I'd like to make a contribution to this thread on the topic of accuracy.
A lot of people seem to get confused between accuracy and precision when discussing instruments and measuring devices. Even my copy of Chambers Concise English Dictionary erroneously equates the two. The point I wish to make is that 'precision' and 'accuracy' are not synonyms - least not in scientific disciplines.
I'll try to justify that claim with an example.
Suppose that I have two large-scale ammeters, each with an F.S.D. of 5 amps.
The first meter is of the moving-coil type with a linear scale. That scale is divided into 1 amp and 0.1 amp divisions in a clear manner. However, the mechanical quality of the meter is poor: cheap bearings, no attempt to compensate for temperature variations affecting the indicted readings is prone to give errors when a substantial current has been left flowing through that meter for any substantial length of time or when situated in the presence of a magnetic or an electrostatic field. The nature of the fine graduations of the scale provides for high precision, but on account of its mechanical limitations, the accuracy of any reading is low.
Hence, this meter has high precision, low accuracy.
The second meter is of the moving-iron type with the usual non-linear scale for a meter of this type. The scale has divisions in 1 amp and 0.5 amp. steps: nothing finer. The quality of the mechanics is very high: double-jewelled bearings, built-in temperature compensation and it is hermetically-sealed with a metal case for screening which carries an earthing stud. The accuracy is high: when a known current is passed through it, the needle always goes to exactly the same place on the dial, irrespective of the environment that the meter is in. But on account of the relatively coarse and non-linear scale calibrations, the precision of any reading is low.
Hence, this meter has low precision, high accuracy.
And finally: an extract from 'The Radio Laboratory Handbook', 6th. edition, 1954, page 36; author is M. G. Scroggie, B.Sc., F.I.E.E.
It is inadequate to state that the accuracy of an instrument is so-much percent. It would obviously be more sensible to state a figure for the inaccuracy, since then there would be no doubt about whether a greater accuracy meant a greater percentage or a smaller one. But other questions arise: Does the stated figure [for accuracy] mean the greatest possible difference between the value shown by the instrument [now] and the absolute value [indicated] at the time of calibration, or does it mean the greatest subsequent variation in value - and if so, over what range of temperature, humidity, frequency, etc. The latter, by itself, can be distinguished by the term constancy.
End of extract: entries such as [xxxxx] in the above are my additions to the original text in order to make the author's meanings abundantly clear.
End of post: thanks for wading through it.
Al. / Skywave / April 18, 2013 //
A lot of people seem to get confused between accuracy and precision when discussing instruments and measuring devices. Even my copy of Chambers Concise English Dictionary erroneously equates the two. The point I wish to make is that 'precision' and 'accuracy' are not synonyms - least not in scientific disciplines.
I'll try to justify that claim with an example.
Suppose that I have two large-scale ammeters, each with an F.S.D. of 5 amps.
The first meter is of the moving-coil type with a linear scale. That scale is divided into 1 amp and 0.1 amp divisions in a clear manner. However, the mechanical quality of the meter is poor: cheap bearings, no attempt to compensate for temperature variations affecting the indicted readings is prone to give errors when a substantial current has been left flowing through that meter for any substantial length of time or when situated in the presence of a magnetic or an electrostatic field. The nature of the fine graduations of the scale provides for high precision, but on account of its mechanical limitations, the accuracy of any reading is low.
Hence, this meter has high precision, low accuracy.
The second meter is of the moving-iron type with the usual non-linear scale for a meter of this type. The scale has divisions in 1 amp and 0.5 amp. steps: nothing finer. The quality of the mechanics is very high: double-jewelled bearings, built-in temperature compensation and it is hermetically-sealed with a metal case for screening which carries an earthing stud. The accuracy is high: when a known current is passed through it, the needle always goes to exactly the same place on the dial, irrespective of the environment that the meter is in. But on account of the relatively coarse and non-linear scale calibrations, the precision of any reading is low.
Hence, this meter has low precision, high accuracy.
And finally: an extract from 'The Radio Laboratory Handbook', 6th. edition, 1954, page 36; author is M. G. Scroggie, B.Sc., F.I.E.E.
It is inadequate to state that the accuracy of an instrument is so-much percent. It would obviously be more sensible to state a figure for the inaccuracy, since then there would be no doubt about whether a greater accuracy meant a greater percentage or a smaller one. But other questions arise: Does the stated figure [for accuracy] mean the greatest possible difference between the value shown by the instrument [now] and the absolute value [indicated] at the time of calibration, or does it mean the greatest subsequent variation in value - and if so, over what range of temperature, humidity, frequency, etc. The latter, by itself, can be distinguished by the term constancy.
End of extract: entries such as [xxxxx] in the above are my additions to the original text in order to make the author's meanings abundantly clear.
End of post: thanks for wading through it.
Al. / Skywave / April 18, 2013 //
