This full text paper was peer-reviewed at the direction of IEEE Instrumentation and Measurement Society prior to the

acceptance and publication.

Calibration Standard for Impulse Energy of

Defibrillator and Defibrillator Analyzer
Wei ZHAO Haiming SHAO Huanghui ZHANG
National Institute of Metrology National Institute of Metrology Fujian Metrology Institute
Beijing, China Beijing, China Fuzhou, China
[email protected] [email protected] [email protected]

Abstract—Impulse energy is one of the most important analyzers and their experiment result shows that the system
parameters that need to be calibrated for defibrillator and can achieve an energy measuring uncertainty of better than
defibrillator analyzer. The measurement accuracy of 0.7 %.[4] Currently, the measurement accuracy of defibrillator
commercial defibrillator analyzer could reach ± 1 % of reading analyzer, such as FLUKE 7000DP, could reach ± 1 % of
plus 0.1 Joule nowadays. So, it placed a greater demand on the reading plus 0.1 Joule. So, it placed a greater demand on the
measurement standard used for the calibration. In the National
measurement standard used for the calibration of the
Institute of Metrology (NIM) China, we had developed a set of
defibrillator and defibrillator analyzer.
national measurement standard. The calibration uncertainty on
the defibrillator analyzer is evaluated to be better than 0.06% of In National Institute of Metrology (NIM) China, we had
reading plus 0.06 Joule. In this manuscript, the components of developed a set of national measurement standard. In 2015,
the system and the uncertainty evaluation of the measurement we had introduced the characteristic of the components and
standard is introduced. the consistency between two impulse energy calculation
algorithm.[5] Then we calibrated the components with plenty
Keywords—Defibrillator, Defibrillator analyzer, Impulse
of experiments for system uncertainty evaluation and
Energy, Calibration, Uncertainty
improvement. We had also set up a traceability system to
I. INTRODUCTION calibrate impulse energy of defibrillator and defibrillator
analyzer, as Fig 1 illustrated. Currently, the calibration
Medical devices are vital parts of modern medical care. In uncertainty on defibrillator analyzer is evaluated to be better
many cases they are associated with the quality of health care. than 0.06% of reading plus 0.06 Joule.
Defibrillator is a medical device designed to resume normal
heart impulse of a patient by discharging energy impulse to a
patient who is suffering from an emergency cardiac attack. It
is commonly available in hospital, ambulances, and medical
institutes, and it is generally assumed that the physiological
effect of a defibrillator impulse is governed by the impulse
energy in Joule (J). This is one of the most important
parameters of defibrillator that need to be calibrated.
Defibrillator analyzer is designed to measure the energy of the
discharge impulse delivered by a defibrillator, commercial
products such as FLUKE Impulse 6000D/7000DP,
DATREND Phase 3 etc., had been developed by companies
to calibrate the defibrillators. Calibration according to the
metrology traceability system of the medical devices are
essential to ensure these instruments serving the public
accurately and reliably.[1, 2] The standard instruments used
for calibration of medical instruments must have appropriate
and acceptable accuracy and it must be traced back to the SI
Units.
Given the particularity of defibrillator analyzers’
measuring parameters (e.g. impulse energy), the measurement
criteria that various metrological services follow are unable to
guarantee the traceability of calibration. Calibration of the
analyzer could only be realized by following the delivery
calibration method of manufacturers, namely, by opening
enclosure in accordance with schematic circuit diagram of
instrument in the past. Since it is required to open enclosure,
this method still fails to meet the demand for whole system
calibration. Additionally, the inconsistency in different types Fig. 1 Traceability chart of defibrillator to SI Units
of schematic circuit diagram from various manufacturers
II. DEFIBRILLATION IMPULSE WAVEFORM
makes it difficult to develop a uniform standard and criteria.
In terms of defibrillation impulse energy measurement, the A. Defibrillator principle
National Physical Laboratory in UK compared the A typical defibrillator includes high voltage supply
measurement results obtained by micro-calorimeter method (including power supply, transformer and rectifier), charge
and data collection method, and the consistency reaches 1%.[3] circuit, capacitor, inductor, discharge circuit, patient paddles,
The Standards and Calibration Laboratory in Hong Kong had etc. The power supply can be from a battery or the main supply.
also done research work on the calibration of defibrillator A simple circuit is shown as Fig. 2.

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

978-1-7281-5386-5/20/$31.00 ©2020 IEEE
Authorized licensed use limited to: CMU Libraries - library.cmich.edu. Downloaded on August 17,2020 at 20:22:28 UTC from IEEE Xplore. Restrictions apply.
 between 200 V and 3000 V. The peak current is about 4 A to
60 A. The duration time is around 10 milliseconds to 20
milliseconds. The formula of energy calculation is:
𝐸 = ∫ 𝑃(𝑡)𝑑𝑡 = ∫ 𝑈(𝑡) 𝐼(𝑡)𝑑𝑡 (2)
Where, E is impulse energy, P(t) is impulse power, U(t) is
impulse voltage, I(t) is impulse current, and t is impulse
duration.
Fig. 2 A simple circuit of a defibrillator The developed calibration standard for impulse energy of
defibrillator analyzer was mainly composed of voltage divider,
In which, the high voltage supply is used to convert the current sensor, sampling digitizer, control software and
power line voltage to high voltage DC to charge the capacitor. calculation algorithm.
The capacitor is used to store a large amount of electrical
charge, then release it over a short period of time in a patient A. Voltage divider
heart. The inductor minimizes the rapid decay of current flow Voltage divider is used to convert high voltage into low
(delivered energy), it prolongs the duration of current flow to voltage that oscilloscope or digitizer can measure. For impulse
allow for optimum time. The delivered impulse energy must signal, resistance voltage divider is often used as it has the
be maintained in several milliseconds for a successful advantage of simple structure, high measurement accuracy
defibrillation. Effective defibrillation depends on released and long-term stability. As the bandwidth of the resistive
energy at the heart. The current and charge delivered by divider is related to the stray capacitances to earth, it is
discharging capacitor decay rapidly in exponential function. necessary to optimum the capacitances in order to get better
B. Type of defibrillation waveform dynamic response performance. In order to minimize the size
of voltage divider with wide bandwidth, techniques such as
In general, defibrillator performs electric defibrillation inductance compensation have also been used to improve the
according to the principle of RLC damped oscillation response performance of voltage divider. The schematic
discharge. Defibrillators produce monophasic shocks where diagram of voltage divider is shown in Fig. 3.
the current flows in one direction and it is generally
recommended that the maximum of energy deliver is 360 J.
Modern defibrillators could also generate biphasic impulse to
a patient with an energy level in the range of about 50 J to 300
J. With biphasic shocks, the current flows in one direction for
specified time before reversing to another direction for the
remainder of the electrical discharge time. By analyzing the
power spectrum of the defibrillator waveform, we found that
more than 99% of the impulse energy was within 5000 Hz
frequency range.
C. Defibrillator analyzer
A defibrillator analyzer is designed to measure the
defibrillation impulse energy generated by a defibrillator. The Fig. 3 The schematic diagram of voltage divider
defibrillator discharge is applied through the application discs
of the analyzer with which the user should couple the paddles 𝑘 = (3)
of the tested defibrillator. The impedance between discs
simulates the impedance of a patient. The discharged Where, k1 is voltage ratio of voltage divider, U is input
defibrillation impulse is attenuated and filtered before voltage of voltage divider, U1 is output voltage of voltage
measuring by the A/D converter. The A/D converter is linked divider
to the MCU for calculating the energy. The digital signals are The input impedance of voltage divider must be large
then processed, and the energy and other defibrillator impulse enough to minimize the loading effect to the impedance of
parameters are measured. UUC. By applying input voltage across terminal U+ and U-,
and then measure the output voltage across terminal U1, ratio
III. MEASUREMENT STANDARD FOR IMPULSE ENERGY
of the divider k1 could be measured. Dynamic response could
CALIBRATION IN NIM be check by frequency response and rise time in step response.
A set of measurement standard for impulse energy of
defibrillation is developed for solving the problems of Table 1 Frequency response of the voltage divider
calibration defibrillator and defibrillator analyzer. The
formula for calculating the defibrillation energy error is: Frequency Ratio Error Phase Error
(Hz) (%) (degree)
𝛿 = × 100% (1)
50 0.01 0.01
Where, 𝛿  is the relative error, E is the impulse energy 100 0.01 0.01
measured by the standard, E0 is the energy measured by the 200 0.01 0.01
unit under test (UUT). 400 0.01 0.05
Energy output signal of defibrillator is a short-time large- 500 0.02 0.07
voltage impulse signal. The peak voltage is approximately 1000 0.03 0.02

Authorized licensed use limited to: CMU Libraries - library.cmich.edu. Downloaded on August 17,2020 at 20:22:28 UTC from IEEE Xplore. Restrictions apply.
 2000 0.04 0.04 IV. DEFIBRILATOR CALIBRATION SETUP
5000 0.03 0.11 A. Measuremnt Setup
10000 0.07 0.18

B. Current sensor
The current sensor was based on zero flux detection
technology and has a small-signal bandwidth up to 500 kHz
as well as minimal phase angle and amplitude error. As the
output has no electrical connection with the input, it could
prevent the interaction between voltage channel and current
channel. The ratio is measured by applying input current and
then measure output voltage. The resistor shunt is connected
to converse current to voltage at output terminal. The
schematic diagram of current divider is shown in Fig. 4. Fig. 5 Wiring Diagram of Defibrillation Analyzer Calibration
With the defibrillation impulse energy testing system as
the measurement standard and defibrillator as energy
generator, we could calibrator the unit under test, which is a
defibrillator analyzer.
B. Calibration Procedure
The calibration procedure include following steps:
a) Measure impedance of defibrillator analyzer (UUC)
by Ohmmeter, normally it is 50 Ω, and this is used in cross
check equation.
Fig. 4 The schematic diagram of current sensor
b) Set Voltage ratio, Current ratio (from certificated or
Where, k2 is the current ratio of current transducer, I is the checked result) and impedance of the defibrillator analyzer to
input current of current divider, U2 is the output voltage of software of calibration system.
current divider.
c) Connect circuit for measuring (see Measurement set
C. Digitizer and Algorithm up as Fig 5.)
Multi-resolution acquisition card PXI-5922 (NI, USA)
d) Set the function of the UUC to DEFIB function
was adopted for synchronously record the two voltage signals.
The sampling rate was set at 1MHz. The system’s resolution e) Set energy range of the UUC (at low, medium and
was 22bit and the error was digitizer was calculated to 0.01% high range cover range from min to max energy)
accuracy.
f) Set energy of generator and apply energy (by
The software adopted two energy calculation algorithms, charging and discharging) to the UUC and Defibrillation
one is symbolled as VR-calculation and the other as VI- Impulse Energy Testing System for measuring
calculation. And we had also done numerical simulations to
compare the performances of discrete integration algorithms, g) Record reading of the UUC and Defibrillation
including rectangular, Simpson and Newton-Cotes methods. Impulse Energy Testing System
Rectangular calculation result is 0.02% smaller than the other h) Repeat step (e) to (g) for different energy levels (such
two methods. So, Simpson or Newton-Cotes algorithm is as 2 J, 5 J, 7 J, 10 J, 20 J, 30 J, 50 J, 70 J, 100 J, 150 J, 200 J,
chosen for integrate calculation. 300 J, 360 J).
If the load resistor is considered as pure resistance, then V. UNCERTAINTY EVALUATION
the impulse energy is calculated with the voltage on the load
and the impedance of the load. The calculation formula is; A. Uncertainty of calibration of impulse energy
measurement standard
∑ (4) 1) Mathematic Model
On the other hand, the energy could also be calculated with According to the calculation model of impulse energy, the
the voltage on the load and the current through the load, then impulse voltage and current are attenuated and then
the formula is; synchronously sampled by the digitizer, and then integrated to
measure the impulse energy. Formula (2) is discretized to
∑ 𝑈 ×𝐼 (5)
𝐸=∑ 𝑘 𝑉 ⋅ ⋅ 𝛥𝑡 (6)
As the resistor is not pure resistance, the VI-calculation
algorithm is more reliable, and VR-calculation serve as a In which, V1j is the output voltage of the voltage divider,
cross-checking value. V2j is the voltage on the shunt connected to the secondary of
the current transformer, 𝛥𝑡 is the sampling interval, 𝑁 = is
the sampling points.
Assuming 𝑝 = ∑ (𝑉 𝑉 ), then we get

Authorized licensed use limited to: CMU Libraries - library.cmich.edu. Downloaded on August 17,2020 at 20:22:28 UTC from IEEE Xplore. Restrictions apply.
 𝐸= ⋅ (7) f) Expanded uncertainty is
𝑈(𝐸) = 𝑘 ∙ 𝑢 (𝐸) = 0.05%, 𝑘 = 2 (19)
It is not easy to accurately analyze the cross-correlation
between each component in Equation (7), but it is relatively 4) Uncertainty Budget of the measurement standard
easy to reduce the influence on the overall uncertainty. As the Table 2 Uncertainty budget of impulse energy measurement
input and output of current sensor is isolated, the influence standard calibration
between voltage channel and current is less significant. The Source Coefficient Value
output of voltage divider and current sensor are both 𝑢 (𝑘 ): Voltage Divider Ratio 1 0.013%
transferred by coaxial cable, as this could reduce the 𝑢 (𝑘 ): Current Sensor Ratio 1 0.013%
environmental noise. And we measured the ratio of the 𝑢 (𝑟): Sampling Resistor 1 0.006%
voltage divider, the ratio of current sensor, the impedance of 𝑢 (𝑝).: Production of
the shunt and digitizer and the accuracy of the digitizer with measured voltages by the 1 0.018%
different multifunction calibrators. Although this method is digitizer
complicated, it could lessen the influence of cross-correction.
𝑢 (𝐸): combined standard uncertainty 0.026%
Then, we get the mathematic model for calculating the
𝑈(𝐸): Expanded uncertainty, k=2 0.05%
uncertainty of the energy E
𝑢 (𝐸) = 𝑢 (𝑘 ) + 𝑢 (𝑘 ) + 𝑢 (𝑟) + 𝑢 (𝑝) (8)
Expanded uncertainty is B. Uncertainty of calibrating defibrilator analyzer with the
𝑈(𝐸) = 𝑘 ∙ 𝑢 (𝐸) (9) impulse energy measurement standard
2) Uncertainty components 1) Mathematic Model
a) Ratio of the voltage divider 𝑢 (𝑘 ) ，combined 𝐸 = 𝐸 +𝛿 (20)
by the uncertainty caused by DC ratio 𝑢 and AC ratio 𝑢
In which, 𝐸 is the energy measured by the defibrillator
𝑢 (𝑘 ) = 𝑢 + 𝑢 (10) analyzer, 𝐸 is the energy measured by the reference
b) Ratio of the current sensor 𝑢 (𝑘 ), combined by measuring system, and 𝛿 is the difference.
the uncertainty caused by DC ratio 𝑢 and AC ratio 𝑢
2) Uncertainty Components
𝑢 (𝑘 ) = 𝑢 + 𝑢 (11)
a) Resolution of the defibrillator analyzer 𝑢
c) Resistor of the resistor 𝑢 (𝑟)
d) Uncertainty of the product of two voltage measured Introduced by the resolution of the analyzer under test,
by the digitizer 𝑢 (𝑝). It could be proved by mathematical and this is a constant value, which did not change with the
analysis that, defibrillation energy. Commonly, it is 0.1 J, so this
𝑢 (𝑝) < √2 ∙ 𝑢 (𝑉) (12) uncertainty component is
In which, 𝑢 (𝑉) is the uncertainty of voltage . J
measurement by the digitizer, which is combined by DC 𝑢 = = 0.03 J (21)
√
accuracy 𝑢 and AC flatness 𝑢 , so
b) Uncertainty of the reference impulse energy
𝑢 (𝑝) < √2 ∙ 𝑢 + 𝑢 (13) measuring system 𝑢
3) Evaluation of the uncertainty component
a) According to the voltage divider calibration According to the result of calibration certificate, the
certificate, the maximum DC permissible error is 0.01% and expanded uncertainty is 0.05%𝐸, k=2, So this component
AC error is 0.02%, so the combination is 0.022%. According value is
to normal distribution, 𝑢 =
. %
= 0.025%𝐸 (22)
. %
𝑢 (𝑘 ) = = 0.013% (14)
√
b) According to the current sensor calibration c) Uncertainty caused by repeatability 𝑢
certificate, the maximum DC permissible error is 0.01% and We did ten experiments with the test point of 360 J, to
AC error is 0.02%, so the combination is 0.022%. According measure the repeatability and the uncertainty component.
to normal distribution, Result of the experiments is as follows,
. %
𝑢 (𝑘 ) = = 0.013% (15)
√
c) According to the resistor calibration certificate, the maximum Ratio of the
Standard
measurement permissible error is 0.01%. According to deviation to Uncertainty
deviation of Freedom
normal distribution, reference component
the difference
. % value
𝑢 (𝑟) = = 0.006% (16)
√
d) According to the digitizer calibration certificate, DC 0.1 J 0.03 % 9 0.01 %
voltage accuracy of the digitizer is 0.02% and AC flatness is
0.01%, so the combination is 0.022%。According to normal 3) Combined of Standard Uncertainty
distribution,
√ ∙ . %
𝑢(𝐸 ) = 𝑢 + 𝑢 + 𝑢 = 0.03𝐽 + 0.027%𝐸 (23)
𝑢 (𝑝) < = 0.018% (17)
√ Expanded uncertainty
e) According to Equation (8), combined standard
uncertainty is 𝑈(𝐸) = 𝑘 ∙ 𝑢(𝐸 ) = 0.06𝐽 + 0.06%𝐸 (24)
𝑢 (𝐸) = 0.026% (18)

Authorized licensed use limited to: CMU Libraries - library.cmich.edu. Downloaded on August 17,2020 at 20:22:28 UTC from IEEE Xplore. Restrictions apply.
 4) Uncertainty Budget of the Calibration of the [2] JJF, “Calibration Specification for Cardiac Defibrillators & Cardiac
Defibrillator monitors,” JJF 1149, 2006.
Defibrillator Analyzer
[3] Stockton JR and Smith RC, “Traceable calibration of a cardiac
Table 3 Uncertainty budget defibrillator analyzer calibration defibrillator energy meter,” Med. Biol. Eng. Comput., 1981, 19(1): 40-
48.
Source Coefficient Value
[4] Yang S and Yan A. “Calibration of Defibrillator Analyzers”, NCSLI
𝑢 : Resolution of Measure, 2015, 10(4): 50-55.
the Defibrillator 1 0.03J [5] Zhao W, SHAO HM, etc. “Energy calibration technique of defibrillator
Analyzer analyzer”, 2015 12th IEEE International Conference on Electronic
𝑢 : Uncertainty of Measurement & Instruments (ICEMI), Qingdao, 2015, pp: 392-396.
the Reference [6] SHAO HM, Zhang HH, Zhao W, etc, "Development of Calibration
1 0.05%𝐸 System for Defibrillator Analyzers," 2018 Conference on Precision
Impulse Energy Electromagnetic Measurements (CPEM 2018), Paris, 2018, pp: 1-2.
Measuring System [7] International Organization for Standardization, ISO/IEC Guide 98-
𝑢 : Repeatability 1 0.01%𝐸 3:2008 – Uncertainty of Measurement – Part 3: Guide to the Expression
𝑢 (𝐸): Standard Uncertainty 0.03𝐽 + 0.027%𝐸 of Uncertainty in Measurement (GUM), ISO, 2008.
𝑈(𝐸): Expanded Uncertainty, [8] International Organization for Standardization, ISO/IEC Guide
0.06𝐽 + 0.06%𝐸 99:2007 – International Vocabulary of Metrology – Basic and General
k=2 Concepts and Associated Terms (VIM), ISO, 2007
[9] M. Grabe, “Estimation of measurement uncertainties – an alternative
to the ISO guide”, Metrologia, vol. 38, pp.97-106, 2001,
VI. CONCLUSION [10] M. Cox, P. Harris, B.L.-R. Siebert, “Evaluation of measurement
In this manuscript, we introduced the measurement uncertainty based on the propagation of distributions using Monte
Carlo simulation”, Measurement Techniques, vol. 46, pp 824-833,
standard and the traceability system we have built for 2003.
calibrating the impulse energy of defibrillator and defibrillator [11] M. Cox, P. Harris, “An outline of Supplement 1 to the guide to the
analyzer. The uncertainty for the defibrillator analyzer expression of uncertainty in measurement on numerical methods for
calibration could be 0.06J plus 0.06% reading, which is the propagation of distributions”, Measurement Techniques vol. 48, pp.
enough for current defibrillator analyzer calibration. 336-345, 2005.
[12] F. Adamo, F. Attivissimo, A. Di Nisio, M.Savino, M. Spadavecchia,
We had also built a transfer standard with the accuracy of “A spectral estimation method for nonstationary signals analysis with
0.2% reading plus 0.1 J that could be supplied to other application to power systems”, Measurement, vol. 73, pp. 247-261,
calibration institutes for defibrillator analyzer calibration.[6] 2015.
We would also be glad provide the impulse energy calibration [13] C. Elster, B. Toman, “Bayesian uncertainty analysis under prior
service to global institute with ISO/IEC 17025 certificate that ignorance of the measurand versus analysis using the Supplement 1 to
the guide: a comparison”, Metrologia, vol. 46, pp.261-266, 2009.
could be traced to NIM China’s measurement standard.
[14] J. Berger, “The case for objective Bayesian analysis”, Bayesian
Analysis, pp.385-402, 2006.
[15] F. Attivissimo, A. Cataldo, L. Fabbiano, N. Giaquinto, “Systematic
ACKNOWLEDGMENT errors and measurement uncertainty: An experimental approach”,
Measurement, vol. 44, pp. 1781-1789, 2011.
The authors wish to thank the National Key R&D Program
[16] P.M. Ramos, F.M. Janeiro, P.S. Girao, “Uncertainty evaluation of
of China(2017YFF0205701) and National Science & multivariate quantities: A case study on electrical impedance”,
Technology Pillar Program (2011BA102B04) for funding Measurement, vol. 78, pp. 397-411, 2016.
support. We would also like to acknowledge Mr. Sittisak [17] C. Elster, “Bayesian uncertainty analysis compared with the
Pimsut for his contribution during his visit in NIM. application of the GUM and its supplements”, Metrologia, vol. 51,
2014.
REFERENCES
[1] IEC, “Medical electrical equipment - Part 2-4: Particular requirements
for the basic safety and essential performance of cardiac defibrillators,”
IEC 60601-2-4, 2010.

Authorized licensed use limited to: CMU Libraries - library.cmich.edu. Downloaded on August 17,2020 at 20:22:28 UTC from IEEE Xplore. Restrictions apply.