This paper deals with the continuous development of digital signal processing in the field of test and measurement. . Continuous process of development of filters to meet he challenges of bandwidth and accuracy. The real time oscilloscope has been the mainstay of electronic and R &D applications.
DSP is a well established discipline. It has become the enabling toll to extend the oscilloscope bandwidth beyond the current analogue limits and to improve the overall measurement accuracy.
This above said extension in bandwidth of oscilloscope is made to reality with the help of filtering technique. In addition, the various developments in digital filters in the field of digital signal processing is discussed.
Digital signal processing is a processing of signals on a digital computer, the operations performed on a signal consists of a number of mathematical operations as specified by a software program. In a broader sense the digital system can be implemented as a combination of digital hard ware and software, each of which performs its own set of specified operations.
This rapid development is a result of the significant advances in digital computer technology & integrated circuit fabrication. The rapid developments in integrated circuit technology such as VLSI of electronic circuits has spurred the development of powerfull smaller & cheaper digital computers & special purpose digital hardware. These inexpensive & relatively fast digital circuits have made it possible to construct highly sophisticated digital systems capable of performing complex DSP functions & tasks, which are usually too difficult and/or too expensive to be performed by analog signal processing systems
As most of the signals are analog, so these are to be converted into digital signals for carrying out processing in digital form. Processing of a input signal is nothing but performing specified operations on it according to the requirement . After processing , signals can be reconverted to analog form if desired.
Filter is the most vital system in the DSP technology. Digital filters are classified by their use & implementation. These are called time domain or frequency domain based on their use, and finite impulse response(FIR) and infinite impulse response(IIR). Now a days WAVELET TRASFORMS have played tremendous role in DSP technology.
Most of the signals encountered in science & engineering are analog in nature. That is , the signals are functions of a continuous variable, such as time or space, and usually take on values in a continuous range. Such signals may be processed directly by appropriate analog systems(such as filters or frequency analyzers) or frequency multipliers for the purpose of changing their characteristics or extracting some desired information. In such a case we say that the signal has been processed directly in its analog form. Here both the input signal & the o/p signal are in analog form.
ANALOG I/P ANALOG O/P
To perform the processing digitally, there is a need for an interface between analog signal and the digital processor. This interface is called A/D converter. The o/p of the A/D converter is a digital signal that is appropriate as an input to the digital processor.
ANALOG I/P SIGNAL ANALOG
SIGNAL O/P SIGNAL
CLASSIFICATION OF DIGITAL FILTERS:
Digital filters are classified by their use and implementation. These are called time domain & frequency domain based on their use, and FIR & IIR by their implementation.
Classification of digital filters by their use & by implementation:
VARIOUS DEVELOPMENTS IN DIGITAL FILTER IN THE FIELD OF DSP
Filter is the most vital system in the digital signal processing technology. A thorough treatment of multirate digital filters & filter banks including quadrature mirror filters was given by vaidya Nathan in 1990. In the case of IIR filters a new algorithm for the design by approximating specified magnitude & phase responses in frequency domain has been proposed by Mathias C lang in 2000. The proposed algorithm minimizes the mean square approximation error subjected to a constraint on pole radii. Consequently stability of such systems can be guaranteed for such filters. Some times it is difficult to estimate frequencies in a desired signal of multiple sinusoids buried in additive noise . For processing such signals, spectral estimation techniques based on DFT are used. Such processing is termed as off line processing. But such methods are costly. Online processing of such signals is carried out by using adaptive notch filtering technique. For direct frequency estimation , any new adaptive algorithm is developed for constrained pole zero notch filter by gang li in the year 1997.
Adaptive algorithm is constructed for solving the stochastic envelope constrained filtering problem. Some times input signals get corrupted by an additive random noise. Therefore, envelope constraint filters are deigned to minimize the noise enhancement while the o/p of noiseless filter lies within an o/p pulse shape envelope. This formulation has advantage over least mean square algorithm which is the conventional approach of filter design.
Echo casncellation is the specific field where the application of this filter is of immense importance. Transmission of message over band limited or dispersive channel leads to the distortion of massage in communication. Kalman filtering based on channel equilization is generally used in this context. Wiener filters are generally used for a stationary random process. An analytical technique to design zero phase FIR digital pass band filter with the evaluation of errors and having high accuracy by using a fast non iterative algorithm,is recently suggested by k nowzynski on the year 2000.This is the further improvement over the optimization of FIR filters by using Remezâ„¢s algorithm.
Digital filters are widely used to get a better performance. In amteur radio, these achieve high â€œfidelity music reproduction. Aaptive filters remove fading in tele communication by altering the sampling rate.
Interpolation filters are used to increase the sampling rate, while decimation filters are used to decrease the sampling rate.
DSP TO BOOST SCOPE PERFORMANCE:
The real time oscilloscope has been the mainstay of electronic design and R&D applications for decades. oscilloscope performance has always risen to the challenges of bandwidth and accuracy.
Ideally, a measuring instrumentâ„¢s bandwidth should exceed that of the target device being observed. Yet, the basic metric of an oscilloscopeâ„¢s performanceâ€the analog band width- is bound by the same technologies as that of say , a digital network switching element. Both platforms use the fastest available semiconductor devices. Both relay on custom ICs. Given these realities, how can the oscilloscope performance make leap to the next level? How can it support next-generation technical advances?
The answer lies in the digital signal processing. It turns out that the raw analogue band width can be extended and ,in fact enhanced, using DSP. It has become the enabling tool to extend the oscilloscope bandwidth beyond the current analogue limits and to improve the overall measurement accuracy. The top tier of todayâ„¢s oscilloscopes includes a host of models offering multigigahertz bandwidth. In the simplest terms, DSP creates a filtering function that counteracts roll off at the top end of the specified frequency range. Fig 1 shows a pair of frequency response curves for a digital storage oscilloscope (DSO) with 4GHz true analogue bandwidth. The dashed line defines a text book â€œperfect band width envelope, while the other line approximates a real world oscilloscopes frequency response curve . Wherever this line departs from the ideal envelope, the deviation becomes part of the measurement. To obtain the best possible signal fidelity, it is essential to keep this deviations to a minimum. Now consider the Fig.2 . This is DSP extended frequency response curve for the same oscilloscope, now offered as a 5GHz oscilloscope. The 3db point is indeed at 5GHz.
What happens beyond the 4GHz boundary depends largely on the quality of the DSP implementation.
DSP frequency extensions are a form of filtering.. It a very astute filter design to create a usable bandwidth extension while minimizing magnitude aberrations at the extremes and elsewhere in the range, as well as controlling the phase shift and distortion. FIR and DSP :
Todayâ„¢s leading high bandwidth oscilloscopes employ a finite impulse response (FIR) filter scheme.Unlike the IIR filters FIR filters are guaranteed to be stable and can deliver perfectly linear phase response. Moreover it is the most suitable for applying equilization of phase and magnitude as needed over almost the entire bandwidth of an oscilloscope channel. The FIR filter is tuned for its optimum step response. Its exact transfer function is proprietary to each manufacturer, but is often based on a Gaussian algorithm.
Using FIR filter approach requires a rigorous calibration process during manufacturing. Each oscilloscope channel and attenuator setting that will receive filtering must be calibrated. FIR filter coefficients based on the measured response of the oscilloscope channel associate with each individual supported attenuator setting and channel. These coefficients are mathematically convolved with the acquisition data when the oscilloscope is running. The result , known as channel matching, is a very closely matched phase and magnitude response across all channels.
ADDITIONAL ADVANTAGES OF USING DSP:
Frequency extension is just the beginning of a well designed DSP filter can do for an oscilloscope. Other benefits of DSP are
i) It can enable accurate comparison of signals across multiple channels. Because each channel is specifically calibrated with its own permanent filter coefficients at the factory, there is a close match in phase and magnitude response between channels.
ii) It can improve rise time sensitivity of the oscilloscope as well as the accuracy of the rise time measurements
iii) Because exceptional magnitude & phase linearity ,DSP filtering can support more accurate frequency domain measurements when using the oscilloscopeâ„¢s spectral acquisition features
DSP filtering can deliver sharper eye diagrams. It removes noise, jitter and aliasing and reduces the amount of overshoot.
OTHER SIDE OF DSP IMPLEMENTATION:
We discussed what a well DSP filter can do for an oscilloscope. It can be shown that magnitude consistency suffers if the DSP implementation does not take variables such as sample rate into sufficient consideratation.
DSP can effect the oscilloscopes equivalent time modes, attenuating and distorting eye diagrams when aquired with high ET sample rates. Another potential side effect is an inconsistent amplitude response that varies with the trigger source selection. A sophisticated DSP implementation can avoid all these short comings.
The industryâ„¢s ever escalating bandwidth needs will mandate continuing evaluation in oscilloscopes. It is safe to assume that increase in analogue band width and DSP enhanced performance will continue to go hand in hand. One cannot advace with out the other. The instruments innate analogue band width is the plat form upon which the DSP frequency extensions must stand. This base bandwidth depends on good analogue engineering in the areas of probing, vertical input amplification and analogue to digital conversion.
To conclude, Digital Signal Processing has brought us to a new age in the technical field. DSP boosts the scopes performance, it will grow as the underlying technologies permit oscilloscope users will become more familiar with what DSP can and cannot do for them, and will demand more than just band width oscilloscope innovators will focus on the across-the-board performance benefits that DSP can deliver and will continue to push the bandwidth boundaries to match user needs. Also it contribution specially in the field of communication systems and aerospace has brought technological revolution .
SCOPE FOR THE DEVELOPMENT: The areas like filter banks , wavelet transforms, adaptive filtering , discrete chirp fourier transform have an anarmous scope of research due to the relevance in number of applications like spectral analysis, control and system identification, channel equilisation, echo cancellation, reconstruction of signal and communications etc.,. To develop new algorithms, to improve the permonce of systems and to reduce the complexity of DSP based sysyems are the various critical issues in digital signal processing.