TechArchive

Original article date: November 1997

Norman MacDonald, Sales Director of Status Instruments compares the traditional measurement techniques that have been developed for measuring accurately the resistance of platinum resistance temperature detectors with the use of “Smart” instruments.

Using the intelligence of microprocessors it is now possible to eliminate some of the drawbacks of three-wire Pt100 measurement systems. A properly designed and manufactured three-wire measuring system is as good as a four-wire system for the vast majority of applications, although it is still inferior when measuring via a zener barrier.

The traditional method of accurately measuring a resistance is to incorporate the resistance into a Wheatstone bridge circuit. A voltage is used to excite the bridge and the voltage across the bridge is proportional to the resistance of the measuring element.

A problem occurs when lead resistances are introduced. Any resistance in the lead appears as though an additional resistance is in the element to be measured. When this is a platinum resistance sensor with a fundamental resistance of 100ohms and a fundamental interval of 38.5ohms (as per BS1904), 0.4ohm of lead resistance or 0.2ohm/leg lead resistance would result in an effective measurement error of approximately 1C. In practice several ohms can be expected and hence the errors are unacceptable in what is considered to be a high accuracy measurement system.

To minimise these errors, the three-wire compensated bridge was introduced. This has the effect of removing the error introduced by the lead resistance, as long as the lead resistance RL1 and RL2 are matched.

However, the effect of the lead resistance can be to cause less current flow in the detector leg and hence a small but possibly significant span error can be introduced. This can be eliminated by exciting the bridge from a constant current source, rather than a constant voltage. Then, whatever the lead rsistance, the same current always flows through the detector. With this method, there are no lead resistance errors introduced, provided the lead resistances are equally matched. In practice, they are very closely matched if the wire used is part of the same multi-core cable, even over quite long runs, hence any lead errors introduced are so small as to be insignificant. The exception to this is when the sensor is used in a hazardous area and connected to the bridge circuit via a zener barrier. Any mis-match in the resistance of the two legs of the zener barrier can appear as a sensor error. Although still small, this error can be as much as 0.15ohm, or approximately 0.3C.

All transmitters and indicators form Status Instruments have traditionally used a variation of this technique, using the Status active bridge circuit. The exception to this is the new Smart series of instruments which use a different technique.

A more accurate method of measuring Pt100 elements is to use a four-wire current and voltage method. The detector is excited by a constant current and the voltage across the detector is measured by an amplifier with a high impedance input. If the current source is perfect and the input impedance of the voltage measuring circuit is infinite, then no error whatsoever is introduced by the lead errors, even if they are mis-matched. In practice, current sources are never perfect and impedance never infinite, but they can be made of a quality such that lead resistance effects can be reduced to insignificant levels.

The method used to measure PT100 on the Smart instruments is to have a universal input capable of supporting a wide range of inputs. This means that the input electronics needs to be as general purpose as possible. It is inconvenient and unnecessary to dedicate input pins and electronics to support a constant current supply and a bridge arrangement. The input circuit measures voltage to a high degree of accuracy, computes the resistance and translates it to an accurate temperature reading. The method is based on a resistor Rc to limit the current flowing and a stable reference resistor Rs.

In addition, the microprocessor can determine whether any one of the RTD inputs has become disconnected or burned out and detect other errors, such as RT short circuit. This is an improvement over both conventional three- and four-wire circuits.

The technique is very effective at removing lead resistance effects, as long as they are equal. Again, the problem when using zener barriers is that, if the legs of the barrier are not matched accurately, then a small error can be introduced. The SEM 230X, however, has addressed this problem by having an intrinsically certified, isolated front end which can be connected directly to sensors in the hazardous area without barriers, hence removing this one last disadvantage over a four-wire system.

• Status Instruments
• Norman MacDonald
• 01684 296818

November 1997