2Copyright 2003 Schaffner EMC - All rights reserved Virtually every EMC laboratory has one or more ESD simulator.Almost none are equipped to verify the ESD simulators’ performance.We will cover:Verification techniques, including ISO, SAE, ANSI and IEC standardsProposed changes in the measurement setupPractical aspects of measurement setup and performanceLive demonstrationCopyright 2003 Schaffner EMC - All rights reserved
3What parameters must be measured? Tip voltageCurrent waveformPeakRiseCurrent at 30nsCurrent at 60nsTime Constant (air discharge, auto manf)Current derivative - ANSI Draft (gives indication of smoothness)Positive peakNegative peak
4Measuring tip voltage Measured at standard test levels: ±2kV, ±4kV, ±6kV, ±8kV, ±15kV and ±25kVMeasured using Electrometer or Giga-ohm meterMost standards don’t specify requirementsISO specifies 100 GOhm minimum input impedanceThe simulator’s tip voltage not affected by the measurementIf a Giga-ohm meter is used, the simulator must continuously chargethe high-voltage capacitor- Many older simulators provide an initial charge only, which can bleed off with time or with load
5Tip voltage measurement using Giga-ohm meter (Brandenburg Model 139D)
8How do we measure the current waveform? A low impedance shunt (ESD target) is used to represent adischarge into a large metallic objectThe shunt impedance is < 2.1 OhmsBlock diagram:ATTENUATOROSCILLOSCOPETARGETGROUND PLANECABLEOptional Attenuator for > 8 kV (20 dB)
9Typical ESD current measurement system NOTE:The reason a Faraday cage was written into the original standard was that the analog phosphorstorage oscilloscopes were generally susceptible to the high field energy produced by simulators.The digitizing oscilloscopes today are much more immune and the Faraday cage is no longer a must. You must confirm your measurement system is unaffected, however!
10Typical ESD current measurement system ESD measurement system at Schaffner, Switzerland.
11Typical ESD current measurement system (Agilent lab) 1.2m ground plane clamped to ESD table.
13Performing a contact discharge into the older ESD target Keytek MZ-15EC MiniZap Simulator.
14Target design history IEC 801-2: 1991 No longer referenced by any current ESD standardNo performance specificationsPoor design - lots of ringingIEC : 1995Referenced by virtually all current ESD standardsTransfer function “zero” at 5-6 GHzANSI C63.16 Draft 9Proposed new design (uses sm resistors and tapered transitions)Flat to 6GHz“Driving” adapter to evaluate high frequency performance
15IEC 801-2 target “ball tip” Old design is no longer specified
16IEC 61000-4-2 target Presently specified in standards The large flat disk tends to build up a pre-corona discharge, which slows the risetime and leads to variable results for air- discharge measurements.Example: EMCO CTC-3, and others
17ANSI C63.16 target Proposed design Example: Schaffner MD-102, Amplifier Research CTR-2, and others
18Old versus new ESD targets EMCO CTC-3 (left) Schaffner MD-102 (right).
19New target with “driving” adapter to measure transfer characteristics Schaffner MD-102
20ANSI C63.16 target specifications Reflection coefficient of target and adapter < 0.1Equivalent to VSWR < 1.22Insertion loss < 0.3dB up to 4 GHzVariation of attenuation of the target-attenuator-cable chain< ±0.3dB from DC to 1GHz (< ±3.51%)< ±0.8dB from 1GHz to 4GHz (< ±9.65%)
21Waveforms of IEC 801-2 target vs. ANSI target less HF ringingand shows truepeak shape
22Actual waveform measurement (Agilent 54855A, 1.5 GHz BW) Old target New Target
23Actual waveform measurement (Agilent 54855A, 6 GHz BW) Old target New Target
24Choosing attenuatorsTarget transfer function is ~1V/A when loaded by 50 OhmsContact mode peak current at 8kV is ~30AInput range of most oscilloscopes is < 10V in 50 Ohm modeTherefore, an attenuator is needed to reduce the signal level20dB is typically chosen for 10:1 ratioContact mode to 25kV may require additional attenuation
25Choosing attenuatorsLow power attenuators may damaged by the short term peak powerAttenuators are available with 1kW peak power ratingsUse an 18GHz attenuator with low SWR, < 1.25 to 8GHzThe attenuator accuracy requires that the entire chain be calibratedAccuracy variation dB Percentage%%%%%
26Choosing cables A low loss cable is required Cable length < 1m is required by most standardsDouble shielding is required by most standardsThe ANSI standard recommends RG 400RG 214 is twice the dia, 1/2 the loss and is commonly available
27Oscilloscopes - Bandwidth All standards require at least 1GHz bandwidthThe BW/risetime of the oscilloscope is the single most limitingfactor to accurately measure the pulse risetimeThe true risetime is related to the observed risetime as follows:The above correction is proposed in the ANSI draft standard andassumes a Gaussian rolloff in frequency response. However mostdigitizers use a sharper cutoff filter, 20dB/decade or higher.
28Oscilloscopes - Bandwidth How does bandwidth affect observed risetime? Let‘s assume a Gaussian rolloff
29Oscilloscopes - Sampling rate Single-shot sampling rate is the keyA fast-edge triangular peak requires fast sample rateRisetime of 800 ps from 10%-90% is 80% of waveform10 Gs/s = 100 ps/sample8 samples in 800 ps or 10%/sample!Since peak is symmetrical and somewhat rounded actualerror is < 5% (assumes a triangle shape)Effective sampling rate increased by capturing multiple shotsMust have stable waveformUseful for contact mode only - never for air dischargeShot to shot variation is low for most simulatorsShould be used for verification - not for calibration
30Shot-to-shot variation - 20 shots 33.3 A peakStd dev .425±0.64% ofpeak898 ps RiseStd dev 11.9±0.66% ofrisetime
31Oscilloscopes - Sampling rate SAE and ISO recommend 4Gs/s minimum 2 Gs/s A-16.0%10 Gs/s A-2.9%5 Gs/s A-3.8%20 Gs/s A
32Oscilloscopes – Interpolation - sin(x)/x ON or OFF? Interpolation ONInterpolation OFF2 Gs/s A-16.0%2 Gs/s A-9.6%5 Gs/s A-3.8%5 Gs/s A-2.6%
33Calibrating the target-attenuator-scope chain It is recommended that the DC transfer function of the entire chainbe measured as follows:Inject a known currentMeasure the resulting voltage at the oscilloscopeThe attenuation factor = Injected current / observed voltageAttenuation factor is used to correct waveform amplitudeATTENUATOROSCILLOSCOPETARGETGROUND PLANECABLECURRENT SOURCEOptional Attenuator for > 8 kV (20 dB)
34Other factors - Do’s and don’ts ShieldingDo we need it?Position of ground cableWill it affect waveform?Orientation of simulatorAutomatic MeasurementsMust use Min and Max values to calculate 10% and 90% pointsOther cablesKeep them well separated
35Oscilloscope shielding - Do we need it Oscilloscope shielding - Do we need it? Standards say yes, but probably not necessary - use distance testScope inside Faraday cageScopeat cornerof planeScopenext tosimulator
36Ground cable position Does affect results - peak, rise and duration 20 Gs/s A,891psNatural loop20 Gs/s A,926psLoop closerto plane
37Simulator orientation to target Does affect results - peak, rise and duration 20 Gs/s A,891psSimulator on axis20 Gs/s A,913psSimulator tip down 10º20 Gs/s A,945psTip down 30º
38Air discharge - What risetime/peak do you want? Approach speedand environmentalfactors will greatlyaffect results - notRepeatable!Obtaining a passingwaveform is amatter of patience!
39Measurement uncertainty The estimated bounds of the deviation of a measured quantity from itstrue valueList all the possible error sources and compute the uncertaintyUncertainty budget for each measured parameterStatement of confidence that can be placed in the value of uncertaintyDoes measured result truly fall within acceptable limits?National Association for Measurement and Sampling publicationNIS81, The Treatment of Uncertainty in EMC MeasurementsLink to CE-Mag site
40Target plane size ANSI - 1. 2m x 1. 2m, IEC 1. 5m x 1 Target plane size ANSI - 1.2m x 1.2m, IEC 1.5m x 1.5m, ISO - N/A, SAE - N/A20 Gs/s AMini TargetPlane20 Gs/s A1.2m2TargetPlane
41Demonstration equipment Simulator: Schaffner NSG 435 / Keytek Minizap MZ-15ECNew Target: Schaffner MD 102 (designed to new ANSI stnd)Old Target: Emco CTC-3 (designed to meet IEC stnd)Target Plane: Small sized plane for demo purposesAttenuator: Weinschel Model 2-20, 20dB, 5W, 1000W peakCable: RG-214 1mOscilloscope: Agilent Infiniium 54855A 6GHz BW, 20Gs/s scopeESD Monitor: Credence Technologies CTC034-3 (counts and beeps for each ESD event)
42Thank you for your attention Your feedback is welcome Greg SenkoBusiness Manager - EMC Test EquipmentSchaffner EMC(603)Ken Wyatt Sr. EMC Engr Hardware Test Mgr Agilent Technologies(719)