In general, it is impossible to offer a statement of total instrument accuracy that is both straightforward and valid. Any true indication of "total accuracy" would, in fact, have to take a large number of error components into account. Even when errors contributed by the transducer/cable/fixture system are ignored (and these can, of course, be considerable), such a statement would have to represent a summation of all initial internal inaccuracies—plus the worst-case effects of ambient temperature, time, line-voltage variation, random noise, etc., on these inaccuracies—plus digital resolution of the displayed measurement, if applicable.
"Internal inaccuracies" can arise from a number of sources, including drift of amplifier offset and gain, multiplexer effects, gain drift of the ranging amplifier, A/D nonlinearity, and inevitable limitations of software linearization. For individual Daytronic analog conditioner instruments and cards, we specify offset and gain accuracies (when applicable), plus the effects of temperature and time on these accuracies. In the case of Temperature-Signal Conditioners, a "Typical System Accuracy" is also given, along with a "Maximum Expected Error."
Though in some cases a statement of "total instrument accuracy" might be desirable, in most cases it would not be all that informative, since what it actually states is the worst possible combination of all possible errors. In practice, the achievable accuracy will greatly exceed that specified by any such "absolute" statement.
With regard to Daytronic 3500 Series and 4000 Series conditioner instruments and to individual Conditioner Cards for use in System 10, SPS6000, and/or SPS8000, it may be said that, following initial calibration of a given transducer-based data channel, the overall stability of the instrument or card will normally allow measurements by that channel to an accuracy of within 0.02% of full scale, except when limited by engineering-unit resolution considerations.
[Top
]
