This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
This document proposes a possible analog circuit design for a low cost programmable hearing aid solution. Along with the design methodology this document throws light on the practical issues such as aliasing, noise, filter selection for design, amplifier selection etc., that are vital considerations for analog circuit of hearing aid design. This design is suitable (but not limited) for a programmable hearing aid design requiring digital signal processing.
Keywords - Antialiasing filter, Noise, Hearing Aid, filters, programmable hearing aid design, Low cost, Deafness, hearing impairment, Automated Gain control.
Our body is made up of different natural sensors which help us in observing our environment. Our measure of surrounding is based on the capabilities of these sensors. Hearing is one such characteristic measure of our surrounding. Ear, which is part of auditory system of the body, detects change in pressure of the medium or vibrations to perceive sound. Although the Auditory system looks same for everyone but the hearing capacity is not same for all. We perceive sounds differently based on our hearing capabilities. A spectrographic analysis of the hearing capacity of a person gives information about hearing profile of that person. Partial or complete incapability of a person to perceive sound without any external aid is hearing loss or hearing impairment. A study done by World Health Organization titled "Global burden of hearing loss in the year 2000" states that more than 250 million people in the world suffer from Hearing impairment and it is spread across the ages. The study  also states that only 10% of the hearing impaired people use hearing aid where in developed countries it is used by around 49% of hearing impaired population but in developing countries less than 1% hearing impaired population is using hearing aid. There is huge need of low cost hearing aid device that can be used in developing countries.
Hearing aid solution are of two types: Analog and Digital hearing aids. Analog hearing aids cheaper than digital hearing aids but are less programmable. Also the noise level as compared with digital hearing aids is high. Analog hearing aids provide very less flexibility for adjustments with the hearing profile of the person. Whereas, Digital hearing aids sometimes known as "true hearing aids" provides a great flexibility for adjustments as per the hearing profile. Digital hearing aids use noise reduction algorithms to reduce the noise. Digital hearing aids use digital signal processing algorithms but still require an analog circuit to capture sound from microphone. In the design, we are trying to process signals in both analog and digital domain in order to get a suitable trade-off between cost, power consumption and quality. The design utilizes the capacity of digital noise reduction algorithms and programmability along with the reduction of burden of digital processing by using analog circuit helping in selecting cheaper digital processor. This paper focuses on the analog circuit required for the design. Rest of the design are covered in  and .
Design Constraints and Methodology
Following are the major constraints that were considered for the design of low cost hearing aid design
Degree of hearing loss
Degree of hearing loss defines the level of hearing loss suffered by an individual. Degree of hearing loss also defines the gain requirement and transfer function of digital Hearing loss compensation filter. Based on degree, hearing loss are of following types
Mild hearing loss (25 and 40 dB SPL),
Moderate hearing loss (40 to 70 dB SPL),
Severe hearing loss (70 to 95 dB SPL).
A general hearing aid should be able to cater the need of all four types of hearing loss but in order to reduce the cost and power the design can be modified and limited to provide a hearing aid till moderate hearing loss. A design should be able to provide 95 dB SPL to cater hearing loss of all degrees.
Hearing loss profile:
As we all perceive sound differently, we have different spectrographic response to sound. For a normal person this doesn't vary much (0 - 20dB) for almost all frequencies. For Hearing impaired people the response is nonlinear and dependent on frequency, which means for some of the hearing loss profiles, low frequency range is not audible and for some high frequency range is not audible. In order to compensate for the loss, the filter should provide enough adjustment capability that can cover all possible hearing loss profiles.
Hearing aids available in market are featured to take care of need of different types of hearing profiles. These features add cost to the device. There are device in which one can set modes like speech, music, noise etc. to adjust the device function as per the need. These features are available only in digital hearing aids and require separate memory for each hearing mode, which also makes the device costly. Also for processing needs these device require digital signal processors for the processing, which are again costly.
A Low cost hearing aid cannot provide the above said features but it should cater the basic need of the ability to hear speech signal. Speech signals range from 125 Hz to 4 KHz.
Digital hearing aid design requires antialiasing filter to satisfy sampling theorem. As microphone can capture a wide spectrum of frequency range, which can cause aliasing because of the signals captured above Nyquist frequency. The purpose of Antialiasing filter is to remove the signals above Nyquist frequency. A filter cannot remove the signal completely but can attenuate the signal. An attenuation of the order of 60 dB is quite safe for hearing aid design. As our frequency of interest lies under 4 KHz, so our sampling frequency should be above 8 KHz. But keeping 8 KHz as sampling frequency will cause higher order of antialiasing filter as a sharp 60 dB attenuation after 4 KHz. Which will require a very high order filter (more than 10), which is practically not advisable. A good design can be where Analog Antialiasing filter can be relaxed by increasing sampling frequency. In our design we identified 800 KHz as a suitable sampling frequency which can relax analog filter up to a level where power consumption by analog part could be reduced to minimum and processing power of digital filter could have maximum budget of clock cycles for a digital antialiasing filter after other digital signal processing .
A preamplifier circuit is required to amplify the signals captured by microphone to the level it can efficiently utilize the voltage reference of Analog to digital convertor. If the voltage reference of the Analog to digital convertor is 5V, then signal should be amplified till it reaches 5V peak to peak. Electret condenser microphones produces a voltage of range 5mV to 10mV. A preamplifier amplifies this signal to the voltage level that can utilize all the bits of Analog to Digital Convertor when given input in order to use ADC efficiently, i.e. of the order of reference voltage. Generally, the gain required is of the order of one thousand. Though the overall gain of preamplifier modified by the automated gain control.
Heterogeneous sound in the surrounding:
Signal source for a hearing aid device is highly heterogeneous, power level of sound in surrounding can change at any point of time. Hearing aid should be smart enough to vary the gain sensing the power level of the surrounding in order to protect any harm to the ear. Automatic Gain Control is used to compress the loud transitions of the sound and amplify the weak sound signals to that of range of signal processing range. Automated Gain Control amplifies the weak signal when the signal is weaker than the normal signal processing level and attenuates when signal power is more than the normal signal level.
Fig. 1 Block diagram for AGC
Circuit design and simulation results
Hearing Aid Components
A general hearing aid circuit consists of components as described in Fig. 2.
Fig. 2 Hearing aid block diagram
Blocks with outlined are explained in detail in following sections
Generally Electret condenser microphones are used in hearing aids, due to their simple design they are cheap. They also provide a stable frequency response, equivalent noise levels across frequencies and sensitivities as per the requirement of the hearing aid application .
A preamplifier required for this circuit must be able to provide high gain (~1000 or 60dB).
Fig. 3, a, general Preamplifier circuit
Fig.3.b AC analysis showing Gain of Preamplifier circuit
Though it is possible based that the theoretical gain of the circuit is not equal to the practical observed gain. As it depends on the following parameters of the operational amplifier.
Unity Gain bandwidth
It's the frequency bandwidth of the operational amplifier when the gain of an Opamp is kept 1. The product of gain and bandwidth is always constant for an Opamp. So as we increase the gain from 1, the bandwidth starts reducing. A 1MHz Unity gain bandwidth opamp is sufficient for the purpose of a low cost hearing aid design.
Large Signal Voltage gain
Large signal voltage gain limits the gain of an opamp for higher ranges. This value ranges from 100dB to 250dB for general opamps. A Cascaded circuit may be required for the higher voltage gain requirement.
Fig. 4. Cascaded design for preamplifier
Antialiasing Filter Selection
Antialiasing filter design requires selection of best possible filter solution which can provide faster roll off during Transition band and lowest possible ripples during Passband and Stopband. Antialiasing a low pass filter.
Following are the possible designs of filters that are available.
Bessel filters provides smoothest roll off but increased transition band which can be an issue for hearing aid design. There are very less ripples in passband and stopband. It also causes very less distortion to the phase of the signal.
Chebyshev filters have faster roll off compared to the Bessel filters, hence transition band is smaller, which is favourable for hearing aid design. There are some ripples in passband and stopband but can be minimized as per the requirement.
Elliptic filter provides fastest roll off but causes a lot of ripples in passband and stop band which makes the transition band smallest. There are ripples in passband and stopband and there is also an issue of large phase distortion.
Thus in our design we used Lowpass, Multiple Feedback, Chebyshev with 1 dB passband ripples. There are many Filter design tools available which can be used to find the best possible design as per the requirements. Following possible parameters were considered to generate the design in Fig 5.
comparison of possible low pass multiple feedback filter design
Order and Stages
Attenuation (4KHz to 5KHz) (in dB )
Attenuation (4KHz to 10Khz)(in dB)
6 & 3
4 & 2
8 & 4
10 & 5
Filter type: Lowpass, Multiple Feedback, Chebyshev 1 dB Order: 4
Stages: 2 (so that only 2 opamps are used reduces power consumption)
Gain: 1 V/V (0 dB)
Allowable PassBand Ripple: 1 dB
Passband Frequency: 4 kHz
Corner Frequency Attenuation: 0 dB
Stopband Attenuation: -10 dB
Stopband Frequency: 5 kHz
Fig. 5. Antialiasing filter Lowpass, Multiple Feedback, Chebyshev with 1 dB passband ripples
Generated using FilterPro Desktop Version 126.96.36.19946 software by Texas Instruments 
Noise in the circuit can adversely affect the quality of sound produced by the device. Based on the sources noise following type of noise can be of found in circuit: Shot noise, Thermal noise, Flicker noise, Burst noise, Avalanche noise . Thermal noise is the major source of noise in the circuit being larger than all other sources of noise, it dominates over others. Thermal noise due to a resistance can be calculated using following equation.
E = âˆš4kTBR (1)
Where k is boltzman constant
T is temperature in Kelvin
B is bandwidth
R is Resistance
Thermal noise values contributed by circuit components
Calculated Thermal Noise (considering Ideal Opams) (in nV2) assuming 20KHz Bandwidth
Antialiasing filter Stage 1
Antialiasing Stage 2
The noise calculated above is considering opamp to an ideal opamp. But opamp itself contributes towards the noise. Hence we should to select our opamp carefully.
Selecting Opamp for the circuit was based on following measures
Power consumption by the opamp
Power consumption by other circuit components
Cost of the opamp when ordered in bulk
Power supply range: 2.1 V to 5V
Availability of Quad packages
Selected Opamp comparision
Cost (in $)
Current consumption (microampere /opamp)
Unity Gain Bandwidth
Noise (nV/âˆšHz )
1.8 to 5.5V
2.1 to 5.5 V
2.5 to 5.5 V
1.8 to 5.5V
In our design we used LMV344 in designing based as the power consumption by the other components was still higher than LMV344
I would like thank Prof. Subhajit Sen for guiding me during this project. I would also like to acknowledge Ms. Megha Tak and Mr. Nityam Vakil, who were working on the other parts of hearing aid design, for helping me in understanding the whole hearing aid design. I would also like to thank the Lab staff Mr. Krunal Patel and Mr. Naresh Patel for giving us access to Lab 101 and Lab 211 for the course of this project.