Time-domain Techniques in EMI Measuring Receivers. Technical and Standardization Requirements

Transcription

1 Time-domain Techniques in EMI Measuring Receivers Technica and Standardization Requirements

2 CISPR = Huge, Sow, Compex, CISPR = Internationa Specia Committee on Radio Interference Technica committee within Internationa Eectrotechnica Commission (IEC) CISPR was estabished in 1933 Today CISPR is one of the 95 technica committees of the IEC 7 sub-committees: A,B,D,F,H,I,S and 15 working groups CISPR has iaisons and specia working arrangements with technica committees of the IEC, ISO, CENELEC, CEPT, ETSI and ECMA. CISPR is composed of the foowing Member bodies: Nationa Committees of the IEC (IEC has 59 fu members and 21 associated members) European Broadcasting Union (EBU) European Teecommunication Standards Institute (ETSI) Internationa Conference on Large Eectric Systems (CIGRE) Internationa Amateur Radio Union (IARU) Internationa Teecommunications Union, Radio Sector (ITU-R) Internationa Teecommunications Union, Teecom Sector (ITU-T) Thousands of peope, committees and documents to expain What to test / against which imits How to test, environment and equipment February 2012 CISPR 16 FFT-based measuring receiver 2

3 CISPR Pubication Leves CISPR pubications are structured into 3 eves Specia or specific conditions appy, supersedes the generic standards Appy to everything not covered by product standard Foundation items that appy to everyone February 2012 CISPR 16 FFT-based measuring receiver 3

4 CISPR 11: Industria, scientific and medica (ISM) radio-frequency equipment - Eectromagnetic disturbance characteristics - Limits and methods of measurement. CISPR 22: Information technoogy equipment - Radio disturbance characteristics - Limits and methods of measurement CISPR 24: Information technoogy equipment - Immunity characteristics - Limits and methods of measurement. IT / ISM CISPR 12: Vehices, boats and interna combustion engine driven devices - Radio disturbance characteristics - Limits and methods of measurement for the protection of receivers except those instaed in the vehice/boat/device itsef or in adjacent vehices/boats/devices. CISPR 25: Vehices, boats and interna combustion engines - Radio disturbance characteristics - Limits and methods of measurement for the protection of on-board receivers" autos CISPR 14: Eectromagnetic compatibiity - Requirements for househod appiances, eectric toos and simiar apparatus - Part 1: Emission / Part 2: Immunity - Product famiy standard. appiances and toos CISPR 16: Specification for radio disturbance and immunity measurement apparatus and methods Part 1: Radio disturbance and immunity measuring apparatus Part 2: Methods of measurement of disturbances and immunity Part 3: Reports and recommendations of CISPR Part 4: Uncertainties, statistics and imit modeing meas devices February 2012 CISPR 16 FFT-based measuring receiver 4

5 Structure of Basic Standard CISPR 16 (2010) CISPR 16-1 Radio disturbance and immunity measuring apparatus CISPR CISPR CISPR Measuring apparatus Anciary equipment Conducted disturbances Anciary equipment Disturbance power CISPR Anciary equipment Radiated disturbances CISPR Antenna caibration test sites for 30 to 1000 MHz *in deveopment* *CISPR *Antenna caibration* CISPR 16-2 Method of measurement of disturbances and immunity CISPR CISPR CISPR Conducted disturbance measurements Measurement of disturbance power Radiated disturbance measurements CISPR Immunity measurements CISPR In situ measurements of disturbing emissions produced by physicay arge equipment CISPR 16-3 CISPR technica reports CISPR 16-4 Uncertainties, statistics and imit modeing CISPR CISPR CISPR Uncertainties in standardized EMC tests Measurement instrumentation uncertainty Statistica considerations in the determination of EMC compiance of mass-produced products CISPR Statistics of compaints and a mode for the cacuation of imits for the protection of radio services CISPR Conditions for the use of aternative test methods February 2012 CISPR 16 FFT-based measuring receiver 5

6 FFT Type Receivers- Can You Use Them? Amendment 1 to CISPR (3rd Ed.) Use of FFT-based measurement instruments for compiance measurements Specific requirements for FFT-based measuring instruments The standard was pubished on 21 June 2010 Appicabiity CISPR 13:2001 (Radio + TV) Appicabe since CISPR 32:2012 (Mutimedia) Appicabe since CISPR 15:201x (Lighting) Pubication of Ed. 8 expected in 2012 CISPR 12:201x (Automotive) Pubication of Ed. 7 expected in 2013 CISPR 25:201x (Automotive) Pubication of Ed. 4 expected in 2013 Not appicabe (not referenced) for CISPR 11:2009, CISPR 12:2007, CISPR13:2009 (5th Ed.), CISPR 14-1:2005, CISPR 15:2005, CISPR 22:2005 (5th Ed.), CISPR 22:2008 (6th Ed.) and CISPR 25:2008 February 2012 CISPR 16 FFT-based measuring receiver 6

7 FFT Type Receivers- Can You Use Them? Rea Answer is yes, use the FFT mode if you have it Verify resuts FFT vs Swept on goden device Know the time domain characteristics of the DUT At east use it during initia scan for time savings Can aways use cassic method for QP / Fina meas February 2012 CISPR 16 FFT-based measuring receiver 7

8 CISPR A1 (3rd Ed.) Measurement receiver CISPR compiant measurements Definition of measuring receiver added: instrument such as a tunabe votmeter, an EMI receiver, a spectrum anayzer or an FFT-based measuring instrument, with or without preseection, that meets the reevant parts of this standard Specific requirement for FFT-based measuring instruments for EMI measurements, FFT-based measuring instruments sha sampe and evauate the signa continuousy during the measurement time February 2012 CISPR 16 FFT-based measuring receiver 8

9 Amendments to CISPR 16-2-x Amendments to CISPR , and incude measurement methods for FFT-based receivers (6/7/2010) The duration of a disturbance must be known Can be measured using time domain outputs Zero span Oscioscope on IF output Time Domain output of FFT Minimum measurement times Tabe for CW signas has been added Same requirements as for scanning receivers February 2012 CISPR 16 FFT-based measuring receiver 9

10 CISPR 16 FFT-based measuring receivers Amendments to CISPR 16-2-x FFT-based instruments (may) combine the parae cacuation at N frequencies and a stepped scan frequency range of interest is subdivided into severa segments, which are measured sequentiay The scan time T scan is T m is to be seected onger than the puse repetition interva T p February 2012 CISPR 16 FFT-based measuring receiver 10

11 Fourier Transforms Frequency Time Domain The discrete Fourier transform (DFT) is a numerica mathematica method that cacuates the spectrum for a periodic signa The fast Fourier transform (FFT) is an efficient agorithm to compute the DFT using symmetry and repetition properties FFT is much faster than DFT due to reduced number of mutipications EMI signas incude both periodic and transient signas and noise Singe FFT cacuation wi not be sufficient to mode the EMI receiver Short-time FFT (STFFT) with a Gaussian window function is used Shows a discretization in both the frequency and time domains IF bandwidth requirements in the frequency domain are ideay met (Fourier transform of a Gauss function in the time domain is a Gauss function in the frequency domain) Gaussian window in time domain minimizes the eakage effects February 2012 CISPR 16 FFT-based measuring receiver 11

12 CISPR 16 FFT-based measuring receivers FFT-based receivers digita signa processing Frequency domain Frequency range to be measured is sub-divided in consecutive frequency segments and fitering Time domain Tempora samping of the fitered signas with high samping rate/resoution F(s) f(t) Discrete Fourier transform (DFT) Signa transformation of the fitered signas from time domain to frequency domain Frequency domain Merging the spectra distributions of a partia frequency ranges February 2012 CISPR 16 FFT-based measuring receiver 12

13 FFT Cacuations Not Perfect Exact cacuation of the frequency spectrum woud require Periodic / Repeating signas Measurement time equa to integer mutipe of the signa period Disturbance characteristic of the signa is unknown The frequency of the signa is not known The signa might not be periodic The measurement time cannot be set as integer mutipe of the signa period Spectra ines can exist between two discrete DFT frequency bins Signa distortions appear eakage effect (signa spectrum becomes more wider), picket fence effect (ampitude error sine wave signas) Ampitude error for isoated puse and a sequence of puses February 2012 CISPR 16 FFT-based measuring receiver 13

14 FFT Errors Leakage Effect Convoution with window function yieds wider spectrum i.e. shows additiona spectra components Sideobes (referred to as eakage effect) These sideobes shoud be suppressed by at east 40 db A suitabe windowing reduces the eakage effect (Gauss, Kaiser-Besse) Rectanguar window and magnitude of the Fourier transform 1) Gaussian window and magnitude of the Fourier transform, σ=2 1) Sideobes when using rectanguar window or Gaussian window 1) Timan Butz, Fouriertransformation fuer Fußgaenger, ISBN February 2012 CISPR 16 FFT-based measuring receiver 14

15 FFT Errors Picket Fence Effect The FFT cacuates a discrete ine spectrum at the frequency bins If the samped sine wave signa is at a frequency that doesn t aign with a cacuated frequency point an ampitude error appears The ampitude error is known as picket fence effect Like stepped-frequency scan with wide IF bandwidth vs step size time domain frequency domain window width February 2012 CISPR 16 FFT-based measuring receiver 15

16 FFT Errors Singe Puse Errors Measurement must be ong enough to capture singe puses Sampe/cacuate process must be gapess during the meas time Without time domain overap ampitude / detection probems An overapping factor of >75% in the time domain is necessary to meet the puse ampitude specification of CISPR February 2012 CISPR 16 FFT-based measuring receiver 16

17 FFT Errors Puse Sequences 0% overap 25% overap Overapping aso fixes puse sequence ampitude errors Exampe shows the recacuated IF signa for different overapping factors 75% overap 90% overap February 2012 CISPR 16 FFT-based measuring receiver 17

18 R&S EMI Instruments with FFT Functions FSW FSV / FSVR ESU ESR February 2012 CISPR 16 FFT-based measuring receiver 18

19 EMI Receivers with FFTs Heterodyne EMI receivers with FFT appied to the wideband IF signa offer the foowing advantages: High dynamic from 16-bit A/D converter Frequency range not imited by the Nyquist criterion (Mixers + FFT) Long maximum dwe time and high resoution (16k FFT ength) Use of receiver preseection, preampifier, RF-attenuation and detectors Use the best mode for test: standard receiver spectrum anayzer IF Receiver + Anayzer FFT scan Rea-Time Power meter EMI Receiver R&S ESR February 2012 CISPR 16 FFT-based measuring receiver 19

20 Bock Diagram (ESR) Receiver preseection and mixer Wideband IF fiter 20,4 MHz ADC Resampe UMOD RAM Main processor Wideband switchabe IF bandwidth preceding the A/D converter (ADC) 16-bit / 128 MHz samping ADC in combination with the preseector yieds high dynamic range to fufi CISPR Resamper for data reduction where needed (at narrower RBW) Universa digita modue (UMOD) saves the data in the 32 M-words RAM for measurement times up to 100 s without any gap (depending on RBW) Main processor accesses the RAM, appies a Gaussian window to the time domain signa (to avoid eakage effects) and cacuates the FFT February 2012 CISPR 16 FFT-based measuring receiver 20

21 EMI receiver with FFT appied to the wideband IF EMI Test Receiver R&S ESU A f start window width 7 MHz Gaussian-type fitering Maximum 90% overapping in time domain windowing (Gauss) t meas DFT ADC 7 MHz f "virtua step width" = ¼ x RBW t response inearisation t resut t "weighting" (detectors) f t meas f February 2012 CISPR 16 FFT-based measuring receiver 21

22 Minimizing the FFT Errors Leakage Effect Capturing an impuse-type disturbance signa by an Gaussian-type FFT window reduces the signa ampitude at the window edges Overap of the window function in the time domain minimizes this error and ensures that no impuse-type disturbance signa is missed R&S ESU offers two different settings for time-domain scan Auto CW mode 20% overap in the time domain narrowband signas are anayzed in the shortest possibe time Auto Puse mode approx. 90% overap in the time domain for broadband-impusive and mixed signas ensures that even very short impuse signas at the edge of the Gaussian-type time-domain window are cacuated without significant ampitude error t t meas DFT ADC February 2012 CISPR 16 FFT-based measuring receiver 22

23 Minimizing the FFT Errors Pickett Fence The FFT agorithm incudes Cacuations using IF bandwidths for commercia and miitary standards A virtua step size that considers the spacing between two adjacent frequency bins The step size is 25% of the seected IF bandwidth Optimum vaue in terms of the amount of samped data Optimum compromise between reduction of ampitude and measurement speed Lower picket fence error than conventiona stepped scan "virtua step width" = ¼ x RBW 7 MHz t meas f February 2012 CISPR 16 FFT-based measuring receiver 23

24 Minimizing the FFT Errors Time-Domain Scan versus Stepped Scan R&S instruments provide both scan methods using the same hardware and firmware Measurement was done using a puse generator for CISPR bands C and D <0.5 db Overa frequency response (Detector = Max.Peak) shows that the differences between the two scan modes are negigibe Trace 1: Time-domain scan (bue) Trace 2: Stepped frequency scan (back) February 2012 CISPR 16 FFT-based measuring receiver 24

25 Why Bother with FFT Scans? Measurement Times (peak detector) FFT vs stepped scan February 2012 CISPR 16 FFT-based measuring receiver 25

26 Why Bother with FFT Scans? Comparison of typica Measurement Times (peak detector) R&S ESU (Stepped Scan) 9 to 150 khz; RBW 200 Hz; f= 80 Hz; T m = 20 ms: 74 s 150 khz to 30 MHz; RBW 9 khz; f= 4 khz; T m = 20 ms: 155 s 30 to 1000 MHz; RBW 120 khz; f= 40 khz; T m = 10 ms: 247 s 30 to 1000 MHz; RBW 9 khz; f= 4 khz; T m = 10 ms: s R&S ESU (Time-domain Scan) 9 to 150 khz; RBW 200 Hz; f= 50 Hz; T m = 20 ms: <1 s 150 khz to 30 MHz; RBW 9 khz; f= 2,25 khz; T m = 20 ms: 3 s 30 to 1000 MHz; RBW 120 khz; f= 30 khz; T m = 10 ms: 13 s 30 to 1000 MHz; RBW 9 khz; f= 2,25 khz; T m = 10 ms: 20 s Note: The frequency step of the FFT is RBW/4 to reduce the picket fence effect. Step mode = Auto Puse with an overap of >90% of the FFT window in the time domain. February 2012 CISPR 16 FFT-based measuring receiver 26

27 Summary: Benefits of FFT / Time Domain Scans Huge time savings vs conventiona frequency (stepped) scans 80 seconds vs. 9 hours for MHz / 120kHz IF / QP scan.82 s vs s for MHz / 9kHz IF / peak detector QP, CISPR-AV, RMS-AV are appicabe in time-domain scan modes Pus sti have cassic modes to compare if in doubt Preseection preserves the fu dynamic range for band of interest Measurement time up to 100 s without any gaps UMOD chip sti aows zoom on any time window inside the capture Measurement of puses without significant error 90% overap (Auto Puse) of the window function and gapess samping February 2012 CISPR 16 FFT-based measuring receiver 27

28 Reatime anaysis (option on ESR) Provides new insights for EMC diagnostics Spectrogram for seamess spectrum dispay in the time domain Persistance mode Vauabe aid for examining signas that change over time. Impusive interferers are ceary distinguished from continuous interferers (see screen shot) Frequency mask trigger Responds to events in the spectrum. Comparison with a predefined frequency mask. Vioation of the frequency mask activates the trigger event February 2012 CISPR 16 FFT-based measuring receiver 28

29 Thank you for your interest! February 2012 CISPR 16 FFT-based measuring receiver 29