1 测量阻抗的方法
主要有电桥法、谐振法和矢量阻抗法。
电桥法以电桥的平衡原理为基础，是最通用和最准确的测量方法。因电桥平衡指示器的灵敏度和标准兀件的正确度都比较高，故用这种方法测量兀件参数能得到很高的精度。电桥的研究己经有一百多年的历史，其类型也非常多。但是电桥法测量需要反复进行平衡调节，测量时间长，因而很难实现快速的自动测量。
谐振法以LC回路的谐振特性为基础，通过测定谐振频率和己知的电感(或电容)计算出被测兀件的参数值，谐振法主要用于对高频(300kHz～300MHz)回路参数的测量。这种测量方法不仅简单，而且被测兀件又大都应用在谐振电路中，用谐振法测量更能使测量条件和具体的应用条件相一致。然而从另一方面讲，当信号频率较高时，低失真度的正弦信号很难获得，这就限制了测量精度的提高和测量范围的扩大。谐振法还要不断的调节回路兀件参数使回路达到谐振，故而调节麻烦测量速度慢，不能适应批量生产。
矢量阻抗法以阻抗的定义为基础，利用流过量程电阻和被测阻抗的电流相等的原理和运放的比例特性－－被测阻抗和量程电阻上的电压矢量之比就等于他们的阻抗之比。通过测量这两个阻抗上的电压矢量，利用微机计算，即.丁求得被测阻抗参数。显然要实现这种方法，仪器必须能够进行矢量测量及除法运算，这就使微处理器的应用成为必然。矢量阻抗法有固定轴法和自山轴法两种实现方案，相比之下自山轴法更具优越性。
以上提到的这些测量方法都要求激励信号是低失真度的正弦波信号。然而频率较高的低失真度的正弦信号很难获得，这就限制了测量精度的提高和测量范围的扩大。 （基于DSP的高精度阻抗参数测量新方法及其实现）
The impedance measurement is one of the most important parts of the measurements of electrical quantities. Different impedance measurement techniques, such as oscillators, frequency domain techniques, and digital AC bridges, etc., have been developed in the last decades to satisfy the increasing requirements. For precise measurements, however, these measuring methods need precise measuring hardware[1].
An intuitive idea of how to measure the amplitude and the phase of a signal is to use a peak detector for the amplitude and a phase detector, for instance, based on zero crossings [6], for the phase. However, this is not a good approach in the case of bioimpedance since the injected current is very low and the environment is quite noisy (the sample can generate voltages by itself). Thus, it is advisable to use some kind of demodulation to reject the noise
西安交通大学博士学位论文
or the interferences that are not in the frequency range of interest. Generally, the bioimpedance changes are very slow and this implies that the frequency range of interest has a spectral width of some Hz and is centered on the frequency of the reference signal. Because of such signal features, coherent demodulation, also called synchronous demodulation or detection, is employed in most cases[2]
（Instrumentation for electrical bioimpedance measurements）
1.1 桥路法与解调法
There are two basic techniques for measuring impedance of biologic tissue: bridge and phase sensitive detector methods. Traditionally, bridge methods [48] have been the most commonly used. Their major advantage is high measurement resolution and accuracy. However, measurements are quite time consuming, which is a problem during in-vivo work where tissue impedance changes in time due to ongoing physiological processes. Recent improvements in the design of bridge systems have decreased the measurement time, however, their accuracy was reduced too.
Nowadays, phase sensitive detector methods are gaining popularity. Our IS design is based on phase sensitive detector technique, for the following reasons:
• its ability to perform accurate measurements throughout the frequency range; • its ability to perform fast automated measurements; and, • simplicity of design and operation.[3]
(Chapter 5)
The bridge method traditionally has been the most popular method for ischemia measurements. There are a number of disadvantages to this system. This method is time consuming and in commercial models often lacks su_cient accuracy for phase measurements in biological tissue. Measurement accuracy is also poor at low frequencies[3].
（Tissue Ischemia Monitoring Using Impedance Spectroscopy - Clinical Evaluation）
Electrical impedance measurements have been widely to study biological systems. These studies generally aim to correlate the electrical structure or physiological events. Both bridge and phase-sensitive detectors have been parameters with tissue bridge techniques and applied in with either two- or four-electrode systems. The combination major advantages of the bridge techniques are the high resolution and accuracy that can be obtained. However, these measurements are time-consuming and not well suited in situations where impedances change rapidly. Therefore, in case of physiological measurements, phase-sensitive detector methods (lock-in amplifiers) are generally used.
（The Four-Electrode Resistivity Technique in Anisotropic Media: Theoretical Analysis and Application on Myocardial Tissue in Vivo）
128
声明
Measurement of the relation between the modules and phase angle of bio-impedance and frequency, called bioimpedance spectroscopy, is very interesting and important for fundamental research particularly in physiology and pathology, as well as for clinical· practice. The applications of this method comprise the assessment of extracellular and intracellular water in the body[3]. Another example is examination of body composition (water, fat, bone, muscle) in various regions [1]. There are several methods for measurement of modules and phase angle or components of impedance [2,4]. Three techniques are usually used: the bridge method; the method based on measurement time shift between signals of current and voltage; the coherent detection method.
(Multifrequency device for measurement of the complex electrical bio-impedance-design and application)
1.2 正交相干解调法（two-phase reference coherent demodulation
method）
The impedance measurement circuit: From the IA a sinusoidal waveform with amplitude and phase proportional to the impedance of the tissue is obtained. For decoding this information, two phase reference coherent demodulation method is employed [6-8]. This is illustrated in Fig. 7, where two in-quadrature clocks signals must be used to perform the multiplier operator. The excitation circuit has a bandpass filter based on a TTB that implements simultaneously BP and LP functions, both in 90˚ phase shift [6]. These signals are used to generate clocks for demodulation purpose: the in-phase and in-quadrature clocks. This is done by means of two differential comparaters placed in each filter output (LP and BP). For a given phase shift α between I and Vx, when the multiplication has been performed, the DC levels obtained at the demodulator outputs are proportional to the cos(α) and sin(α) functions, being these the real and imaginary parts of Zx. A four order lowpass filter, with very low cutoff frequency (about 200Hz) is necessary to select these DC signals. This filter
西安交通大学博士学位论文
has been implemented by cascading two TTB circuits, with input pair transistors working in weak inversion. In order to reduce the capacitor area, bias currents of 2nA have been used to obtain low transconductance to capacitance ratios. [4]
We use the two-phase reference coherent demodulation method [8], which gives simultaneously the real (Re(Zx)) and imaginary (Im(Zx)) component of a complex impedance Zx .
1.3 模拟解调法
相敏解调器(phase sensitive demodulator,PSD)采用与输出激励信号同步的参考信号解调出被测信号的幅值和相位信息(或实部与虚部信息), 也称为相敏检测器(phase sensitive detector)或同步解调器(synchronous demodulator)。实现相敏解调可以采用模拟或数字解调技术。模拟解调有开关解调器、模拟乘法器及我们提出的脉冲采样解调器等, 相应的解调参考信号为方波、正弦波和窄脉冲。相敏解调的最大优点是可以有效抑制与激励频率不同频的干扰及噪声, 但它需要与激励信号同频、同相的解调参考信号[5]。
在生物电阻抗检测、阻抗血流图和电阻抗断层成像(EIT)中，在安全的交流电压或电流激励下，被测生物组织的电阻抗信息调制在激励信号上，通过对调制信号的幅值或相位进行测量，可以得到被测目标电阻抗的实部或虚部。要测量调制信号的幅值和相位信息，可以采用模拟解调或数字解调技术。模拟解调器的基本原理是将调制信号与参考信号相乘, 以得到与调制信号的幅度或相位成正比的解调输出信号。
在多频生物电阻抗模拟解调技术中，广泛采用双路模拟乘法器和低通滤波器实现正交解调，获得与生物电阻抗实部及虚部相关的模拟信号，在参考信号幅值精确已知的条件下，可以求得被测生物电阻抗的实部和虚部或幅值和相位D\"E。在频带较宽时，要精确测得不同频率下参考信号的实际值较困难，常采用固定值或通过定标间接求取，从而引入了测量误差[5, 6]。
在正弦交流激励进行生物电阻抗测量的系统中，常用的模拟解调技术主要基于四象限乘法器实现实部、虚部解调或幅值、相位解调。实现原理是将与激励信号同频同相的正弦参考信号及同频正交的余弦参考信号与被测信号进行四象限相乘，被测信号、参考信号分
130
声明
由上式看出，解调输出信号不仅与被测信号的幅值和相位有关，还与参考信号的幅值和相位有关。在求取信号幅值和相位时，除了要用到解调输出信号外，还需要精确知道参考信号的幅值。在单频系统中，参考信号的幅值可以预先测到，而多频系统中要测量所有频率点的幅值是相当困难的，一般采用对有限个频率点进行测量或通过定标的方法间接获得实际条件下的幅值，因而可以认为是已知量。参考信号幅值已知的前提下，对上述（1）、（2）式进行三角函数变换可得：
由此说明，目前的正交解调方法中， 参考信号幅值的测量精度直接影响被测信号幅值和相位的测量精度[6]。
相敏解调法phase-sensitive detection method
西安交通大学博士学位论文
In order to meet the requested characteristics, biological impedance measurement systems should be based on the 4 electrode phase-sensitive detection method [2]. The basic idea of this approach is presented in Fig. 1. If a voltage on the electrodes is v(t)=Vcos(ω0t+θ)= Vcosθcosω0t
－Vsinθsinω0t (impedance time
variation, and noise are neglected), then the resistive component of the tissue impedance is R=Vcose/I, and reactive component of the impedance is X=Vsine/I. To measure these components, a phase-sensitive demodulation method is applied on the measured voltage on the electrodes (Fig. 1). Demodulated in-phase and quadrature-phase signals are proportional to the resistive and reactive components of the impedance, respectively. These signals are sampled with the A/D converters for further processing. [7]
Ackmann and Seitz [1] have reviewed the methods for complex bioelectric impedance measurement. The separation of in-phase and quadrature impedance components is usually done by phase-sensitive techniques relaying on analog demodulators. Signals are processed by two analog channels that must be closely matched; otherwise, large phase errors result (Fig. 1). Each channel includes an analog demodulator, followed by a low-pass filter, a sampler, and, usually, an ADC. Alternatively, a single sampler and ADC can be shared by both channels if an analog multiplexer precedes them. [8]
132
声明
A prototype system was designed and constructed. This system includes a four-electrode catheter described in detail by Sacrist´an (2001), and a complex impedance spectrometer by parallel demodulation described by Othman (1999). The system uses a PC as interface and platform for spectral processing. The spectrometer generates an excitation current at different frequencies within the range of interest, from 0.05 to 300 kHz. This current is injected into the tissue through the external electrodes of the catheter (figure 1). The internal electrodes of the catheter measure the potential generated in the tissue by the excitation current. A reference potential is also generated across a known resistance in series with the tissue. The instrument uses the demodulation method to transfer the information of any excitation frequency to the same low frequency for analogue–digital conversion, without losing the amplitude and phase information (figure 2). The tissue signal and the reference signal are processed and digitized by identical parallel circuits. The amplitude and phase are calculated in software for each 280 C AGonz´alez et al excitation frequency from the tissue signal relative to the reference signal. The complex impedance spectrum is then calculated from the
西安交通大学博士学位论文
impedance measurements at pre-programmed frequencies sweeping the range of interest.[9]
1.3.1 缺点：
传统的复包络采样方法[20]是采用模拟正交双通道采样，即对中频信号乘以，井滤除分量后得到正交的双路基带信号，再进行数字采样。如图2-1所示，此法需要产生完全正交的模拟木振信号和，这在实际中很难做到，同时模拟双通道存在漂移和通道不一致性，得到的双路正交信号存在较大的误差[21]，反映在频域是附加了镜频分量[22]。此方法的正交性能一般为:幅度平衡在0.5 dB左右，相位正交误差在3度左右，即幅相误差引入的镜像功率在－3 0 dB左右[[23]。这就限制了信号处理器性能的提高。 （基于DsP的高精度阻抗参数测量新方法及其实现）
21 F. Churchill, G. Ogar and B. Thompson. The Correction of I and Q Errors in a Coherent Processor. IEEE Trans. on AES,1981，17 (1):131一137
22 S. J. Roome. Analysis of Quadrature Detector Using Complex Envelope Notation. IEEE Proc.,1989,136(2):95一100
23 A. L. Sinsky, P. C. Wang. Error Analysis of a Quadrature Coherent Detector Processor. IEEE Trans. on AES,1974,145(10):880一883
传统的模拟混频正交解调方法很难保证正交两路具有精确的幅度一致性和相位正交性, 从而引入了不希望有的镜频分量。此外, 模拟元件的老化特性是一项随时间变化的过程特性, 难以控制。模拟电路的温度特性, 生产性能重复性等方面都是与数字电路无法相比的。
（中频数字化及数字式正交解调）
传统的雷达模拟正交接收机将输入的中频带通信号分别与正交的两路本振信号相乘，然后通过低通滤波滤除倍频分量，得到I, Q两路正交基带信号。其本振、混频、低通滤波均采用模拟技术实现，数字化在I, Q基带信号生成之后进行。由于模拟器件的一致性及稳定性都较差，所获得的两路正交通道很难在大的动态范围内保持高度的幅度一致性及相位正交性，一般需要对两通道间的偏差做大量的校正工作。
（一种宽带数字化雷达正交解调接收机的设计与实现）
If a high measurement accuracy is required, practical realization of this circuit becomes very difficult - the components must be extremely well selected and matched.
However, the existence of two detecting channels (in-phase and quadrature-phase) still remains a problem. Any mismatch between the analog circuits of the two channels will cause a systematic error. [7]
The classical quadrature coherent detector is implemented by analogy circuits and it has the problems such as the amplitude and phase of the two channels can not be kept constant
134
声明
for wideband signal and the steady of the circuit is greatly influenced by temperature.
（Analysis and Experimental Results of Digital Quadrature Coherent Detector）
In a quadrature radio receiver, it is well known that the outputs of I and Q channels may not have the same amplitude and their respective phases are not exactly 90\" apart, due to channel mismatches generated in the fabrication process. These mismatches are distributed on every analogue part of the channels, including the frequency synthesizer, mixer, anti-aliasing filter and analogue-to-digital converter. These mismatches can generate an image signal which will limit the dynamic range of a receiver [l]. This problem becomes more severe in a wideband receiver, since the mismatch can be a function of frequency[10].
（Wideband digital correction of I and Q mismatch in quadrature radio receivers）
Quadrature demodulation techniques are useful in communications, radar and electronic warfare systems. The classical approach for obtaining in-phase (I) and quadrature (Q) signals from an analog bandpass signal involves quadrature mixing and lowpass filtering. Analog implementations have problems with DC offsets and mismatches in the gain and phase of the I and Q channels. These errors result in the generation of spurious signals (e.g., spurious amplitude and phase modulations in the time domain or spectral images in the frequency domain). Digital approaches for performing quadrature demodulation on a sampled and digitized bandpass signal have potential advantages in these respects, but their computational cost is often an issue[11].
（The Design of Digital Quadrature Demodulators Using the Parks-McClellan Algorithm）
The processing of a real bandpass signal nominally centred on an intermediate frequency, fiF, to form an inphase (I) and quadrature (Q) signal representation is useful in coherent radar, communication and electronic warfare systems. The classical quadrature mixing and lowpass filtering approach for obtaining I and Q signals from an analog bandpass signal, shown in Figure 1, has been widely used. However, dc offsets and mismatches in the gain and phase of the I and Q channels result in the generation of spurious signals (e.g., spurious amplitude and phase modulations in the time domain or spectral images in the frequency domain). Analog quadrature demodulators often require careful matching of components andor subsequent post-processing to achieve acceptable accuracy [ 11. Digital approaches for performing quadrature demodulation on a sampled and digitized bandpass signal have potential advantages in these respects and can provideL ideal group delay characteristics if linear phase finite impulse response (FIR) filters are used. However, the computational cost of FIR filters is often an issue[12].
（Novel FIR Filter Designs for Digital Quadrature Demodulation）
The classical analog quadrature demodulation appproach has the problem that accurate
西安交通大学博士学位论文
amplitude and phase matching of the inphase and quadrature channels cannot be easily achieved [1]-[3]. Another error results from DC offsets introduced during the analog-to-digital conversion of the I and Q signals. These errors are particularly serious in coherent radar systems where they yield spurious sidebands at the image doppler frequency [2]. Consequently, digital approaches for performing quadrature demodulation on a digitized IF signal have attracted attention. Problems with the matching of analog components are eliminated and it is straightforward to achieve high accuracy through the use of filters having a sufficiently large number of coefficients. However, the computational cost can be a significant disadvantage when wideband signals must be processed in real-time[13].
（High Accuracy Digital Quadrature Demodulation）
Based on analog circuit techmques, traditional QCD methods lead to significant amplitude and phase errors, and low stability, as well as large zero-shift, so that it is rather hard to improve their performance further. This situation obstructs the improvement of over-all radar performance, i.e., the decrease of output signal-clutter ratio of MTI and the reduction of antenna side-lobe, etc. Benefitted from the fast advancement of digital techniques, DSP speed has been accelerated so much that it makes direct sampling possible from carrier frequency signals, which makes analog QCD evolve to digital QCD[']. Digital QCD becomes popular owing to its good agreement in amplitude and phase, as well as its good rejection to mirror frequency[14].
（Analysis of Mirror Frequency Rejection in Quadrature Coherent Detectors via Interpolation Method）
In radar, sonar and communication systems, there is often a need to translate a received bandpass signal to baseband and produce from it the I and Q components. This process is known as quadraturc demodulation and Fig. 1 is a block diagram of a conventional demodulator. It first downconverts the received signal to some intermediate frequency (iF) f0, and then multiplies the signal by cos伪at) and sin和pt) to generate baseband outputs with } deg phase difference. The I and Q samples are formed after lowpass filtering and sampling. This method has been adopted extensively in many applications. There is, however, an inherent difficulty in obtaining perfect balance because severe gain and phase errors can occur. These errors were investigated in [t) and the phase error can be as high as 2 to 3 deg }2}. This amount of error is not acceptable for some signal processing tasks and other quadrature demodulation techniques that can reduce these errors arc required[15].
（A Digital Quadrature Demodulation System）
1.4 数字正交解调
随着信号处理技术的提高，近年来出现了直接对中频信号进行采样并对采样信号进行数字信号处理的正交采样方法［24－28］。其原理框图如图2-2所示。即对中频信号采样后，用数字方法形成I(In-phase), Q(Quadrature)信号。与传统的模拟正交采样方
136
声明
法比起来，这样得到的正交信号的一致性好、精度高，而且具有数字电路的其他优点，从而在很大程度上提高了系统性能。
24 C. Ho, Y. T. Chan. A Digital Quadrature Demodulation System. IEEE Trans. on AES,1996,32(4):1218一1226
2 _5雷文，龙腾，曾涛等.脉冲雷达中频采样系统的镜频抑制性能分析与参数估计.电子学报，2001,29(12):1585-1588
26 H. L. Liu, A. Ghafoor. A New Quadrature Sampling and Processing Approach. IEEE Trans. on AES,1989,25(5):733一747
27林云松，黄勇.宽带数字正交技术及性能分析.电r科技大学学报，1999,(1):20-23 28苏涛，强生斌，吴顺君.数字正交采样和脉压的高效算法及实现.现代雷 达，2001,23(1):39-41
西安交通大学博士学位论文
1.4.1 缺点：
In principle, digital demodulation does not imply the use of high sampling rate ADCs since undersampling techniques can be applied. That is, it is not necessary to employ ADCs with sampling frequencies above 2fo. However, the aperture time of the ADC is a critical parameter that is usually specified in accordance to the ADC sampling rate and, at the end, ADCs with high sampling rates are required. A constant aperture delay of the ADC affects in the same manner to the resistance and reactance components and causes a constant phase error that can be canceled by calibration. However, the aperture fitter, due to the clock fitter or to ADC sample-and- hold fitter, will cause a random error that can be regarded as noise. This source of error will affect seriously to the phase measurement. For instance, Figure C. 7 shows the standard deviation of the phase angle that the results from a Matlab simulation of the measurement of a 100 kHz signal when the fitter rangers from 1 ns to 100 ns.
It must be mentioned that in the case that the system response is linear, in principle, it is also possible to obtain the impedance at any frequency by injecting other signals than sinusoids, for instance, rectangular pulses. This can be achieved by performing the Fourier transform to the obtained time response after applying the pulse. Such strategy has been used by some authors in the bioimpedance field [13-15] and it can be beneficial in terms of measurement time.[2]
（Instrumentation for electrical bioimpedance measurements）
138
声明
1.5 基于相敏解调的矢量阻抗法
1.5.1 缺点：
从以上的分析可以看出，传统的矢量阻抗法在进行测量时，量程电阻和被测阻抗上某一时刻的电压是分四次测出的，因此必须保证开关切换和坐标矢量旋转前后输入电路的电流相等，这就要求测试信号源具有相当高的频谱纯度和幅值稳定度。目前信号源电路多采用数字合成技术，此方法通常是在ROM(Read Only Memory)中存储一个周期的正弦曲线采样点表，每一个存储单元存储的样点数据和地址之间的关系与正弦波的正弦幅值和时间轴的关系是一致的。这样，当按顺序逐单元读出ROM的样点数据时，就能得到量化了的正弦曲线，若周期的重复这一过程，少}将数字量经D/A转换与平滑滤波后输出，从而得到连续的正弦波信号。一个周期的样点数据个数和D/A转换器的量化误差都会影响输出的正弦波的质量，尤其是在频率很高时，低失真度的正弦信号很难获得，这就限制了测量精度的提高和测量范围的扩大。
西安交通大学博士学位论文
1.5.2 synchronous sampling
One way to avoid this problem is by using synchronous sampling detection. This method was described by Pallas-Areny and Webster [3] - see Fig. 3.
This method reduces the number of analog components. Sampling frequency can be lower than the frequency of the measured signal, which is important at higher measurement frequencies.
The main disadvantage of this method is a very large aperture-time relative error of the S/H - ADC at higher measurement frequencies [3]. Also, for wide frequency range measurements a variable sampling frequency and a variable high pass filter are required[16].
140
声明
1.6 The frequency response analysis technique (sine correlation)
The frequency response analysis technique (sine correlation) is excellent at extracting the required signal component from noise. This is achieved by correlating the input signal with reference sine waves and integrating the result over a number of complete cycles of the sine wave. Harmonics are rejected by the
correlation process, and noise is rejected by averaging the signal over a number of cycles. Using this technique very small signals can be identified in the presence of very high levels of harmonics and noise. Frequency Response Analyzers (FRA’s) usually have separate analysers for each input. This allows fast simultaneous measurements to be achieved. Also, since measurements are taken at exactly the same time on each input (in parallel) any errors due to variations in the signals with time are cancelled.
(Impedance Measurement Techniques- Sine Correlation)
西安交通大学博士学位论文
1.6.1 \"zero-crossing\" method
There are several methods for measurement of modulus and phase angle or components of impedance [2.4]. Three techniques are usually used:
the bridge method;
the method based on measurement time shift between signals of current and voltage; the coherent detection method. [17]
Measurement of impedance is based on tetrapolar current method, in which two different electrodes are used for injection of current and two another electrodes for impedance signal detection. Separation of the current application and signal detection electrodes virtually eliminates the influence of the electrode tissue impedance. The current injected to the tissue is independent of the measured impedance. Sinusoidal current is generated by a specialised voltage-controlled oscillator. This solution allows to change the w e n t frequency manually or automatically.[17]
Block scheme of the device is presented in the figure 1. Voltage controlled oscillator VCO ( 1) generates sinusoidal signal with constant voltage over the entire range of applied frequencies. Voltage from oscillator is fed to voltage-current converter (2) which provides high output impedance and practically eliminates the influence of the electrode-tissue contact. Instrumentation amplifier (3), receiving signals from the detection electrodes, has a frequency bandwidth 2-200 kHz. Amplitude of the sinusoidal signal from amplifier, proportional to modulus of the measured impedance lZo\\, is fed to the peak detector (4). Phase angle (p) is measured as a time of current - voltage shift (7) in relation to cycle duration (T). It can be expressed by the formula: p=k×IT where k stands for a coefficient of proportionality.[17]
output signal from the detectmg amplifier in phase with the voltage and the other signal in phase with the current are fed to the inputs of fast comparators (5). These comparators switch at the zero-crossing of the input signal. After further processing pulses of duration 7 and periodicity f (lA') are obtained. Next stages of the processing allow to obtain average amplitude of the pulse signal, its amplification and transferring it over the isolation barrier. The isolation barrier is designed for transfemng the two independent analogue signals of the modulus Zo 1 and the phase p in the bandwidth of 0 to 100 Hz. The range of barrier output signal is 0 to 5 V. Power supply gives voltages +15 and - 15 V and is hlly isolated.
The specifications of designed device are as follows:
method of the measurement: tetrapolar, - application cwrent source: amplitude 1 mA, output
impedance > 100 kD (for load impedances up to 5 kD),
frequency control infinitely variable in the range of 2 to 200 kHz, in two sub-ranges 2-20 kHz and 20
- 200 Hz,
measurement range of phase: 0 to 45\" with the resolution of 0.1 ', result shown on a 3 digit display, isolating barriers of power supply and output signals p and IZoJ. Analogue output signals are
subsequently analysed by a PC computer. The computer calculates the real and imaginary parts of the
142
声明
bio-impedance (R &) and theirs changes (.hR e). There are two additional modules providmg reference signals: ECG amplifier and impedance cardlograph with a peak detector. The electrical safety parameters of the device are conform to the requirements of the IEC-601 Electrical Safety Standard.[17]
西安交通大学博士学位论文
1.6.2 Driven shield connections to the sample
Accurate measurements at high frequency are a specific problem due to errors introduced by input and cable capacitance. One technique which has been widely used to reduce the effects of capacitance is to position electrometer buffer amplifiers close to the sample. These buffers are able to drive the cable and input capacitance of the equipment and therefore reduce errors in voltage measurements. However, these external amplifiers require a power supply which usually involves extra cabling. In addition, where the temperature of the sample is to be varied, the accuracy of measurements may be effected since the buffers must be positioned close to the sample and are therefore subject to the same temperature variations. An alternative solution is the use of driven shield cables. This technique replicates the signal waveform (which appears on the cable inner), onto the cable shield in order to minimise leakage current flow between the cable inner and the shield. Since no current flows between the cable inner and shield, the impedance appears to be very large and therefore the effects of the cable and input capacitance are minimised。
This method allows the high impedance buffers to be kept within the instrument. As an external power supply unit is not required for the electrometers, the cabling is reduced and the sample temperature may be changed without affecting the accuracy of the results.[18].
1.6.3 Balanced Generator
One problem associated with four terminal measurements on bio-materials is the requirement to study the impedance of the sample in the presence of relatively high electrode impedances (i.e. to reject the voltages across the electrodes in order to obtain accurate measurements of the voltage across the sample itself).
Figure 5 shows a typical measurement situation where the electrode impedance is ten times higher than the sample impedance which is required to be measured. In the example shown, the sample impedance is simulated by a 1Kohm resistor, and the electrode impedances are simulated by two 10kohm resistors. Using conventional measurement techniques, the AC stimulus voltage is applied to GenHi while GenLo is grounded, and the current through the sample and voltage drop across the sample (between V Hi and V Lo) are measured in order to compute the impedance of the sample.
144
声明
In this example, the FRA is required to measure a small voltage difference between two relatively high voltage signals which appear at V Hi and V Lo. For example, if 1 volt is applied between GenHi and GenLo, less than 50 mVolts appears across the sample (V Hi - V Lo) whereas the voltage on the V Lo measurement connection is approximately 500mV. The FRA is therefore required to measure a 50 mVolt “difference” signal in the presence of 500 mVolts which leads to errors which are referred to as “common mode” errors. In addition, the relatively high voltage on the V Lo connection causes some current which has passed through the sample impedance to leak to earth via the input / cable capacitance on V Lo instead of being measured by the current measurement circuit.
The use of a balanced generator (see Fig. 6) reduces errors due to common mode voltages and earth leakage allowing more precise measurements of the sample impedance. This is achieved by adjusting the GenHi and GenLo signals in order to provide a balanced stimulus to the sample which has the effect of making the voltage which appears at V Lo as close as possible to earth voltage, (i.e. zero volts).
The balanced generator can also cope with extremely difficult measurement situations where for
西安交通大学博士学位论文
instance the electrode impedances are not equal. This is also typical for bio-impedance measurements where it is difficult to obtain reproducable electrode contacts onto skin. In this case the signals on GenHi and GenLo are set to different voltage levels in order to again achieve zero volts on the V Lo connection.[18]
全身阻抗的测量：
From：Influence of various factors on the measurement of multifrequency bioimpedance
1. Liu, J.-G., U. Frühauf, and A. Schönecker, Accuracy improvement of
impedance measurements by using the self-calibration. Measurement, 1999. 25: p. 213-225.
2. Cano, A.I., Contributions to the measurement of electrical impedance for living tissue ischemia injury monitoring. PhD Dissertation of Universitat Politècnica De Catalunya, Barcelona Spain, 2005.
3. Songer, J., Tissue ischemia monitoring using impedance spectroscopy: Clinical evaluation. M.S. thesis of Worcester Polytechnic Institute in USA, 2001.
4. Yufera, A., et al., A tissue impedance measurement chip for myocardial ischemia detection. Circuits and Systems I: Regular Papers, IEEE Transactions on [see also Circuits and Systems I: Fundamental Theory and Applications, IEEE Transactions on], 2005. 52(12): p. 2620-2628.
5. 尤富生, et al., 生物电阻抗模拟解调技术的研究. 北京生物医学工程, 2004(01): p. 21-24.
6. 尤富生, et al., 多频生物电阻抗高精度模拟解调技术的研究. 医疗卫生装备, 2004(07): p. 10-11.
7. Ristic, B., S. Kun, and R.A. Peura, Development of an Impedance
Spectrometer for Tissue Ischemia Monitoring: Application of Synchronous Sampling Principle. IEEE, 1995: p. 74-75.
8. Pallas-Areny, R. and J.G. Webster, Bioelectric impedance measurements using synchronous sampling. IEEE Transactions on Biomedical Engineering, 1993. 40(8): p. 824-829.
146
声明
9. Gonz´alez, C.e.A., et al., Impedance spectroscopy for monitoring ischemic injury in the intestinal mucosa. Phsiological Measurement, 2003. 24(3): p. 277-289.
10. PUN, K.-p., J.E. FRANCA, and C. Azeredo-Leme, Wideband digital correction of I and Q mismatch in quadrature radio receivers. Proceedings of IEEE International Symposium on Circuits and Systems, 2000: p. V661-V664.
11. Inkol, R.J. and P. Truong Npuyen Duc. The design of digital quadrature demodulators using the Parks-McClellan algorithm. in Electrical and Computer Engineering, 1998. IEEE Canadian Conference on. 1998.
12. Inkol, R.J., Novel FIR Filter Designs for Digital Quadrature
Demodulation. Proceedings of the 1999 IEEE Canadian Conference on Electrical and Computer Engineering, 1999: p. 733-738.
13. Inkol, R.J., M. Herzig, and R. Saper. High accuracy digital quadrature demodulation. in Microwave Symposium Digest, 1995., IEEE MTT-S International. 1995.
14. Peng, S., Z. Yang, and W. Dong. Analysis of mirror frequency rejection in quadrature coherent detectors via interpolation method. in Environmental Electromagnetics, 2003. CEEM 2003. Proceedings. Asia-Pacific Conference on. 2003.
15. Ho, K.C., Y.T. Chan, and R. Inkol, A digital quadrature demodulation system. Aerospace and Electronic Systems, IEEE Transactions on, 1996. 32(4): p. 1218-1227.
16. Ristic, B., S. Kun, and R.K. Peura. Development of an impedance spectrometer for tissue ischemia monitoring: application of synchronous sampling principle. in Bioengineering Conference, 1995., Proceedings of the 1995 IEEE 21st Annual Northeast. 1995.
17. Palko, T., F. Bialokoz, and J. Weglarz. Multifrequency device for measurement of the complex electrical bio-impedance-design and application. in Engineering in Medicine and Biology Society, 1995 and 14th Conference of the Biomedical Engineering Society of India. An International Meeting, Proceedings of the First Regional Conference., IEEE. 1995.
18. Hinton, A.J. and B. Sayers. Advanced Instrumentation for Bioimpedance Measurements. http://www.solartronanalytical.com/ 1998 [cited.
因篇幅问题不能全部显示，请点此查看更多更全内容