Добавил:
Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:

Sb97789

.pdf
Скачиваний:
4
Добавлен:
13.02.2021
Размер:
577.44 Кб
Скачать

sensitive to the condition of the skin and are highly susceptible to motion artifacts. An easier to use and less obtrusive technology is called for to match the advancements made in wireless body sensor networks.

In contrast to wet and dry contact sensors, non-contact capacitive electrodes do not require an ohmic connection to the body. For body sensor applications, this offers numerous advantages since non-contact electrodes require zero preparation, are completely insensitive to skin conditions and can be embedded within comfortable layers of fabric. While the concept of non-contact biopotential sensors is not new, with the first working device reported decades ago, a practical device for patient use has yet to be realized. More recently, several authors have presented results from designs utilizing the latest in commercially available discrete low noise amplifiers, including some wireless designs. In all cases, the challenges in noncontact sensing have lead to many clever, and often-times, proprietary circuit designs in an effort stabilize the electrode’s input.

In this paper, we expand on the work previously presented by building a sensor with much improved noise performance. In addition, the full design and schematics for a wireless, non-contact EEG/ECG system with features designed for specifically for practical body sensor networks is described.

System design

The system contains a set of non-contact biopotential electrodes connected along a single common wire. The sensors can be either in direct contact with the skin or embedded within fabric and clothing. A small base unit powers the entire system and contains a wireless transmitter to send data to a computer or other external device. Near the base unit, a single adhesive or dry contact sensor placed anywhere convenient is used to establish the ground reference for the system.

А. Electrode Construction

Each electrode is constructed from two, US quarter sized, PCBs. The upper PCB contains a low noise differential amplifier and a 16-bit ADC. Rather than outputting a single analog signal, the electrode outputs the digitized value, which can be carried in serial daisy chain to drastically reduce the number of wires needed. A miniature 10-wire ribbon cable carries the power supply, digital control as well as analog common mode reference from electrode-to-electrode.

The lower PCB contains the INA116 configured as an ultra-high input impedance amplifier. The bottom surface of the PCB is a solid copper fill, insulated by solder mask, that functions as the electrode. This surface forms a coupling capacitor with the body. An active shield formed by in a solid inner plane protects the

21

electrode from external noise pick-up. To minimize the shield capacitance, an extra thick PCB is used for the electrode.

B. Front End Amplifier

Designing an ultra-high input impedance amplifier with low noise levels is the main challenge in implementing non-contact electrodes. Signal sources from the body (EEG/ECG) can be thought of as a voltage source, Vs, connected to the input of an amplifier via a small coupling capacitance, Cs. All real amplifiers will also have some finite resistance, Rb, and input capacitance. A small amount of positive feedback can be applied through, Cn, to neutralize the effect of the input capacitance for better channel matching and CMRR.

Important noise sources include the input referred voltage noise of the amplifier, Vna, the input current noise, Ina and the additional current noise, In b, due to the leakage and conductance of the biasing element. The current noise contribution will either 4kTR thermal noise for a resistive device or 2qI shot noise for a PN junction. Bootstrapping can be used to electronically boost the effective impedance of the biasing element, but the noise contribution depends only on the physical resistance or leakage current, illustrating the challenge in finding suitable components for a non-contact sensor.

The equation clearly shows the effect of the parasitic input capacitances and leakage currents on the noise performance of the amplifier and the difficulty in designing a non-contact electrode. Any excess input capacitance will directly multiply the effect of the amplifier’s input voltage noise as Cin+Cn> Cs.

Furthermore, since biopotential signals are at low frequencies (.1-100 Hz), even small amounts of current noise become integrated into large amounts of input voltage noise. This necessitates an amplifier with both very low input and guard capacitance as well as almost zero leakage currents.

The INA116 by Burr-Brown is an amplifier that is well-known for ultra-high input impedance applications by virtue of its extremely low current noise. However, any circuit introduced to bias the inputs will significantly degrade the noise performance of the amplifier. An extremely difficult to obtain resistor would be required to match the current noise specification of the INA116. Fortunately, it was found during experiments that the INA116 would reliably charge a floating input to a point inside the allowable input range shortly after power-up purely through leakage currents, removing the need for any external bias network. To remove drift and DC offsets, a low-passed version of the input signal was taken from the noninverting input’s guard and connected to the inverting input. This effectively performs

22

AC coupling without degrading the input impedance and centers the output to mid-rail for maximum signal swing.

It is also worth noting that a similar ’bias-free’ technique can also be applied successfully to rail-to-rail input/output operational amplifiers configured in unity gain by simply leaving the non-inverting input floating and AC coupling the output to the next stage. As long as DC measurements are not needed (as in EEG/ECG applications), the amplifier is guaranteed to operate somewhere inside the supply rails. We have successfully tested this with the LMC6081 and LMP7702 operational amplifiers. The overall performance is comparable to the INA116 circuit. More detailed characterization of the design will be explored in a future paper.

To ensure that the electrode’s gain is constant over a wide range of coupling distances, a small amount of positive feedback, adjusted byR3, is applied back to the input throughC2. Each electrode is carefully calibrated at a test-bench byadjustingR3until the gain is constant for different coupling distances. In practice, however, the capacitance neutralization circuit is not needed since the system is wireless with a floating ground – no significant 50/60 Hz mains interference is observed even with mismatched electrodes.

C. Differential Amplifier and ADC

The output of the INA116 is coupled to a differential gain amplifier through an additional high-pass filter with a cutoff of :1 Hz to remove the relatively high DC offset of the INA116. The LTC6078 micro power operational amplifier was chosen for it’s excellent low noise and low offset characteristics. A simple noninverting differential gain stage of 40:1 dB was implemented by connecting the electrodes together through the node Vcm, which is carried in the daisy chain. This constructs what is essentially a multi-channel instrumentation amplifier to remove the common-mode noise while amplifying the local biopotential signal.

Here it is shown the full measured transfer function of the front-end and differential amplifier. A gain of 46 dB and cut off frequencies of 0:7 Hz and 100 Hz were obtained as expected.

The total in-band input referred noise was measured to be 3.8µ VRMS. At present, it appears that the noise pickup from external sources is as problematic, if not more so, as the intrinsic noise sources in the circuit, even with the active shield layer. This is not surprising due to the sensitivity of the ultra-high impedance input node. Future versions of the electrode will incorporate more comprehensive shielding strategies than a simple inner PCB plane.

23

Conclusion

We present a wireless body sensor network for high quality EEG/ECG recordings utilizing non-contact electrodes. The full schematics for building the simple, low noise capacitive electrode are presented. Future work will focus on miniaturizing and better packaging the electrode as well as reducing the power consumption of the digital and wireless transmitter components.

Дополнительные задания к тексту:

1.Грамматические. Найдите предложения в пассивном залоге. Объясните значение глаголов. Попытайтесь перестроить предложения так, чтобы использовать активный залог. Найдите в тексте причастия. Объясните их использование. Найдите в тексте инфинитивные обороты, объясните особенности их перевода на русский язык.

2.Лексические. Выпишите слова, относящиеся к базовой лексике по теме статьи. Выпишите слова, относящиеся к общетехнической лексике. Выпишите специализированные термины и профессионализмы, использованные

встатье.

24

Тема 2. Кардиограф

Прочитайте текст и перескажите его. Переведите текст и выпишите ключевые слова.

Blood Pressure Monitor. Fundamentals and Design

This application note demonstrates the implementation of a basic blood pressure monitor using Freescale products. The blood pressure monitor can be implemented using any of the Freescale medical oriented MCUs: Kinetis MK53N512 and Flexis MM members MC9S08MM128 and MCF51MM256 embedding a 16bit ADC, 12-bit DAC, 2 Programmable-Gain Op-Amps, 2 TRIAMPS, Analog Comparators, and Vref generator. The K50 family can also perform DSP instructions for signal treatment and MCF51MM can perform multiply and accumulate (MAC) instructions. This document is intended to be used by biomedical engineers, medical equipment developers, or any person related with the practice of medicine and interested in understanding the operation of blood pressure monitors. Nevertheless, it is necessary to know fundamentals of electronic, analog, and digital circuits.

Arterial pressure

Arterial pressure is defined as the hydrostatic pressure exerted by the blood over the arteries as a result of the heart left ventricle contraction. Systolic arterial pressure is the higher blood pressure reached by the arteries during systole (ventricular contraction), and diastolic arterial pressure is the lowest blood pressure reached during diastole (ventricular relaxation). In a healthy young adult at rest, systolic arterial pressure is around 110 mmHg and diastolic arterial pressure is around 70 mmHg. Blood flow is the blood volume that flows through any tissue in a determined period of time (typically represented as ml/min) in order to bring tissue oxygen and nutrients transported in blood. Blood flow is directly affected by the blood pressure as blood flows from the area with more pressure to the area with less pressure. Greater the pressure difference, higher is the blood flow. Blood is pumped from the left ventricle of the heart out to the aorta where it reaches its higher pressure levels. Blood pressure falls as blood moves away from the left ventricle until it reaches 0 mm Hg, when it returns to the heart’s right atrium.

25

Fig. 1. Pressure changes on blood vessels

Blood pressure monitor operating principle

Blood pressure monitor operation is based on the oscillometric method. This method takes advantage of the pressure pulsations taken during measurements. An occluding cuff is placed on the left arm and is connected to an air pump and a pressure sensor. Cuff is inflated until a pressure greater than the typical systolic value is reached, then the cuff is slowly deflated. As the cuff deflates, when systolic pressure value approaches, pulsations start to appear. These pulsations represent the pressure changes due to heart ventricle contraction and can be used to calculate the heartbeat rate. Pulsations grow in amplitude until mean arterial pressure (MAP) is reached, then decrease until they disappear.

Oscillometric method determines the MAP by taking the cuff pressure when the pulse with the largest amplitude appears. Systolic and diastolic values are calculated using algorithms that vary among different medical equipment developers. Freescale Blood Pressure Monitor calculates the systolic and diastolic pressure by considering that systolic pressure is approximately equal to the pressure measurement taken in the cuff when a pulse with 70% of the amplitude of the MAP pulse appears while the cuff pressure is above the MAP value. Diastolic pressure is approximately equal to the cuff pressure value registered when a pulse with 50% of the MAP pulse amplitude appears while the cuff pressure is under the MAP value.

Blood pressure monitor implementation

Blood pressure monitor is implemented using Freescale medical-oriented Kinetis K53 MCUs and Flexis MM devices, which feature the following characteristics:

16-bit ADC;

12-bit DAC;

2x programmable gain operational amplifiers (OpAmps);

2x transimpendance amplifiers (TRIAMPS);

26

Vref generator;

Set of DSP instructions including MAC (Only K5X Family);

Multiply and Accumulate (MAC) instruction on MCF51MM.

Freescale medical-oriented MCUs reduce the Bill Of Materials (BOM) required for medical applications and provide great processing capabilities ideal for medical equipment. Nevertheless, some external circuitry is needed for pressure sensing and cuff control.

MED-BPM analog front end

MED-BPM Analog Front End (AFE) is a demo board designed for work as a blood pressure monitor in conjunction with a Freescale medical-oriented MCU. MED-BPM communicates with the MCU using the medical connector, and allows for easy prototyping and reduced time to market by using the Freescale Tower System.

Arm cuff pressure control

MED-BPM works using an oscillometric method for blood pressure measurements. This is a noninvasive method which requires an external arm cuff in order to occlude the patient’s arm and detect the systolic and diastolic arterial pressure. The arm cuff is inflated using an external air pump controlled with an MCU GPIO pin, and deflated by activating an escape valve with another GPIO pin. Because the current provided by the USB port (500 mA) is not enough to activate the air pump and the valve (600 mA), those external components are activated by using an external power source which provides sufficient current. An optocoupler is needed for coupling MCU control signals with the components to activate.

Output from the optocoupler is connected to a MOSFET working as a switch, so the air pump and valve mechanisms can be activated successfully.

Signal filtering and amplification

This stage is composed of three filters, one buffer circuit, and one noninverting amplifier.

Filters are first order RC passive type, and the cut-off frequency is described by the following formula.

1.A signal is passed through a 10 Hz RC low-pass filter (LPF) composed of a resistor and a capacitor in order to remove high-frequency noise.

2.Then the signal is passed through a buffer circuit consisting of a single OpAmp in buffer mode to couple the signal to the sensor. The output from the buffer circuit is where the arterial pressure measurements are taken.

27

3.The signal is then filtered again with a 2.2 Hz RC high-pass filter which removes high-frequency noise and gets a cleaner signal for amplification.

4.This signal is amplified using a non-inverting amplifier composed by a second Op-Amp and two resistors, (100 kΩ and 1 kΩ) generating a gain of 101 so cuff oscillations can be distinguished better.

5.After this stage, the signal is filtered again with another 10 Hz RC LPF so high-frequency noise can be removed.

Задания к тексту:

1.Выпишите ключевые слова. Разделите их на две группы – профессиональной лексики и общетехнической. Отберите самые важные для описания содержания статьи.

2.Найдите в тексте статьи предложения, содержащие грамматические обороты с инфинитивом, герундием, причастием в активном и пассивном залоге. Объясните особенности их использования и перевода.

3.В чем заключается научная новизна исследования?

4.Подготовьте свой вариант заключения представленного в статье мате-

риала.

28

Тема 3. Глюкометр

Задание 1.Прочитайте аннотацию. Составьте глоссарий слов и выражений, которые, на ваш взгляд, могут встретиться в статье, посвященной данной теме.

This tutorial introduces the different types of glucose meters. It explains the different types of calibration used for test-strips, and discusses other variables that designers must consider when selecting products.

Задание 2. Ознакомьтесь со вступительной частью статьи. Что вы знаете о глюкометрах? Какие они бывают? Для чего нужны глюкометры? Какие могут быть усовершенствования в производстве глюкометров?

Blood glucose meters and other home medical devices today are small, portable, and easy to use. The mark of a good meter is one that the patient will use regularly and that returns accurate and precise results. Over the past few years the trend with blood glucose meters has been to maximize patient comfort and convenience by reducing the volume of the blood sample required. The blood sample size is now small enough that alternate-site testing is possible. This eliminates the need to obtain blood from the fingers and greatly reduces the pain associated with daily testing. Accurate and precise results have been increased by using better test strips, electronics, and advanced measurement algorithms. Other conveniences include speedy results, edge fill strips, and illuminated test strip ports, to name just a few.

Задание 3. Прочитайте всю статью. Переведите ее и перескажите основное ее содержание. Сформулируйте выводы.

Blood Glucose Meter

There are continuous and discrete (single-test) meters on the market today, and implantable and noninvasive meters are in development. Continuous meters are by prescription only and use a subcutaneous electrochemical sensor to measure at a programmed interval. Single-test meters use electrochemical or optical reflectometry to measure the glucose level in units of mg/dL or mmol/L.

The majority of blood glucose meters are electrochemical. Electrochemical test strips have electrodes where a precise bias voltage is applied with a digital-to- analog converter (DAC), and a current proportional to the glucose in the blood is measured as a result of the electrochemical reaction on the test strip. There can be one or more channels, and the current is usually converted to a voltage by a transimpedance amplifier (TIA) for measurement with an analog-to-digital converter (ADC). The full-scale current measurement of the test strip is in the range of 10µA

29

to 50µA with a resolution of less than 10nA. Ambient temperature needs to be measured because the test strips are temperature dependent.

Electrochemical blood glucose meter

Optical-reflectometry test strips use color to determine the glucose concentration in the blood. Typically, a calibrated current passes through two light-emitting diodes (LEDs) that alternately flash onto the colored test strip. A photodiode senses the reflected light intensity, which is dependent on the color of the test strip, which, in turn, is dependent on the amount of glucose in the blood. The photodiode current is usually converted to a voltage by a TIA for measurement with an ADC. The full-scale current from the photodiode ranges from 1µA to 5µA with a resolution of less than 5nA. Ambient temperature is required for this method.

Optical-reflectometry blood glucose meter

Test-Strip Calibration

The test strips usually need to be calibrated to the meter to account for manufacturing variations in the test strips. Calibration is done by entering a code manually or by inserting a memory device from the package of test strips. An EPROM or EEPROM memory device enables additional information to be transferred, which is a significant advantage over manually entering a code. A 1-Wire® memory device provides an additional benefit, because the unique serial identification number in each 1-Wire device ensures that the proper test strip is used.

Some meters use test strips that do not require any coding by the user. This "self-calibration" can be accomplished three ways: with tight manufacturing controls, built-in calibration on each test strip, or built-in calibration on a pack of test strips loaded into the meter. A pack of test strips inside the meter also facilitates testing because these often small strips do not need to be handled and inserted by the user.

Glucose Meter Solutions

Accuracy and Precision

Both optical-reflectometry and electrochemical meters need to resolve currents in the single-digit nano-amp range. To meet the error budget for a meter, components must have extremely low leakage and drift over supply voltage, temperature, and time once the meter has been calibrated during manufacture. An operational amplifier's key specifications are ultra-low input bias current (< 1nA), high linearity, and stability when connected to a capacitive electrochemical test strip. The operational amplifier is typically configured as a TIA for both types of meters. A voltage

30

Соседние файлы в предмете [НЕСОРТИРОВАННОЕ]