If the number of measurements (N samples) is small, any one
data point can distort our mean value. Therefore, we need to
take enough samples to be representative of the population.
...
As N ? ? (or the entire population), the effect of any one point
is minimized, and our measured average approaches that of
the entire population.
...
Random Errors
� Caused by a lack of repeatability
� Can originate from the measuring system itself, from the experimental system, or from the environment.
� Generates a scatter of measurement results
� Some sources: temperature drift, electrical noise
Random Error = Re
Calibration
Test during which known measurand values are applied to the measurement
device under specified conditions and the corresponding output readings are recorded.
Interpolation
is an estimation of a value within two known values in a sequence of numbers
Intrusive Measurement
the sensor is immersed or embedded
into the material being measured (and may affect the outcome).
Ex. Bulb-type thermometer
Non-Intrusive Measurement
the sensor does not physically
interact with the material being measured. Ex. Infrared
Thermometer
Systematic Errors
Consistent, repeatable errors
� calibration errors
� loading errors
� spatial errors
Systematic Error = Average of Readings - True Value
Replication
Independent duplication of a set of measurements under
similar operating conditions
Sensitvity
The sensitivity of an instrument is the ratio of the change in
magnitude of the output to the change of magnitude of the
measurand
sensitivity = (d (output))/(d input) = (change in output )/change in input
Range
Measurand values, which a measurement device is intended to measure, specified by
the upper and lower limits.
Repetitions
Repeated measurements we make during any single test run or on a single batch
Hysteresis
A phenomenon that may degrade accuracy.
� If the measurand was increasing prior to taking the reading, different output readings may be obtained than if the measurand was
decreasing.
� Hysteresis is caused by such effects as friction, mechanical flexure
o
Output Span
If output operating range is ymin to ymax, output = ymax - ymin
Accuracy
The closeness in agreement between a measured value and the
true value.
� Note that although measuring device manufacturers frequently state
a value for "accuracy" as part of the device specifications, it is really
the inaccuracy that is specified.
Precision
� Another term frequently used to describe a measuring system
or component and characterizes the random error.
� Note that a highly precise measuring system will give the same
value each time it is read, but it may not be very accurate - it
may simply giv
Variables
Changing aspect (independent, dependent, extraneous, controlled)
Noise
Random variation of the value of the measured signal caused by variation of extraneous variables that can increase scatter and can be reduced or removed using statisitical techniques.
Interference
Effect tgar introduced a determinstic trend on the measured value. i.e trend that is predictable in time or space, such as a ramp or sine function.
Sensor Stage
...
Sensor Stage
Transducer
Device that converts physical parameters such as temperature, pressure, motion,
and so on, into an electrical output.
Signal conditioner
Device that modifies an electrical signal to make it more acceptable as an
input to other devices.
Extraneous Variable
...
Extraneous Variable
Independent Variable
...
Dependent Variable
...
Random Test Strategy
...
Resolution
Smallest detectable change in a measurement.
K
K = 1.87(N-1)0.40 + 1
Central Tendency
mean, median, mode
68.27%
x' +- ?
95.45%
x' +- 2?
99.73%
x' +- 3?