullAgilent’s LF/RF Vector AnalyzersE5100A/B 180, 300 MHz economical fast, small target markets: crystals, resonators, filters equivalent-circuit models evaporation-monitor-function option4395A/4396B 500 MHz (4395A), 1.8 GHz (4396B) impedance-measuring option fast, FFT-based spectrum analysis time-gated spectrum-analyzer option IBASIC standard test fixturesLFCombination NA / SAnullSpectrum Analyzer / Tracking GeneratorTracking generatorRF inTG outf = IFSpectrum analyzerIFLODUTKey differences from network analyzer: one channel -- no ratioed or phase measurements More expensive than scalar NA (but better dynamic range) Only error correction available is normalization (and possibly open-short averaging) Poorer accuracy Small incremental cost if SA is required for other measurementsnullAgendanullCalibration TopicsWhat measurements do we make? Network analyzer hardware Error models and calibration measurement errors what is vector error correction? calibration types accuracy examples calibration considerations Example measurements AppendixnullSystematic errors due to imperfections in the analyzer and test setup assumed to be time invariant (predictable) Random errors vary with time in random fashion (unpredictable) main contributors: instrument noise, switch and connector repeatability Drift errors due to system performance changing after a calibration has been done primarily caused by temperature variationMeasurement Error ModelingnullSystematic Measurement ErrorsABSourceMismatchLoadMismatchCrosstalkDirectivityDUTFrequency response reflection tracking (A/R) transmission tracking (B/R)RSix forward and six reverse error terms yields 12 error terms for two-port devicesnullTypes of Error Correctionresponse (normalization) simple to perform only corrects for tracking errors stores reference trace in memory, then does data divided by memory vector requires more standards requires an analyzer that can measure phase accounts for all major sources of systematic errorSHORTOPENLOADnullWhat is Vector-Error Correction?Process of characterizing systematic error terms measure known standards remove effects from subsequent measurements 1-port calibration (reflection measurements) only 3 systematic error terms measured directivity, source match, and reflection tracking Full 2-port calibration (reflection and transmission measurements) 12 systematic error terms measured usually requires 12 measurements on four known standards (SOLT) Standards defined in cal kit definition file network analyzer contains standard cal kit definitions CAL KIT DEFINITION MUST MATCH ACTUAL CAL KIT USED! User-built standards must be characterized and entered into user cal-kit Reflection: One-Port ModelReflection: One-Port ModelTo solve for error terms, we measure 3 standards to generate 3 equations and 3 unknownsAssumes good termination at port two if testing two-port devices If using port 2 of NA and DUT reverse isolation is low (e.g., filter passband): assumption of good termination is not valid two-port error correction yields better resultsnullBefore and After One-Port CalibrationnullTwo-Port Error CorrectionEach actual S-parameter is a function of all four measured S-parameters Analyzer must make forward and reverse sweep to update any one S-parameter Luckily, you don't need to know these equations to use network analyzers!!!nullCrosstalk: Signal Leakage Between Test Ports During TransmissionCan be a problem with: high-isolation devices (e.g., switch in open position) high-dynamic range devices (some filter stopbands) Isolation calibration adds noise to error model (measuring near noise floor of system) only perform if really needed (use averaging if necessary) if crosstalk is independent of DUT match, use two terminations if dependent on DUT match, use DUT with termination on outputnullErrors and Calibration Standards Convenient Generally not accurate No errors removedEasy to perform Use when highest accuracy is not required Removes frequency response errorFor reflection measurements Need good termination for high accuracy with two-port devices Removes these errors: Directivity Source match Reflection tracking Highest accuracy Removes these errors: Directivity Source, load match Reflection tracking Transmission tracking Crosstalk UNCORRECTED RESPONSE 1-PORT FULL 2-PORTDUTDUTDUTDUTENHANCED-RESPONSECombines response and 1-port Corrects source match for transmission measurementsnullCalibration Summaryerror cannot be corrected*enhanced response cal corrects for source match during transmission measurementserror can be correctednullReflection Example Using a One-Port CalLoad match: 18 dB (.126).158(.891)(.126)(.891) = .100nullUsing a One-Port Cal + AttenuatorLow-loss bi-directional devices generally require two-port calibration for low measurement uncertaintynullTransmission Example Using Response CalRL = 14 dB (.200)RL = 18 dB (.126)Thru calibration (normalization) builds error into measurement due to source and load match interactionCalibration Uncertainty = (1 ± S L) = (1 ± (.200)(.126) = ± 0.22 dBnullFilter Measurement with Response CalSource match = 14 dB (.200)1(.126)(.158) = .020(.158)(.200) = .032(.126)(.891)(.200)(.891) = .020Measurement uncertainty = 1 ± (.020+.020+.032) = 1 ± .072 = + 0.60 dB - 0.65 dB DUT 1 dB loss (.891) 16 dB RL (.158)Load match = 18 dB (.126)nullMeasuring Amplifiers with a Response CalTotal measurement uncertainty: +0.44 + 0.22 = + 0.66 dB -0.46 - 0.22 = - 0.68 dBMeasurement uncertainty = 1 ± (.020+.032) = 1 ± .052 = + 0.44 dB - 0.46 dB1(.126)(.158) = .020 DUT 16 dB RL (.158)(.158)(.200) = .032Source match = 14 dB (.200)Load match = 18 dB (.126)nullFilter Measurements using the Enhanced Response CalibrationMeasurement uncertainty = 1 ± (.020+.0018+.0028) = 1 ± .0246 = + 0.211 dB - 0.216 dBTotal measurement uncertainty: 0.22 + .02 = ± 0.24 dBCalibration Uncertainty = Effective source match = 35 dB!Source match = 35 dB (.0178)1(.126)(.158) = .020(.126)(.891)(.0178)(.891) = .0018 DUT 1 dB loss (.891) 16 dB RL (.158)Load match = 18 dB (.126)(.158)(.0178) = .0028(1 ± rS rL) = (1 ± (.0178)(.126) = ± .02 dBnullUsing the Enhanced Response Calibration Plus an AttenuatorMeasurement uncertainty = 1 ± (.006+.0005+.0028) = 1 ± .0093 = ± 0.08 dBSource match = 35 dB (.0178)1(.0366)(.158) = .006(.0366)(.891)(.0178)(.891) = .0005 DUT 1 dB loss (.891) 16 dB RL (.158)Effective load match = (.316)(.316)(.126) + .024 = .0366 (28.7dB)(.158)(.0178) = .002810 dB attenuator (.316) SWR = 1.05 (.024 linear or 32.4 dB)Analyzer load match =18 dB (.126)nullCalculating Measurement Uncertainty After a Two-Port CalibrationTransmission uncertainty= 0.891 ± .0056 = 1 dB ±0.05 dB (worst-case)Reflection uncertainty= 0.158 ± .0088 = 16 dB +0.53 dB, -0.44 dB (worst-case)nullResponse versus Two-Port CalibrationCH1 S21&Mlog MAG1 dB/REF 0 dB CorCH2 MEMlog MAGREF 0 dB1 dB/ CorUncorrectedAfter two-port calibrationSTART 2 000.000 MHzSTOP 6 000.000 MHzx212After response calibrationMeasuring filter insertion lossnullVariety of modules cover 30 kHz to 26.5 GHz Six connector types available (50  and 75 ) Single-connection reduces calibration time makes calibrations easy to perform minimizes wear on cables and standards eliminates operator errors Highly repeatable temperature-compensated terminations provide excellent accuracyECal: Electronic Calibration (85060/90 series)Microwave modules use a transmission line shunted by PIN-diode switches in various combinationsnullAdapter ConsiderationsTerminationAdapterDUT Coupler directivity = 40 dBleakage signaldesired signalreflection from adapterAPC-7 calibration done hereDUT has SMA (f) connectorsnullCalibrating Non-Insertable DevicesWhen doing a through cal, normally test ports mate directly cables can be connected directly without an adapter result is a zero-length through What is an insertable device? has same type of connector, but different sex on each port has same type of sexless connector on each port (e.g. APC-7) What is a non-insertable device? one that cannot be inserted in place of a zero-length through has same connectors on each port (type and sex) has different type of connector on each port (e.g., waveguide on one port, coaxial on the other) What calibration choices do I have for non-insertable devices? use an uncharacterized through adapter use a characterized through adapter (modify cal-kit definition) swap equal adapters adapter removalnullSwap Equal Adapters MethodDUT1. Transmission cal using adapter A.2. Reflection cal using adapter B.3. Measure DUT using adapter B.Port 1Port 2Adapter AAdapter BAdapter BDUTAccuracy depends on how well the adapters are matched - loss, electrical length, match and impedance should all be equalnullAdapter Removal Calibratio