STSM: On-Site Emission Measurements

Transcription

1 On-Site Emission Measurements Based on Reverberation Chamber Techniques Robert Vogt, STSM - SP, Boras, Sweden

2 CONTENTS On-site emission measurements (Motivation) Test site classification Test site analysis Validation measurements Error analysis Conclusions Work in progress 2

3 1. MOTIVATION ON-SITE TESTING Heavy, complex equipment Cannot move to an EMC lab Measure on-site EMRP project goal 3

4 EMRP Standard laboratory Industrial environment Connection & Correlation Traceability Traceability Uncertainty Uncertainty Precision Precision Reliability Reliability Reflections Reflections Near Field Effects Near Field Effects Line Impedance Line Impedance Field Uniformity Field Uniformity NSA NSA CM Impedance CM Impedance EUT - Antenna Coupling EUT - Antenna Coupling Correlation Correlation - Mustafa ÇETİNTAŞ, 02/09/2013, Brugge, Belgium 4

5 ON-SITE TESTING - STANDARD METHOD IEC standard Rather simplified Based on anechoic chamber technique Line of sight component measurement Problems Influence of the environment Reflections, standing waves Measurement angles Unknown EUT radiation characteristic Obstacles blocking access to all angles Ambient noise presence 5

6 2. TEST SITE CLASSIFICATION ANECHOIC CHAMBER Faraday cage lined with absorbers Fully Anechoic Chamber Semi Anechoic Chamber Acts like free space No reflections (Q-factor 0) Only LOS component Additional ground reflection in SAC Dependent on EUT pattern 6

7 ANECHOIC CHAMBER Working volume calibration IEC Field uniformity 7

8 REVERBERATION CHAMBER Faraday cage with a stirrer High Q-factor Changing boundary condition Mode tuned / stirred Multipath environment Lots of reflections No LOS component Isotropic Statistically uniform Random polarizations Independent of EUT pattern 8

9 REVERBERATION CHAMBER Working volume calibration IEC Field uniformity Chamber loading (CLF, CVF, IL) 9

10 INDUSTRIAL ENVIRONMENT? Industrial environment? Anechoic Chamber? Reverberation Chamber 10

11 STD. ENVIRONMENT CLASSIFICATION 11

12 IMP. ENVIRONMENT CLASSIFICATION 12

13 ANECHOIC TEST SITE Anechoic environment Rather low Q Some metallic objects Resonant / lossy objects Apertures Multipath environment Some reflections Strong LOS component Not isotropic Not statistically uniform Not random polarizations Dependent on EUT pattern 13

14 ANECHOIC TEST SITE Statistical + deterministic data AC calibration impossible/difficult Influential setup parameters Insertion loss Over-/undermoded Lowest usable frequency Q-factors (Portable) stirrer efficiency Number of independent samples Ambient noise floor 14

15 SCATTERING TEST SITE Reverberant environment Rather high Q Plenty metallic objects No stirrer Resonant / lossy objects Apertures Multipath Plenty of reflections No LOS component Almost isotropic Almost statistically uniform Almost random polarizations Independent of EUT pattern 15

16 SCATTERING TEST SITE Only statistical data RC calibration possible Influential setup parameters Insertion loss Over-/undermoded Lowest usable frequency Q-factors Apertures, objects (losses) EUT loading Random walk technique efficiency Number of independent samples Ambient noise floor 16

17 IMP. ENVIRONMENT CLASSIFICATION Industrial environment Anechoic Chamber Reverberation Chamber?? Anechoic Test Site?? Scattering Test Site 17

18 AVAILABLE TEST SITES SP & INTA (partner) AC + SAC RC Office Workshop 18

19 RC STANDARD METHOD Goal: Evaluate the IEC standard method in imperfect reverberant environments Test site analysis Q-factor K-S GoF test Insertion loss IL Emission measurement Validation measurement using a known source Error analysis Random walk technique instead of a RC stirrer 19

20 3. TEST SITE ANALYSIS Tx: Omnidirectional, static discone antenna Rx: Directional PICA antenna VNA: Complex S frequency points between 300 MHz - 4 GHz Minimal LOS Tx/Rx coupling Crosspolarization Radiation pattern minima 50 spatial positions Randomly distributed Variable height > λ/2 distance 20

21 4. EVALUATION: POWER DELAY PROFILES Inspired by acoustics, decay time of sound pressure level INTA SAC MHz NLoS 600 MHz NLoS 800 MHz NLoS 1 GHz NLoS 2 GHz NLoS 4 GHz NLoS PDP (db) 30 PDP (db) 400 MHz 500 MHz 600 MHz 700 MHz 800 MHz 900 MHz 1 GHz INTA ATS T ime (s) x 10 6 INTA Small Reverberation Chamber T ime (s) PDP (db) PDP (db) 400 MHz 450 MHz 500 MHz 550 MHz 600 MHz 650 MHz 700 MHz 750 MHz 800 MHz 850 MHz 900 MHz 950 MHz 1 GHz MHz 500 MHz 600 MHz 700 MHz 800 MHz 900 MHz 1 GHz x 10 INTA STS T ime (s) x T ime (s) x

22 EVALUATION: Q-FACTOR INTA 22

23 EVALUATION: Q-FACTOR SP 5 10 SP comparison of Q factor in different environments 4 Q Factor 10 ac los ac nlos office los office nlos rc los rc nlos sac los sac nlos workshop los workshop nlos Frequency (GHz)

24 EVALUATION: GOF TEST K-S Goodness-of-Fit test for Rayleigh distribution (perfect RC) 24

25 EVALUATION: INSERTION LOSS Max + SNR - Statistics Avg + Statistics - SNR 25

26 TEST SITE ANALYSIS: CONCLUSIONS Evaluation Q-factor GoF test Insertion loss AC < Office < Workshop < RC Office/workshop do not behave like a standard RC, but they are actually pretty close.. Next step: validation & error analysis 26

27 4. VALIDATION Tx: Known source - comb generator (monopole) Rx: Directional PICA antenna Spectrum analyzer, CISPR settings 21 frequency points between 300 MHz and 1.3 GHz Minimal LOS Tx/Rx coupling Comb generator peaks Crosspolarization Radiation pattern minima 50 spatial positions Randomly distributed Variable height > λ/2 distance 27

28 VALIDATION, RAW POWER Raw power results Highest in RC Workshop Office All above noise < 20 db for W/O > 20 db for RC 28

29 VALIDATION, TRP Calculated TRP Total radiated power Lowest in RC Highest in office (max) (avg) Raw 29

30 VALIDATION, ERROR Absolute error W.r.t. RC (avg) Office > Workshop (avg) > (max) Positive Overestimation Raw TRP Error sources Ambient noise low SNR Poor field statistics 30

31 UNCERTAINTIES, FIELD STATS Uncertainties due to poor field statistics Empirical Imperfect environment Based only on the data RC theory not usable VNA frequency points Frequency stirring possible Generate more data for statistical analysis Central limit theorem Information also usable for power measurement data Same environment, same distributions, just rescaled 31

32 UNCERTAINTIES, FIELD STATS Combined TRPavg unc. From average values Narrow intervals High repeatability Combined TRPmax unc. From max values Wide intervals High spread 32

33 5. ERROR, SNR Uncertainties due to low SNR Maxima Single values Usually high above the noise Less prone to error Averages Calculated from many values Also values very close to noise More prone to error Generator SA Low SNR = high error Sig.gen. (dbm) RBW 120kHz RBW 10kHz

34 CONCLUSIONS Industrial environments can be quite reverberant Anechoic chamber, LOS methods Erroneous due to high amount of reflections Obstacles not allowing to measure from every direction High SNR due to strong LOS component Reverb chamber, no LOS methods Benefits from high amount of reflections EUT characteristic not important Low SNR due to no LOS component Tradeoffs Use maximal values Large spread of peaks wide intervals High SNR Noise error avoidable Use average values Low spread of averages narrow intervals Low SNR Noise error unavoidable 34

35 THANK YOU 35