際際滷

際際滷Share a Scribd company logo
Engineering Testing and Measurements
Alex Liberzon
Tel Aviv University
March 22, 2014
Outline
1. Backup of the measurement system, types of signals
2. errors vs uncertainty
3. Calibration and calibration errors
4. Regression
5. Log-log
6. calibration errors
Generalized measurement system
Types of sensors/transducers
Displacement type of sensors - the change in the measured
quantity causes a change of the sensor/transducer
and creates a proportional signal.
Zero displacement sensor - a change of some quantity
through the closed loop control adjusts the sensing
point
 Advantages: very small deviations can be measured as these
are amplified by the control loop (passive or active) and it is
easy to compare the signal to the zero value.
 Disadvantages: frequency f [Hz] or the time response  [sec]
of the control loop affects the sampling frequency, fs
Types of sensors/transducers
Direct sensor- the sensor reacts directly to the quantity
measured, e.g. weight is measured by calibrated
weight balance.
Indirect sensor - the sensor reacts to another quantity rather
than the measured one and we can infer the measured
quantity by some physical law, e.g. change of
electrical resistance indicates the change of
temperature of the resistor.
Passive sensor - the only energy is from the measurement,
e.g. temperature measurement by the thermocouple
Active sensor - receives energy from external source, e.g.
piezo-electric pressure transducer, hot-wire
Types of signals
Analog: precise value is important, physical quantities like
length, voltage, temperature
Digital: a combination of N bits (0/1 or On/Off), no precise
value, less sensitive to analog noise, but limited
resolution = FOS/2N
Sensitivity
Error, accuracy, precision
Error: difference e between the measurement x and the true value
x0:
e = |x x0|
Accuracy: the ratio of error e to the true value or to some other
reference, e.g.
a =
e
x
100% a =
e
xmax xmin
=
e
xf
100%
percents of the value or percents of the full-scale
Precision, reproducibility: capability of the system to return the
same value
Example
The known value is 100, we measured: 103, 104, 105, 103, 105
the errror: e = 5, accuracy = 5% , precision: 賊1 around the mean
value of 104
bias, systematic error: 104 - 100
Signal-to-noise ratio: A/B
Full calibration analysis

More Related Content

lecture2.pdf

  • 1. Engineering Testing and Measurements Alex Liberzon Tel Aviv University March 22, 2014
  • 2. Outline 1. Backup of the measurement system, types of signals 2. errors vs uncertainty 3. Calibration and calibration errors 4. Regression 5. Log-log 6. calibration errors
  • 4. Types of sensors/transducers Displacement type of sensors - the change in the measured quantity causes a change of the sensor/transducer and creates a proportional signal. Zero displacement sensor - a change of some quantity through the closed loop control adjusts the sensing point Advantages: very small deviations can be measured as these are amplified by the control loop (passive or active) and it is easy to compare the signal to the zero value. Disadvantages: frequency f [Hz] or the time response [sec] of the control loop affects the sampling frequency, fs
  • 5. Types of sensors/transducers Direct sensor- the sensor reacts directly to the quantity measured, e.g. weight is measured by calibrated weight balance. Indirect sensor - the sensor reacts to another quantity rather than the measured one and we can infer the measured quantity by some physical law, e.g. change of electrical resistance indicates the change of temperature of the resistor. Passive sensor - the only energy is from the measurement, e.g. temperature measurement by the thermocouple Active sensor - receives energy from external source, e.g. piezo-electric pressure transducer, hot-wire
  • 6. Types of signals Analog: precise value is important, physical quantities like length, voltage, temperature Digital: a combination of N bits (0/1 or On/Off), no precise value, less sensitive to analog noise, but limited resolution = FOS/2N
  • 8. Error, accuracy, precision Error: difference e between the measurement x and the true value x0: e = |x x0| Accuracy: the ratio of error e to the true value or to some other reference, e.g. a = e x 100% a = e xmax xmin = e xf 100% percents of the value or percents of the full-scale Precision, reproducibility: capability of the system to return the same value
  • 9. Example The known value is 100, we measured: 103, 104, 105, 103, 105 the errror: e = 5, accuracy = 5% , precision: 賊1 around the mean value of 104 bias, systematic error: 104 - 100