5. The Technical Means of Diagnostics and Balancing.
Question 5.6.
Your diagnostic systems are used by many Russian railways. We
would like to know what problems is it possible to solve with its help, what
specific peculiarities in railway vehicles have you met and how do you take into
account these peculiarities?
(The question was asked by the specialists of Byelorussia car stock maintenance
teams.)
The question is answered by
Oleg Arkhangelsky:
Such features as the power consumption, processing rate and
reliability are very important when you design a portable measurement system.
The usage of a signal processor in such systems optimizes all the three features
without making worse the others (for example, such as the cost).
One of the main reason why we use specialized digital signal processors (DSP) in
such instruments is the necessity to make a real time signal processing. In
alternative case the signal processing is made in two stages. On the first one
we perform real-time acquisition of the signal and only after that we process it.
Such approach not only takes more time, but requires also an additional buffer
memory that is not needed when you use the DSP.
The signal processor can be used for signal spectral analysis, the analysis of
its form, and very often for preliminary signal filtering. This preliminary
filtering enables to decrease the number of anti-aliasing filters up to only one,
the characteristics of which is determined by the highest frequency of
analog-to-digit conversion, without changing it for all types of signal analysis.
Besides this, the order of this filter can be significantly decreased. It means
that its complexity and power consumption will be decreased too.
All successive filtering is performed by DSP using digital technique. It
significantly improves specification features of the instrument. The use of the
DSP enables you to have all the advantages of digital filters against analog
ones:
the high order of the filters;
practically any characteristics of the filters;
stable parameters, independent of time;
the possibility to design filters with varying parameters,
including adaptive filters;
simple filter commutation;
no elements that require tuning;
high noise immunity and power-supply rejection;
self-test can be built in;
fewer components.
Besides this the signal processor is used in our instruments
also to implement frequency band filters and successive enveloping of the
filtered signal.
The computational complexity of the digital signal processing algorithm assumes
intensive use of mathematical operations, and consequently, an instrument
without an internal DSP has to have sufficiently powerful central processor.
When you use a DSP the latter makes the major part of the computations. It
allows to use as a central processor a microcontroller that significantly
decreases the power consumption of the instrument and makes its circuitry much
simpler and, consequently, increases its work with its internal power batteries,
decreases its weight and increases its reliability.
And one more argument in favor of DSP. The use of digital filters provides an
incredible flexibility of the measurement system because the new filters,
detectors and other means of processing can be added to the system not only
without any intervention in the hardware, but by simple reloading of the
firmware.
