Analog Circuit Design For Hearing Aid Computer Science Essay

This papers proposes a possible parallel circuit design for a low cost programmable hearing assistance solution. Along with the design methodological analysis this papers throws visible radiation on the practical issues such as aliasing, noise, filter choice for design, amplifier choice etc. , that are critical considerations for parallel circuit of hearing assistance design. This design is suited ( but non limited ) for a programmable hearing assistance design necessitating digital signal processing.

Keywords – Antialiasing filter, Noise, Hearing Aid, filters, programmable hearing assistance design, Low cost, Deafness, hearing damage, Automated Gain control.

Introduction

Our organic structure is made up of different natural detectors which help us in detecting our environment. Our step of environing is based on the capablenesss of these detectors. Hearing is one such characteristic step of our environing. Ear, which is portion of auditory system of the organic structure, detects alteration in force per unit area of the medium or quivers to comprehend sound. Although the Auditory system looks same for everyone but the hearing capacity is non same for all. We perceive sounds otherwise based on our hearing capablenesss. A spectrographic analysis of the hearing capacity of a individual gives information about hearing profile of that individual. Partial or complete incapableness of a individual to comprehend sound without any external assistance is hearing loss or hearing damage. A survey done by World Health Organization titled “ Global load of hearing loss in the twelvemonth 2000 ” states that more than 250 million people in the universe suffer from Hearing damage and it is dispersed across the ages [ 1 ] . The survey [ 1 ] besides states that merely 10 % of the hearing impaired people use hearing assistance where in developed states it is used by around 49 % of hearing impaired population but in developing states less than 1 % hearing impaired population is utilizing hearing assistance. There is immense demand of low cost hearing assistance device that can be used in developing states.

Hearing assistance solution are of two types: Analogue and Digital hearing AIDSs. Analog hearing AIDSs cheaper than digital hearing AIDSs but are less programmable. Besides the noise degree as compared with digital hearing AIDSs is high. Analog hearing AIDSs supply really less flexibleness for accommodations with the hearing profile of the individual. Whereas, Digital hearing AIDSs sometimes known as “ true hearing AIDSs ” provides a great flexibleness for accommodations as per the hearing profile. Digital hearing AIDSs use noise decrease algorithms to cut down the noise. Digital hearing AIDSs use digital signal processing algorithms but still necessitate an parallel circuit to capture sound from mike. In the design, we are seeking to treat signals in both parallel and digital sphere in order to acquire a suited tradeoff between cost, power ingestion and quality. The design utilizes the capacity of digital noise decrease algorithms and programmability along with the decrease of load of digital processing by utilizing parallel circuit assisting in choosing cheaper digital processor. This paper focuses on the parallel circuit required for the design. Rest of the design are covered in [ 2 ] and [ 3 ] .

Design Constraints and Methodology

Following are the major restraints that were considered for the design of low cost hearing assistance design

Degree of hearing loss

Degree of hearing loss defines the degree of hearing loss suffered by an person. Degree of hearing loss besides defines the addition demand and reassign map of digital Hearing loss compensation filter. Based on grade, hearing loss are of following types

Mild hearing loss ( 25 and 40 dubniums SPL ) ,

Moderate hearing loss ( 40 to 70 dubniums SPL ) ,

Severe hearing loss ( 70 to 95 dubniums SPL ) .

A general hearing assistance should be able to provide the demand of all four types of hearing loss but in order to cut down the cost and power the design can be modified and limited to supply a hearing assistance boulder clay moderate hearing loss. A design should be able to supply 95 dubnium SPL to provide hearing loss of all grades.

Hearing loss profile:

As we all perceive sound otherwise, we have different spectrographic response to sound. For a normal individual this does n’t change much ( 0 – 20dB ) for about all frequences. For Hearing impaired people the response is nonlinear and dependent on frequence, which means for some of the hearing loss profiles, low frequence scope is non hearable and for some high frequence scope is non hearable. In order to counterbalance for the loss, the filter should supply adequate accommodation capableness that can cover all possible hearing loss profiles.

Frequency scope:

Hearing AIDSs available in market are featured to take attention of demand of different types of hearing profiles. These characteristics add cost to the device. There are device in which one can put manners like address, music, noise etc. to set the device map as per the demand. These characteristics are available merely in digital hearing AIDSs and necessitate separate memory for each hearing manner, which besides makes the device dearly-won. Besides for treating demands these device require digital signal processors for the processing, which are once more dearly-won.

A Low cost hearing assistance can non supply the above said characteristics but it should provide the basic demand of the ability to hear speech signal. Address signals range from 125 Hz to 4 KHz.

Antialiasing

Digital hearing assistance design requires antialiasing filter to fulfill trying theorem. As mike can capture a broad spectrum of frequence scope, which can do aliasing because of the signals captured above Nyquist frequence. The intent of Antialiasing filter is to take the signals above Nyquist frequence. A filter can non take the signal wholly but can rarefy the signal. An fading of the order of 60 dubnium is rather safe for hearing assistance design. As our frequence of involvement prevarications under 4 KHz, so our trying frequence should be above 8 KHz. But maintaining 8 KHz as trying frequence will do higher order of antialiasing filter as a crisp 60 dubnium fading after 4 KHz. Which will necessitate a really high order filter ( more than 10 ) , which is practically non advisable. A good design can be where Analog Antialiasing filter can be relaxed by increasing sampling frequence. In our design we identified 800 KHz as a suited sampling frequence which can loosen up parallel filter up to a degree where power ingestion by parallel portion could be reduced to minimum and treating power of digital filter could hold maximal budget of clock rhythms for a digital antialiasing filter after other digital signal processing [ 2 ] .

Preamplifier addition.

A preamplifier circuit is required to magnify the signals captured by mike to the degree it can expeditiously use the electromotive force mention of Analog to digital converter. If the electromotive force mention of the Analog to digital converter is 5V, so signal should be amplified till it reaches 5V extremum to top out. Electret capacitor mikes produces a electromotive force of scope 5mV to 10mV. A preamplifier amplifies this signal to the electromotive force degree that can use all the spots of Analog to Digital Convertor when given input in order to utilize ADC expeditiously, i.e. of the order of mention electromotive force. By and large, the addition required is of the order of one 1000. Though the overall addition of preamplifier modified by the machine-controlled addition control.

Heterogeneous sound in the surrounding:

Signal beginning for a hearing assistance device is extremely heterogenous, power degree of sound in environing can alter at any point of clip. Hearing assistance should be smart plenty to change the addition feeling the power degree of the surrounding in order to protect any injury to the ear. Automatic Gain Control is used to compact the loud passages of the sound and magnify the weak sound signals to that of scope of signal processing scope. Automated Gain Control amplifies the weak signal when the signal is weaker than the normal signal processing degree and attenuates when signal power is more than the normal signal degree.

Fig. 1 Block diagram for AGC

Circuit design and simulation consequences

Hearing Aid Components

A general hearing assistance circuit consists of constituents as described in Fig. 2.

Fig. 2 Hearing assistance block diagram

Blocks with outlined are explained in item in following subdivisions

Microphone

By and large Electret capacitor mikes are used in hearing AIDSs, due to their simple design they are inexpensive. They besides provide a stable frequence response, tantamount noise degrees across frequences and sensitivenesss as per the demand of the hearing assistance application [ 6 ] .

Preamplifier

A preamplifier required for this circuit must be able to supply high addition ( ~1000 or 60dB ) .

Fig. 3, a, general Preamplifier circuit

Fig.3.b AC analysis demoing Gain of Preamplifier circuit

Though it is possible based that the theoretical addition of the circuit is non equal to the practical ascertained addition. As it depends on the undermentioned parametric quantities of the operational amplifier.

Unity Gain bandwidth

It ‘s the frequence bandwidth of the operational amplifier when the addition of an Opamp is unbroken 1. The merchandise of addition and bandwidth is ever changeless for an Opamp. So as we increase the addition from 1, the bandwidth starts cut downing. A 1MHz Unity addition bandwidth opamp is sufficient for the intent of a low cost hearing assistance design.

Large Signal Voltage addition

Large signal electromotive force addition limits the addition of an opamp for higher scopes. This value ranges from 100dB to 250dB for general opamps. A Cascaded circuit may be required for the higher electromotive force addition demand.

Fig. 4. Cascaded design for preamplifier

Antialiasing Filter Selection

Antialiasing filter design requires choice of best possible filter solution which can supply faster axial rotation off during Transition set and lowest possible ripplings during Passband and Stopband. Antialiasing a low base on balls filter.

Following are the possible designs of filters that are available.

Bessel filters

Bessel filters provides smoothest axial rotation off but increased passage set which can be an issue for hearing assistance design. There are really less ripplings in passband and stopband. It besides causes really less deformation to the stage of the signal.

Chebyshev filters

Chebyshev filters have faster roll away compared to the Bessel filters, hence passage set is smaller, which is favorable for hearing assistance design. There are some ripplings in passband and stopband but can be minimized as per the demand.

Elliptic filters

Elliptic filter provides fastest axial rotation off but causes a batch of ripplings in passband and halt set which makes the passage set smallest. There are ripplings in passband and stopband and there is besides an issue of big stage deformation.

Therefore in our design we used Lowpass, Multiple Feedback, Chebyshev with 1 dubniums passband ripplings. There are many Filter design tools available which can be used to happen the best possible design as per the demands. Following possible parametric quantities were considered to bring forth the design in Fig 5.

Table I

comparing of possible low base on balls multiple feedback filter design

Design Name

Order and Phases

Attenuation ( 4KHz to 5KHz ) ( in dubnium )

Attenuation ( 4KHz to 10Khz ) ( in dubnium )

Chebyshev 1dB

6 & A ; 3

-20

-60

Chebyshev 1dB

4 & A ; 2

-10dB

-44dB

Chebyshev 0.5dB

8 & A ; 4

-30dB

-95dB

Chebyshev 0.5dB

10 & A ; 5

-40dB

-120dB

Filter type: Lowpass, Multiple Feedback, Chebyshev 1 dubnium Order: 4

Phases: 2 ( so that merely 2 opamps are used reduces power ingestion )

Addition: 1 V/V ( 0 dubnium )

Allowable PassBand Ripple: 1 dubnium

Passband Frequency: 4 kilohertz

Corner Frequency Attenuation: 0 dubnium

Stopband Attenuation: -10 dubnium

Stopband Frequency: 5 kilohertz

Fig. 5. Antialiasing filter Lowpass, Multiple Feedback, Chebyshev with 1 dubniums passband ripplings

Generated utilizing FilterPro Desktop Version 3.1.0.23446 package by Texas Instruments [ 8 ]

Noise

Noise in the circuit can adversely impact the quality of sound produced by the device. Based on the beginnings noise following type of noise can be of found in circuit: Shot noise, Thermal noise, Flicker noise, Burst noise, Avalanche noise [ 9 ] . Thermal noise is the major beginning of noise in the circuit being larger than all other beginnings of noise, it dominates over others. Thermal noise due to a opposition can be calculated utilizing following equation.

E = a?s4kTBR ( 1 )

Where K is boltzman changeless

T is temperature in Kelvin

B is bandwidth

R is Resistance

TABLE II

Thermal noise values contributed by circuit constituents

Phases

Calculated Thermal Noise ( sing Ideal Opams ) ( in nV2 ) presuming 20KHz Bandwidth

Preamplifier

3.293517

Antialiasing filter Phase 1

0.011040

Antialiasing Phase 2

0.001017952

The noise calculated above is sing opamp to an ideal opamp. But opamp itself contributes towards the noise. Hence we should to choose our opamp carefully.

Opamp choice

Choosing Opamp for the circuit was based on following steps

Power ingestion by the opamp

Power ingestion by other circuit constituents

Cost of the opamp when ordered in majority

Power supply scope: 2.1 V to 5V

Noise

Handiness of Quad bundles

Table IIII

Selected Opamp comparision

Opamp

Operating Voltage

Cost ( in $ )

Current ingestion ( microampere /opamp )

Unity Gain Bandwidth

Noise ( nV/a?sHz )

OPA4379

1.8 to 5.5V

0.85

5.5

90KHz

80

OP4348

2.1 to 5.5 Volts

0.5

65

1MHz

35

LMV344

2.5 to 5.5 Volts

0.38

170

1MHz

20

LMV614

1.8 to 5.5V

0.30

210

1.4 MHz

60

In our design we used LMV344 in planing based as the power ingestion by the other constituents was still higher than LMV344

Decisions

Decision

Recognition

I would wish thank Prof. Subhajit Sen for steering me during this undertaking. I would besides wish to admit Ms. Megha Tak and Mr. Nityam Vakil, who were working on the other parts of hearing assistance design, for assisting me in understanding the whole hearing assistance design. I would besides wish to thank the Lab staff Mr. Krunal Patel and Mr. Naresh Patel for giving us entree to Lab 101 and Lab 211 for the class of this undertaking.