APPENDIX A - TRANSMITTER ACCURACY SPECIFICATIONS
The complex current-transmitter circuitry necessary to amplify, isolate, protect, and offset weak input signals while consuming only small amounts of power can distort the signal in many ways. Additional accuracy limitations occur in RTD transmitters, which require precise excitation, lead-resistance-effect cancellation and large zero-suppression ranges in order to obtain good sensitivity and linearity for high temperatures.

Many transmitter data sheets omit key accuracy factors and/or express performance in percentage values without mentioning the full-scale value. Design limitations can be disguised by such "specsmanship"; the 501 specifications, however, are detailed in order to present the complete performance accuracy.

Input errors are logically expressed in degrees (rather than ohms), and output errors are readily expressed in microamperes, since output is current. Transmitter users are rarely interested in microamperes, however, so these output current errors are translated back to input degrees as a percentage (or ppm) of the selected Span.

Another fundamental division of errors is that of independence or dependence on Zero and Reading. Resistor aging and tempco mismatch in the Zero and Voltage Reference circuits will produce errors which increase with Zero suppression but which are independent of the amount of Reading (value above the Zero). Resistor aging and tempco mismatch in the amplifier gain (feedback) circuits will usually affect both Zero and Reading accuracy; amplifier gain tempco variations are important to just the Reading stability. A complete error specification needs a term proportional to Zero (suppression) and a term proportional to Reading.

For RTD transmitters, the excitation current and amplifier offset tempcos are never precisely zero, even when factory-tailored over wide ambient excursions with curvilinear adjustments, as in the 501. This error component is readily stated as a percentage of the ambient temperature excursion from the nominal temperature at which the Zero was set (assuming, as in the 501, that the Zero potentiometer has ample resolution on all Zero and Span ranges). For transmitters with restricted turndown ratios (low zero-suppression capability), the tempco errors may be lumped into a single error term.

In addition to these three components of tempco (ambient temperature effects), there are other possible errors, often referred to as "time," "hysteresis," "repeatability," or "drift" errors. No statistically-signific-ant errors of these types have yet been observed for the 501, which utilizes a solid-state, band-gap input voltage reference, matched-pair input PNP transistors, integrated-circuit current source and zero tempco control, and matched-tempco bridge resistors. The 501 also provides a variable-tempco output adjustment (factory-set) which eliminates many of the errors lumped in this category for other units. Its specification includes a 0.2°C tolerance for the calibration accuracies.