New Transducer Design Considerations
In the past few decades, factors deciding which transducer to use were simply based more on basic primitive transducer characteristics. With test parameters ever peaking up; speeds are getting faster, weights are getting lighter and frequencies of measured phenomena are getting incredibly high. Accordingly, criteria for transducer selection have to also be reconsidered.
Transducer non-linearity, hysteresis, or output used in the past to decide which sensor is more suitable for the measurement, are gradually giving way to other parameters such as frequency response, critical speeds or number of readings per second a transducer can output.
One such important characteristic is the number of readings per second a transducer can output. This should not be confused with the number of samples per second the transducer can collect. Both are related through the “Sampling Theory” by the “Nyquist Frequency” and are dependent on the design of the sensor and any on board intelligence.
Signal sampling representation. The continuous signal is represented
with a green color whereas the discrete samples are in blue.
Is the reduction of a continuous signal to a discrete signal.
An ideal (theoretical) sampler is one that produces samples equivalent to the instantaneous value of the continuous signal at the desired points.
Nyquist Sampling Rate
Band limited signals can be perfectly reconstructed from their samples if the sampling rate is more than twice the maximum frequency.
Nyquist Frequency fN;
Is the highest signal frequency that can be reconstructed from a certain sampling rate, usually no higher than half the sampling rate.
If fN is the Nyquist frequency of the yellow function, Red sample points
could fall either on the yellow function or any other function whose f > fN
or a function f with frequency f > fN, f cannot be distinguished from other functions with frequencies
n fN + f and n fN – f
Resolution of an A/D Converter (ADC)
The smallest voltage difference that can be detected. Also refered to as the least significant bit (LSB) in the conversion.
Quantization error results in when an analogue signal is represented by a series of discrete steps differing by the resolution of the digitizing process.
Hence, the primary concern in any measurement problem nowadays would be the expected frequency of the phenomenon measured, especially if it is above the 200 Hz ceiling of the past. With more and more electronics aboard modern sensors, the choice of the right transducer is further complicated. At SensorData we believe the appropriate transducer for any measurement problem, is the one that can output a number of measurements per second that is at least twice the frequency of the phenomenon measured to be able to reconstruct the phenomena as accurately as feasible.
Once this is established, sensor natural frequency and dynamic balancing would then determine the transducers speed worthiness.