National Congress on Communications and Computer Aided Electronic Systems (CCAES 2012)

Performance Analysis of FIR Filter algorithm and architecture to design a Programmable PreModulation Filter for Avionic Application
Mohd. Sayeeduddin Habeeb1, U. Naresh Kumar2
1

Asst. Prof. Dept. of ECE, Hi-Point College of Engineering and Technology, Hyderabad, Email: [email protected] 2 Scientist D, DOFI, RCI, Hyderabad, Email: [email protected] aerospace vehicle, hardware modifications are sickly prohibited for the batter reliability. It is necessary to change the PCM bit rate and cutoff frequency of the pre modulation filter with the software programming in order to improve the bit rate in the allowable bandwidth even if the units are already manufactured. In the addition, under the same RF power condition, the reduction of the transmitting data rate will result in an increase of link margin. For this, the pre modulation filter should have the variable cutoff frequency property with software programmable function according to 1.4 times the bit rate. This paper we present the configuration of pre modulation having variable cutoff frequency characteristics. Programmable pre modulation filter consist of three sub blocks DAC (Digital to Analog Converter), FIR (Finite Impulse Response) filter and 2nd order LPF(Low pass filter) to meet the magnitude frequency response of on an analog pre modulation filter (Bessel filter). By analyzing the frequency response of three sub blocks, attenuation requirement of this sub blocks can be calculated and linear phase features are also taken into account for each stage. Each block is simulated and tested and finally implemented on PCB board. 2. PRE MODULATION FILTER DESIGN Programmable pre modulation filter consist of digital filter, DAC and 2nd order low pass filter instate of analog Bessel filter and Bessel filter is used as an analog pre modulation filter. As shown in fig 1 programmable pre modulation filter consist Digital filter, DAC and low pass filter. FIR filter is heart of programmable pre modulation filter, thus FIR filter attenuation frequency response is compared with the attenuation frequency response of Bessel filter. DAC and LPF is used interpolator for desired output sequence from digital filter. Total sum of attenuation of each block of programmable filer must be equal to attenuation frequency response of Bessel filter, this filter will have linear phase characteristics like Bessel filter so thats why FIR filter is used [2].Here in this paper sampling frequency is set at eight times of PCM bit rate, with this we dont have aliasing problem, this DDS

Abstract: The purpose of this study is to design and analyze a programmable Pre-Modulation filter for avionic application. Modulated carriers have more energy in side bands; this side band energy can be reduced by the effect of premodulation filter. But in present technology pre-modulation filter are based upon constant cutoff frequency whereas programmable pre-modulation filter is based upon variable cut off frequency property. So with this programmable filter any hardware modification can be avoided because an onboard unit of launched vehicles has gone through many environmental and functional tests for reliability point of view. Programmable pre modulation filter mainly consist of three sub blocks Digital filter, DAC and 2nd order low pass filter. Digital filter used is FIR filer because it has linear phase as that of Bessel filter, FIR filter different design technique is compared and implemented for our application

Key words FIR filter, PCM/FM, Bessel Filter

1. INTRODUCTION It is very important during the development of flight to monitor the performance of different subsystem of an aerospace vehicle. Monitoring this detailed performance of this sub-system during the flight is of great importance and this requirement is meant by telemetry system. Telemetry is the process of monitoring the real time performance of various subsystem of an aerospace vehicle that occurs in the field of view. The modulation technique used for this system is Bi L PCM (Pulse Code Modulation) style based upon IRIG (Inter Range Telemetry Group) telemetry slanderers which is commonly used in communication system [1]. PreModulation filter here used as pulse shaping filter with the low pass specification for given PCM bit stem. According to IRIG telemetry standards, -3dBcutoff frequency of filter is 1.4 times the PCM bit stem, with we can say that filter bandwidth depends upon PCM bit rate. PCM bit rate is set at initial development stage of aerospace vehicle, various functions and environment test are performed on a designed and manufactured onboard hardware units. For the successfully tested hardware units at the final development stage of

National Congress on Communications and Computer Aided Electronic Systems (CCAES 2012)
(Direct Digital Synthesis) can provide this frequency for our FPGA and DAC. filter have the bilateral symmetrical representation in frequency response at every 2/T interval.

Fig. 2 Pre Modulation filter input and output

Fig. 1 Functional block diagram of a RF interface modal including the variable cutoff frequency premeditation filter TABLE 1 MAJOR SPECIFICATION FOR THE VARIABLE

3. FIR FILTER The fundamental elements for DSP is an FIR Filter, it is type of digital filter with only zeros on Z-plane. FIR filters has finite value because it dont have feedback. The output differential equation in terms of input is given by

y[n] b0 x[n] b1x[n 1] ..............bN x[n N ]

PROGRAMMABLE PRE MODULATION FILTER
Items Modulation scheme PCM presentation style. Range of tunable bit rate (fb) Attenuation frequency response Phase frequency response -3dB cutoff frequency (fc) FIR/DAC sampling frequency (fs) Interpolation LPF specification Specification PCM/FM Bi L 100kbps 10Mbps 6th order Bessel LPF Linear phase 1.4 times fb 8 times fb Tunable 2nd order LPF

x[n] is the input signal, y[n] is the output signal, bi are the filter coefficients, and N is the filter order an Nth-order filter has(N + 1) terms on the right-hand side; these are commonly referred to as taps. Convolution Representation of above equation is given by

y[n] bi x[n i]
k 0

N 1

3.1 TYPES OF LINEAR PHASE FIR FILTER Linear phase FIR filter can be classified in four types. h(n) is non-zero for 0 n N 1 , Filters classified as Type I or II present the positive symmetry, then the symmetry of the impulse response can be represented by

The pre modulation filters smoothest and extends rectangular pulse as shown in Fig 2 Table 1 shows the major specification of programmable pre modulation filter. In summery we can design an analog Bessel filter with software programming by using cascade structure of different sub blocks like FIR filter, DAC and 2nd order LPF in PCM/FM transmission under the IRIG standard. Summation of amount of this sub blocks at the cutoff frequency shall be 3dB. The cutoff frequency will be 0.35/T at 1.4 times of PCM bit rate. Figure 2 shows the input and output of pre modulation filter. We setstop band frequency at 0.9/T because digital

h(n) h( N 1 n)
Filters classified as Type III or IV present the negative symmetry, anti-symmetry can be represented by

h(n) (h( N 1 n))

3.2 FIR FILTER STRUCTURES FIR filter are most commonly implemented in the directform and transpose-form. Recursive implementation needs less computation steps for certain special filters. Occasionally lattice and cascade structures are used. Direct foam FIR filter is most common type of realization method and it is most straightforward

National Congress on Communications and Computer Aided Electronic Systems (CCAES 2012)
structure, [3] here there is no close loop, so it is called non-recursive structure. So it is always possible to implemented FIR filter non-recursively because it can be implemented using the direct-form non-recursive structure, it is also possible to implement FIR filter recursively for some special case of filter coefficients Transposed form is alternative to direct form implementation for FIR filters. Transposed form is selfpipelined, and it takes more area than direct. Transposeform FIR filter, no. delays elements, coefficients, adders and output value. [4]Extra pipeline registers are required in direct form to reduce delays of adder ant to achieve high throughput. Without adding any extra pipeline registers we can achieve high throughput in FIR filter with transposed form. Direct form Symmetric FIR filter structures we can implement symmetry condition to reduce numbers of multiplication by half i.e. reduce the number of multipliers from N+1 to N/2+1. [3] 4. DESIGN OF FIR FILTER Design technique of FIR filter is used to estimating coefficients of the FIR filter. Infinite length impulse response cannot be realized so we are going for design of FIR filter. Starting with the study of the windowing technique. Windowing is simple and convenient, but it is not optimal type of designing. Causal and linear phase FIR filter can be obtained by truncating of Infinite impulse sequence by appropriate window function. Let w(n) be window function and window is twice of the rectangular window. First side loab for Hanning window is one tenth of rectangular window, because of this there is less ripples in stop band and passband region [5]. Hamming window is similar to that of Hanning window except that it has small amount of discontinuities at the boundaries. Compared to Hanning window, stop band attenuation is low at higher frequency. Mostly Hamming window is preferred because it has fewer oscillations in side loabs compares to that of Hanning window. Blackman Window is similar to hamming and hanning window but it has an additional cosine term for the reduction of ripples in pass band and stop band, and improves the width of main loab. We can select suitable window type and widow size based up on given transition width and minimum stop band attenuation of the desired filter information as describe in below Table 2. Fixable family of window is defined by Kaiser. Filter which is design by Bartlett window will reduces the peak amplitude (overshoot) of main loab but spreads transition region considerably. For getting smooth and better truncations of ideal frequency responses, the function like Hamming, Hanning and Blackman are more complicated. Kaiser provides better window results because parameter can be used to compromise between transition region width spreading reduction of peak amplitude (overshoot) and transition region width spreading. Frequency Sampling Design Technique is simplest and most direct technique if the desired frequency response is specified. In this technique [6, 7, 8] desired frequency response can be obtain by sampling, frequency response which is provided by the previous method. Sampling is done at the particular set of equally spaced frequency to obtain N number of samples. For designing liner phase FIR filter till now we used windowing and frequency sampling method, both are simple designing method but they have some disadvantages. Firstly we cannot specify the stop band and pass band frequencies secondly Window design gives the tradeoff between ripples in stop band and stop band and in the frequency sampling design technique we can optimize only ripples in stop band [3]. A set of condition can be derived for the linear phase FIR filter provided its filter response should be optimal in the sense of minimizing maximum approximation error to minimum, sometimes called as Chebyshev error, filter with this property is known as Equiripple filters. Here in the pass band and stop band approximation errors are uniformly distributes which results to filter of lower order [3].

hd (n) be

the ideal impulse response of the filter. The impulse response of the designed filter is given by

h(n) hd (n)w(n)
There are several window functions each window function is have different characteristics in time and frequency domain. Direct truncation or non-uniform convergence results to Gibbs phenomenon. It results to the overshoot and ripples in the spectrum, by multiplying infinite impulse response with finite window w(n) this can be reduced or eliminated Gibbs phenomenon. Rectangular window is simplest window design technique, but at the stop band this window provides worst performance. Sudden transition of rectangular window from 0 to 1 or 1 to 0 which result to Gibbs phenomenon, so triangular window was suggested by Bartlett call as Bartlett window. Triangular function is produces a smooth magnitude repose in pass band and stop band. A disadvantage with this method is transition region is more and in stop band attenuation is less [5]. Because of this problem triangular window is not used. In Hanning window width transition region of filter is double because the width of main loab in Hanning

National Congress on Communications and Computer Aided Electronic Systems (CCAES 2012)
TABLE 2 SUMMARY OF COMMONLY USED WINDOW FUNCTION CHARACTERISTICS

4.1 Parks McClellan algorithm Parks-McClellan algorithm is efficient and more flexible than other design techniques. The Parks-McClellan algorithm was published by James McClellan and Thomas Parks. The goal of the algorithm is to minimize the error in the pass and stop bands by utilizing the Chebyshev approximation. The Parks-McClellan algorithm is a variation of the Remez algorithm or Remez exchange algorithm, with the change that it is specifically designed for FIR filters and has become a standard method for FIR filter design. This method is an iterative procedure to find the optimal Chebyshev finite impulse response FIR filter. Alternation theorem was developed by Parks and McClellan to solve this problem. The alternations mentioned means that the optimum filter is equiripple. The last equation shows that the sign changes M+1 times, resulting in an oscillation or ripple on the band of interest. Thistheorem specifies that M+2 alternations are present; depending upon filter there may more or one. We can get M+3 alternations if some combinations of p and s. Here one ripple is extra, filter having this extra ripple called as Extra-ripple filter FREQUENCY RESPONSE OF FIR FILTERS. Frequency response of FIR filter using different design technique is given as bellow. 5.1 USING HAMMING WINDOW Following is the hamming window of order 6 and length of filter is 7. The MATLAB function used for plotting this window is W=hamming (M) 5.

Fig. 3 Frequency Response of Hamming Window

5.2 USING KAISAR WINDOW Matlab function used for plotting kisar window is W=kaisar(M)

Fig. 4 Frequency Response of Kaisar Window

5.3 FREQUENCY SAMPLING

Fig. 5 Frequency Response of Frequency Sampling

National Congress on Communications and Computer Aided Electronic Systems (CCAES 2012)
5.4 USING PARKS-MCCLELLAN ALGORITHM (5) P. Ramesh Babu, Digital signal processing,3rd edition, Scitech Publication(India) Pvt. Ltd.,2007 (6) T.W. Parks and C.S. Burrus, Digital Filter Design. New York:Wiley,1987 (7) L.R. Rabiner and B. Gold, Theory and Applications of Digital Signal Processing. (8) New Jersey: Prentice-Hall, 1975 J.G. Proakis and D.G. Manolakis, Digital Signal Processing-Principles, Algorithms and Applications New Delhi: Prentice-Hall, 2000

Fig. 6 Frequency Response of Parks-McClellan Algorithm

6.

CONCLUSION Various techniques involved in the design of FIR filters have been discussed in this paper. Every design methods have its own advantages and disadvantages and are selected depending on the type of filter to be designed. Prototype filters like low-pass, high-pass, band-pass etc can be design by using windowing techniques. For the any given frequency response this windowing design technique is not suitable. Other way of designing is by using frequency sampling technique for given magnitude response. In frequency sampling method errors occurs in frequency response at the point where it is not sampled So this can be overcome by using Parks-McClellan Filter Design Algorithm errors in pass and stop band can be minimizing. So in our application this algorithm is preferred for designing Digital filter part in Pre-Modulation filter.

BIBLIOGRAPHY (1) D. Taggart, PCM/FM performance enhancement using Reed Solomon Channel coding, IEEE Aerospace conference, pp13371346, 2004. (2) A. V. Oppenheim and R. W. Schafer, Digital signal processing, 2nd edition Prentice Hall, Inc., 1999. (3) Vinay K. Ingle and John G. Proakis, Digital signal Processing using MATLAB, International Student Edition, Vikas Publishing House, 2003 (4) John G. Proakis and Dimitris G. Manolakis, Signal Processing principles. Algorithms, and Application, 3rd edition, Prentice-Hall of India privet limited, New Delhi, India, 2006.