phase stability study of a hpr
DESCRIPTION
This slide set gives an overview of the typical phase stability tests that NMDG performs on a comb generator as candidate to become a harmonic phase reference (HPR).TRANSCRIPT
Phase Stability Studyof a
Harmonic Phase Reference
Jan 09
© Copyright 2009 2
Outline
● Introduction
● Harmonic phase reference (HPR) and Error Bounds
● Phase alignment
● Measurement Noise
● Warm-up time
● Phase stability study under different external conditions
● Summary
© Copyright 2009 3
Introduction
● Goal● Presenting the different studies, performed on comb generators to establish a
worst-case error bound on their phase characteristic
● Motivation● Accurately measuring the voltage and current at the input and output ports of a
nonlinear device using a Large-Signal Network Analyser (LSNA), which can be based on a Vector Network Analyser (VNA), e.g. ZVA24 [1]
● Requires additional calibrations on top of the relative calibration● Absolute Amplitude calibration: Calibrated Power Sensor ● Absolute Phase calibration: Calibrated Harmonic Phase Reference (HPR)
● The phase stability of the HPR impacts the phase accuracy of the LSNA measurements
© Copyright 2009 4
Harmonic Phase Reference
● What?● a comb generator including proper padding● driven by a synthesiser at frequency f
o
● What is important for the absolute phase calibration?● a fixed phase relationship between the combs● an arbitrary delay can always be applied to the phase characteristic
Comb generator
fo = 600 MHzInput pulse = sine wave
Output = pulse ns
[V]
GHz
Pulse Fundamental and harmonics
f0
2f0
© Copyright 2009 5
Contributions to the Error Bounds on the HPR Phase
● Systematic errors, introduced by the calibration process● Traceability to NIST has been established● Not discussed in this presentation
● Intrinsic device noise in Comb Generator● Always present in the measurements
● Phase stability impacted by external factors● Temperature variations● Variations in applied input power● Variations in applied bias voltage● Impact of the nonlinearities of the source● Impact due to load variations
© Copyright 2009 6
Measurement Setup to study HPR Stability
● Study of stability only requires a comparison between measurements● Using the MT4463 [2] which is proven to be very stable after warm-up time to
measure incident and reflected waves
● Observing the voltage waveforms a1 and b
2 under varying conditions and
studying the phase variation of b2
● Two Problems● Arbitrary delays in repeatable measurements of b2 require proper phase alignment
for correct phase characterisation● Measurement noise converts into phase uncertainty and hides phase error on HPR
● Need to estimate the phase error due to the measurement noise● This also includes the intrinsic noise generated by the comb generator
Pin
a1
50Ω
Combgenerator
Reference planes
fo MT4463
b1
b2
a2
© Copyright 2009 7
Elimination of Arbitrary Delay
● The error bound on the phase as function of frequency under varying conditions is influenced by the method to eliminate the arbitrary delay
● Challenge: for each measurement, determine the delay such that the phase variation across comparable measurements is minimised for all spectral components
● Different methods to eliminate the delay● Apply a delay to the signal such that the phase of the fundamental frequency
becomes zero● Apply a delay to the signal such that the difference between the signals is
minimised for all spectral components● Use system identification for optimal extraction of delay [3],[4]
© Copyright 2009 8
5 10 15 20Harmonic Index
0.05
0.1
0.15
0.2
0.251 degrees
Phase Alignment
To determine phase error specifications, NMDG applies the latter (red) method
Blue = delay signals such that phase of fundamental is zeroRed = delay signals to minimise the error for all spectral components
© Copyright 2009 9
Estimation of Phase Error Due to Measurement System
● Measurement noise converts into phase error
● This phase uncertainty adds up to the phase variations of the HPR
Measurement Noise
One Spectral Componentof comb generator
LargePhase deviation
LargeSpectral Component
SmallSpectral Component
SmallPhase deviation
AmplitudeUncertainty
5 109 11010 1.5 1010 21010 2.5 1010Hz
0.4
0.2
0.2
0.4
0.6
0.8
11 degrees
1000 comb generator measurementsAfter proper phase alignment: Total phase uncertainty Phase uncertainty due to measurement noise
1 - uncertainty
ConclusionPhase variation of comb generator cannotbe distinguished from measurement noise
© Copyright 2009 10
5 10 15 20 25 30 35 40Time min
3
2
1
1
2
3Phase deg
Minimal Warm-up Time
● Comb Generator● Nominal input power: +10 dBm, nominal biasing voltage: +5V, f0 : 600 MHz
● Measurement Procedure● Apply nominal input power and nominal biasing voltage● Perform 100 measurements after {0, 1, 3, 6, 10, 15, 21, 28, 36, 45, 55} minutes
● Calculate an “average pulse” using the proper alignment at each time slot● Compare the 10 “average pulses” against the “average pulse” after 55 minutes after
proper alignment
● Conclusion● After 3 minutes warm-up● Phase deviation < 1 degree
Cold Start
2f0
Phase variations of each harmonic vs. time
f0
© Copyright 2009 11
5 10 15 20 25GHz
1
0.75
0.5
0.25
0.25
0.5
0.75
11 degreesLong term sensitivity 3 days,4 measurements
Long – Term Stability
● Comb Generator● Nominal input power: +10 dBm, nominal biasing voltage: +5V, f
0 : 600 MHz
● Measurement Procedure● Apply input power● Wait 30 minutes● 1000 measurements with interval of 5 seconds
● Calculate an “average pulse” using the proper alignment● Repeat 4 times, spread across 3 days
● This results in 4 “average pulses” being compared after alignment with each other
● Conclusion● Phase Variations of
0.5 degrees are observable
© Copyright 2009 12
9.25 9.5 9.75 10 10.25 10.5 10.75 11Input power dBm
3
2
1
1
2
3Phase deg
Power Sensitivity
● Comb Generator● Nominal biasing voltage: +5V, f
0 : 600 MHz
● Measurement Procedure● Applying nominal input power and nominal biasing voltage● Wait 30 minutes● Stepping input power from +9 dBm to +11 dBm in steps of 0.1 dBm (21 steps)● For each power setting: 1000 measurements
● Calculate an “average pulse” using the proper alignment● Compare the phase of each “average pulse” with “average pulse” for nominal input
power of +10 dBm
● Conclusion● Power variation needs to be
limited to +/- 0.25 dBm
to limit phase variation
within 1 degree
2f0
Phase variations of each harmonic vs. input power
© Copyright 2009 13
4.8 4.9 5.1 5.2Bias voltage V
6
4
2
2
4
6Phase deg
Bias Sensitivity
● Comb Generator● Nominal input power: +10 dBm, f
0 : 600 MHz
● Measurement Procedure● Applying nominal input power and bias voltage● Wait 30 minutes● Stepping biasing voltage from +4.8 to +5.2 Volt in steps of 0.05 Volt (9 steps)● For each bias setting: 1000 measurements
● Calculate an “average pulse” using the proper alignment● Compare the phase of each “average pulse” with “average pulse” for nominal biasing
voltage of +5 Volt
● Conclusion● Bias voltage variation
needs to be controlled
within 50 mV due to
phase variation of second harmonic
2f0
Phase variationsof each harmonicvs. bias voltage
© Copyright 2009 14
Source and Load Sensitivity
● Using the comb generator as HPR requires proper padding● Minimising mismatch effects● Limiting the peak voltage to avoid compression of LSNA receivers● The comb generator used in this study is padded with 20 dB
● Source sensitivity● The HPR is driven by the synthesiser used in the LSNA setup● The phase characteristic can be sensitive to harmonic distortion (HD)● It is assumed that the synthesiser has a HD of -30 dBc in worst case
● Load sensitivity● The HPR is connected to different test sets● A test set presents a certain load match, which can impact the phase
characteristic● It is assumed that the test set has a return loss of 20 dB or better
● A special measurement setup is used to inject signals at fundamental and harmonics to simulate different harmonic distortions and return losses
© Copyright 2009 15
Source and Load Sensitivity: Measurement Setup
Pin a1
b1
b2
a250ΩRF1
fo
RF2
Vbias
Combgenerator
HPR
Second source (RF2) to inject a waveform
Source sensitivity: inject waveform at 2f0 and 3f
0
Load sensitivity: inject waveform at f0, 2f
0, 3f
0
© Copyright 2009 16
Source Sensitivity
● Comb Generator● Nominal input power: +10 dBm, nominal biasing voltage: +5V, f0 : 600 MHz
● Measurement Procedure● Applying nominal input power and nominal biasing voltage● Wait 30 minutes● Injecting different powers at second and third harmonic while rotating the phase
to simulate all possible harmonic distortion situations up to -30 dBc● Processing of measurements in the same way as HPR under nominal conditions
● Calculate an “average pulse” using the proper alignment for all measurements● Determine the uncertainty of all pulses, compared to the “average pulse”
● Conclusion● Harmonic Distortion of -30 dBc
can lead to a phase deviation of
1 degree
2.5 5 7.5 10 12.5 15 17.5 20GHz
0.5
1
1.5
2
2.5
3
3.5
41 degrees Source sensitivity
2f0
Phase uncertaintydue to measurementnoise
Phase deviationdue to harmonicdistortion
© Copyright 2009 17
2.5 5 7.5 10 12.5 15 17.5 20Hz
1
2
3
4
51 degrees Load sensitivity
Load Sensitivity
● Comb Generator● Nominal input power: +10 dBm, nominal biasing voltage: +5V, f0 : 600 MHz
● Measurement Procedure● Applying nominal input power and nominal biasing voltage● Wait 30 minutes● Injecting different powers at fundamental, second and third harmonic while
rotating the phase to simulate all possible load mismatches up to 20 dB return loss
● Same processing as for source sensitivity
● Conclusion● Mismatch can lead to some small
phase deviation
2f0
Phase uncertaintydue to measurementnoise
Phase deviationdue to mismatch
© Copyright 2009 18
Summary
● Discuss the harmonic phase reference and its phase specification● Discuss the phase alignment and error contributions to the phase● Determine the minimal warm-up time● Discuss the sensitivity to external factors
● Power sensitivity● Bias sensitivity● Sensitivity to harmonics generated by the synthesiser● Sensitivity to different load matches
● Explain the measurement procedure for each sensitivity study
● For the commercial Phase Reference, these procedures are repeated on a dense grid
For more technical information, please contact [email protected]
© Copyright 2009 19
References
[1] NM300, ZVxPlus, http://www.nmdg.be/files/ProductNoteNM300.pdf
[2] MT4463 Datasheet, http://www.nmdg.be/files/MT4463DataSheet.pdf
[3] Frans Verbeyst, Rik Pintelon, Yves Rolain, Johan Schoukens and Tracy S. Clement, “System Identification Approach Applied to Drift Estimation”,Instrumentation and Measurement Technology Conference, IMTC 2007,Warsaw, Poland, May 1-3, 2007.http://www.nmdg.be/library/contributions/IM-7123.pdf
[4] Frans Verbeyst, “Contribution to Large Signal Network Analysis”,PhD dissertation, September 2006.http://www.nmdg.be/index_jump.html?goto=library_center