Establishment of calibration equipment for defibrillator analyzer and evaluation of measurement uncertainty

A cardiac defibrillator is a medical electrical device used to defibrillate the heart of patients. In JJF1149-2014, defibrillator analyzers are used to measure the output energy of cardiac defibrillators, whose energy accuracy is an important parameter. This manuscript describes establishment of calibration equipment for defibrillator analyzer, including voltage divider module, sampling module and data processing module. Measurement uncertainty is evaluated for measurement results at last.


Introduction
A cardiac defibrillator is a medical electrical device used to defibrillate the heart by applying electrical pulses to the patient's skin (external electrodes) or exposed heart (internal electrodes) through electrodes. The width of the instantaneous high-energy pulse is (4-10) ms, and the output energy is (40-400) J. The output energy of the cardiac defibrillator is very important. It may not reach the effect if too small, or the patient may be injured if too large.
In JJF1149-2014 Calibration Specification for Cardiac Defibrillators [1], defibrillator analyzers are used to measure the output energy of cardiac defibrillators. Therefore, the energy accuracy of the defibrillator analyzer is an important parameter.
a) comparison method This method requires a defibrillator and a standard defibrillator analyzer as a standard device. Firstly, use the standard defibrillator analyzer to calibrate the energy value of the defibrillator, denoted as E0. Then use the calibrated defibrillator analyzer to measure the energy value of the defibrillator, denoted as E1. Set ∆ the error of energy value, then This method is simple to operate. However, the output energy of the cardiac defibrillator is unstable with poor repeatability. When energy value is 300J, the maximum deviation of the output energy may reach ±10J. When this method is used for calibration, the measurement uncertainty is relatively large, about 4.6%(k=2). b) power integration method According to the energy formula where E is the energy, P is the power, V is voltage, I is electricity, and R is resistance.
This method needs to design an energy test system with high requirements for the standard device. This method solves the problems of poor repeatability of the cardiac defibrillator and traceability of energy value.
We choose this method for measurement.

Establishment of calibration equipment
If the sampling time T of the device is short enough (less than 0.1μs), then The schematic diagram of the equipment is shown in Fig. 1. The voltage divider module uses a voltage divider, and divider ratio r is a fixed value, such as 1000:1. The sampling module uses an oscilloscope, whose voltage is donated as ( ). The defibrillator analyzer has a fixed resistance, usually 50Ω, which can be calibrated by a digital multimeter. Then

Analysis of pulse waveform of cardiac defibrillator
We need to analyze the pulse waveform generated by the defibrillator. The energy of the cardiac defibrillator is 360J, measured by an oscilloscope with a voltage probe shown in Fig. 2.a. The waveform has a width of 5ms and a peak value of 3.16kV.
The Fourier transform of the non-periodic continuous-time signal x(t) can be expressed as The discrete Fourier transform (DFT) of a finite-length discrete signal x(n) is defined as Fig. 2.b shows the curve of the pulse signal in the frequency domain by MATLAB. The energy of the signal is mainly concentrated in the low frequency part, that is, the part from DC to 1kHz. Therefore, the defibrillation analyzer calibration system needs to meet requirements of the AC voltage peak value greater than 3.16kV, DC to 1kHz with high accuracy.

Voltage divider module
After investigating voltage dividers on the market, we found many manufacturers with different performance. Finally, we customized the VD5 voltage divider by ROSS company. The specifications of the instrument are as follows It can be seen from table that the error of the voltage divider ratio at DC is much less than ±0.1%, and the error of voltage divider ratio at 60Hz is less than ±0.2%.
According to specification, the error is about ±3% at 1MHz, and the 3dB bandwidth point is about 7.1MHz. The frequency response curve is relatively flat below 1kHz. The instrument meets the requirements of the equipment.

Sampling module
We need a high-precision real-time sampling oscilloscope, which can be used to process the results in real time. We chose the 14-bit PXI-5122 oscilloscope produced by NI, which has a bandwidth of 100MHz and a sampling rate of 100MS/s. Our institute is authorized for calibration of NI instruments. We can use the calibration program Calibration Executive to adjust the accuracy of the PXI-5122. After adjustment, its AC accuracy can reach 0.37%, which greatly improves the accuracy of the oscilloscope.
To use PXI-5122, we need a NI PXI-1050 chassis, a PC with PCI slots, and PCI-8336 and PXI-8336 for communication.

Data processing module
When using a traditional oscilloscope for measurement, we must save data into a csv file, then copy the file to a computer for data processing. Thus, the results cannot be displayed in real time.
We write LabVIEW programs for NI oscilloscope, and the program can display result in real time.
The front panel of the energy test program is as follows. Resistance is used to set the resistance value, that is, the internal resistance value of the defibrillation analyzer measured by the 3458A multimeter. The Energy indicator displays the calculated energy value.

Measurement results
Set the defibrillator in six energy levels. The test results are as follows.

Evaluation of standard uncertainty
Evaluations of standard uncertainty are as follows.
1)The measurement uncertainty component 1 introduced by ac accuracy of the high voltage divider According to the specification, the voltage divider VD5 is generally better than ±0.1% DC, ±0.2% 50/60 Hz. Take the ac accuracy as the maximum permissible error for calculation. Assuming it is uniformly distributed, we take k1=√3, thus 1rel = 0.2%/√3=0.115%.
4) The measurement uncertainty component 4 introduced by ac amplitude accuracy of the oscilloscope According to the specification of PXI-5122, maximum permissible error of DC is ±0.65%.

5) The measurement uncertainty component 5 introduced by the resolution of the calibrated defibrillator analyzer
The resolution of the defibrillator analyzer is 0.1J, then a5=0.05J. Assuming it is uniformly distributed, we take k5=√3, then 5 =a5/k5=0.03J.
6) The measurement uncertainty component 6 introduced by the measurement repeatability of the calibrated defibrillator analyzer Set the defibrillator in six energy levels, 30J, 50J, 100J, 200J, 300J, 360J, and test each energy level for 10 times. The test standard deviation of a single measurement result is Standard uncertainty is expressed by test standard deviation, so 6 =s.

Uncertainty synthesis
The uncertainty components 1rel , 2rel , 3rel , 4rel are independent of each other. According to the formula Which is = �4 × 4rel 2 + 4 × 1rel 2 + 2rel 2 + 3rel 2 = 0.787%. The uncertainty components 5 and 6 are positively correlated, and the larger one among them is chosen to be participated in the synthesis.
In summary, there are two main sources of uncertainty. One is rel , and one is 6rel . These two components are independent of each other. rel = � rel 2 + 6rel 2 (10) Calculate the relative combined standard uncertainty at different energy level respectively.

Results
We firstly analyzed the defibrillation signal generated by a defibrillator, then designed the measurement equipment, including voltage divider module, sampling module, and data processing module. With LabView program, the equipment can measure the energy value in real time. At last, the uncertainty evaluation results are given.