DSP application on one dimensional signal

Patent Review :
Pulse width modulation control unit of inverter.
Application No:89303871.1
Patent No:033878B1
Date of patent:18.04.89
Inventor: Ohyama,kazunobu c/o shiga plant diakin insdustries limited
Summary:
      Main objective this patent was to relate PWM control unit of an inverter with precise Carrier frequency. The control unit is capable of effective precise waveform control and therefore reduction in electromagnetic noise and improvement in motor efficiency.
     The control unit on PWM by high speed carrier is possible by use of high speed arithmetic operator,such as an exclusive software composed by digital circuit,DSP. There are some disadvantage of use of exclusive software ,that can be overcome by use of one-chip microcomputer which cheap and simple in composition of circuit.
Link: http://www.freepatentsonline.com/EP0338798A2.html



IEEE Paper Review :
Pulse Width Modulation of Three-Phase Switched Boost Inverter.
Author: Ravindranath Adda, Avinash Joshi, and Santanu Mishra
Publisher:IEEE
Published in: 2013
Date of Conference:2013 IEEE Energy Conversion  Congress And Exposition
Summary:
The paper was about the steady state analysis of a three-phase switched boost inverter (SBI). A Pulse Width Modulation (PWM) control strategy suitable for the gate control signal generation of the three-phase SBI is also described in this paper. The proposed PWM control technique has been implemented in digital domain using Texas Instruments TMS320F28335 digital signal processor (DSP). The steady state analysis and the PWM control technique are verified using an experimental prototype of three-phase SBI supplying a balanced three-phase load. The gate control signals for the experimental prototype of SBI are generated using the DSP. A comparison of the SBI with traditional three-phase voltage source inverter (VSI) for the same input and control parameters is also given in this paper. The experimental results presented in the paper show good correlation with the theoretical analysis. 
Link:http://ieeexplore.ieee.org/document/6646780/?denied

Basic operation on DSP Processor

     All previous labs performed by us were either mathematical calculation and its execution using c language on compiler or simulation using SCILAB. But to know what is real time, perform of our system and what is its response for given input, we need to test a real time system. For that we go for Emulation of the system. 
   All previous labs performed by us were either mathematical calculation and its execution using c language on compiler or simulation using SCILAB. But to know what is real time, perform of our system and what is its response for given input, we need to test a real time system. For that we go for Emulation of the system. 

FIR filter (Frequency sampling)

     In the previous lab we performed FIR filter designing using windowing method. This lab was about designing a filter having linear or nonlinear filter. Required input data given was cutoff frequency and order of the filter. Here also to find h (n) we used to have (w) from that we replaced w by (2πk) /N that is sampling of frequency response, where N is range of sampling.
     Therefore Hd (w) is converted into Hd (k) having symmetrical sampled values about N/2 sampling point. And then by taking IDFT of Hd (k) we found required h (n) i.e. characteristic of the filter.

Introduction

      In previous 4 labs of DSPP, we have seen different mathematical calculation on discrete time signal, such as convolution, correlation. That are used to find the output of the system. We did Fourier transform of discrete time signal and used OSM and OAM to give a start to filter designing.

     All the labs discussed above were performed by simple c code. But to study the behavior of a system we need to first simulate it using the appropriate simulator and then process it in the real world, to do that we go for emulation. All upcoming labs are scheduled for simulation of the program Emulation of some of the previously performed labs.

FIR filter (windowing method)

     As we know IIR filter requires infinite no. of samples.But it is difficult to store such many values.One of the solution for it is to go for truncation of these infinite values and making it finite i.e. nothing but FIR filter. For that we took same input as that of Butterworth filter. We used windowing method to do that. There are no. of windows that can be used for designing of FIR filters, selection criteria of a window is to have a maximum attenuation in stopband and transition width should be as small as possible. Procedure to do this is to first find desired h(n) i.e. hd(n) and multiply it with required window function.

   The reason for using a FIR filter was that we can't always have a stable IIR filter. But FIR filters are always stable. And reason for this is position of the poles, they lie on the origin in digital filter.Also linear phase FIR filter reduces distortion in the output. 

Chebyshev Filter

      Our previous lab was about Butterworth filter. There is another kind of filter designing technique for IIR filter. That is Chebyshev filter, the procedure for this was same as Butterworth filter. We can compare these two techniques, We can say that For same input order of the filter is reduced in the Chebyshev filter than Butterworth filter. 
       But ripple is present in the passband and absent in stopband for Chebyshev-I. And what about the position of the poles, It lies on the ellipse. in both the filter designs zero does not exist and it includes a number of calculations. 

Butterworth filter

          We saw two types of filters, those are IIR and FIR. The design of filter means to find H(z) in Digital filter design and H(s) in the case of  Analog signal. Butterworth filters are the type of IIR filter. We used SCILAB for simulation of this filter. Where inputs such as stop band(Ws) and pass band (Wp)frequencies with their respective attenuation(As, Ap) and sampling frequency(Ws) were taken from the user.
           The program was written so that to calculate the order of the filter, transfer function and plot the graph.From that graph, we came to know that, there is no ripple in the passband and stopband of Butterworth filter. Transfer function indicates poles position and order of the filter. It is a stable filter with poles lying on a unit circle. For high pass filter Ws>>Wp, Inverse of this for low pass filter.