Verifying the performance of platinum resistance thermometers, or RTDs, will help processors ensure they are achieving reliable and accurate temperature measurement.
When it is necessary to measure temperature, a resistance temperature detector (RTD) can provide years of reliable, accurate service. As with most high accuracy measurement devices, it is prudent to make periodic accuracy performance checks of an RTD against a known standard or specification to ensure that the integrity of the measurement is maintained. The best way to consistently maintain confidence in the RTD’s accuracy is to implement a periodic performance verification program. To establish a periodic verification program requires the user to set the initial verification frequency, define the verification testing techniques and use a logical method to optimize the verification period based on experience over time.
Determining the frequency of verification is a multi-variable problem that is difficult to solve using theory alone. Only by gaining actual experience with a particular design of probe in a particular installation can an efficient compromise between cost of verification and confidence in the measurement be found. Unfortunately, it is necessary to make an educated estimate of the initial verification period without the luxury of actual experience.
In lieu of actual experience, the RTD user has to use published data for the RTD’s performance and couple this with knowledge of the environment in which the probe is being used. Using good judgment, a conservative educated estimate of a suitable initial verification period can be established.
RTD Performance over Time
A competent, reputable manufacturer of RTDs will be able to provide sufficient data to characterize the performance of the probe under specific test conditions, typically at the extremes of the performance range. The probe specification should include the drift, vibration and other performance criteria.
Drift Due to Time at Temperature. The temperature the probe normally experiences can have an adverse effect on the accuracy of the probe. The higher the temperature, the more accelerated the deterioration in accuracy. The longer the duration at the temperature, the larger the accumulated effect on the overall accuracy. In the long term, this drift is attributable to contamination of the platinum sensing element and, at the design stage, can be mitigated by using appropriate materials and good construction techniques. The drift is somewhat predictable and a graph similar to figure 1 may be available. At a minimum, a single defining point should be provided. For example:
The probe ice point reading will change not more than ±0.13°C for 1,000 hours at the maximum rated temperature.
Drift Due to Temperature Cycling. When a probe is exercised between temperature extremes, the platinum sensing element can become stressed due to the expansion and contraction of the different materials used in the construction of the probe. The affect on accuracy is more detrimental when the temperatures are more extreme and the rate of temperature change is greater. A graph similar to the example in figure 2 may be available for a particular RTD. At a minimum, a single defining point should be provided. For example:
The probe ice point reading will change not more than ±0.13°C after 10 cycles between the maximum and minimum rated temperatures.
Vibration. In many applications, RTDs have to withstand the constant vibration of the equipment or process that it is monitoring. If the vibration is extreme enough, it can cause stress in the platinum sensing element, which can result in erosion of accuracy. RTD manufacturers typically state a vibration test schedule that a probe can withstand without a significant change in accuracy. A typical specification might be:
Less than ±0.075°C ice point shift after 30 minutes at 21 g peak vibration 5 to 350 Hz continuous sweep, at 68°F (20°C) for unsupported lengths of 5.5" or less.
The next three specifications benchmark the probes’ initial performance. If the initial performance is known, it can be used to determine the deterioration of the probe at subsequent verifications.
Initial Ice Point Resistance. The actual initial resistance at the ice point (R0) should be obtained from the manufacturer. If the actual ice point resistance is not provided, the manufacturer should be able to provide a nominal value with an interchangeability tolerance (e.g. 100 Ω ± 0.10 Ω).
Time Response. Time response measures the RTDs’ ability to match a step increase in temperature. A typical industrial probes time response is stated as:
4 seconds for a 63.2 percent response to water moving at 3 ft/sec.
However it may be more useful to benchmark the initial time response of the probe in the application.
Insulation Resistance. Insulation resistance (IR) is the resistance between the electrical circuit containing the sensing element and its outside environment. Often, the probe’s insulation resistance can be measured with the probe installed in its application. Insulation resistance acts as a shunt resistor to the measurement circuit; the lower the insulation resistance, the higher the effect on the accuracy of the probe. The manufacturer should be able to provide a threshold value for initial insulation resistance. A typical statement of insulation resistance is:
Greater than 500 MΩ measured at 500 VDC when the probe is at 68°F (20°C).
The user should also be aware that at elevated temperatures, insulation resistance will be significantly lower than the room temperature value. If an elevated temperature insulation resistance value is provided, this value could be important to the analysis for sensors used at elevated temperatures within their prescribed operating range.
Table 1 shows the specifications as defined by ASTM E1137. You may note that these values are stated as minimums. Most RTD manufacturers strive for much higher initial insulation resistance performance.
The next step is to evaluate the environment the RTD is being used in against the specification.
Temperature Cycles. It is important to understand the nominal temperature the probe will experience and the time held at that temperature. In addition, it is important to know how often the probe will see excursions to extreme temperatures.
Vibration. How much vibration the probe experiences when it is installed is often difficult to calculate, but an assessment of the severity should be made. High vibration environments should be a red flag that results in a relatively short initial verification period.
Corrosion, Erosion and Buildup. Some processes can cause the sheath or thermowell to erode or corrode away (figure 3). Other processes may cause a substance buildup on the sheath. A thermowell can protect the probe from its operating environment but even when a thermowell is used, corrosion, erosion and buildup can still negatively affect the probe operation, so adequate limits need to be set. Corrosion and erosion will - if left unchecked - eventually breach the probe’s protective housing and render the probe inoperable. Buildup will reduce the time response of the sensor and increase the self heating. The self-heating, caused by the measuring current passing through the resistance platinum element, can affect the accuracy of the temperature measurement. In installations with sufficient thermal contact with the process, the self-heating effect generally is negligible, but when buildup insulates the probe, it can become significant.
Infrequent Events. These include planned events such as wash down cleaning or unplanned events such as instrumentation voltage spikes that can very quickly negatively affect the accuracy of the probe.
Initial Verification Period
Armed with a clear understanding of the probe performance and the operating environment, the initial verification period can be established. First, decide on the maximum deviation from actual temperature the measurement can tolerate. This could be based on energy costs, temperature of material spoilage, etc. Once the maximum deviation is known, a predication as to the time when the probe might fall outside the acceptable accuracy limits can be made by using the probes’ published accuracy deterioration rates.
If buildup or corrosion/erosion rates are known, then an estimate as to the time in service when these might start to be a concern can be determined. If this yields a verification time that is shorter than the drift due to temperature, then the criteria should be used to set the initial verification period.
In addition to the time-based approach, consideration should be given to an event-based approach - even if only an abridged verification takes place. For example, in a wash down situation, it might be prudent that after each time the wash down is performed, the probes’ insulation resistance is measured and compared to the established benchmark.
Before any verification of the accuracy the RTD, an insulation resistance measurement should be taken and confirmed that it is still within the manufacturer’s threshold value.
Taking an R0 reading is the most efficient way to determine the accuracy of the probe. Fortunately, it is a relatively easy and low cost task to create an ice/water bath that provides a sufficiently accurate 32°F (0°C) temperature point. By monitoring the R0 value of the probe over subsequent verification intervals, it is possible to build up a history that can be used to predict the future performance of the probe.
Within the resistance vs. temperature curve that fully characterizes the probe, the R0 point is used commonly as a confirmation that the probe is consistent over time. If there is any doubt about the probes’ performance at other temperatures, or if the typical operating temperature is of most concern, then it should be tested at, or close to, those temperatures.
In the case of erosion/corrosion being a concern, then the outside diameter of the probe should be closely measured and monitored. If buildup is a concern, then some type of time response measurement should be tracked. This may be able to be done with sufficient accuracy using the process the probe is measuring, or it may be necessary to rely on industry standard techniques such as those detailed in ASTM E644-09 Standard Test Methods for Testing Industrial Resistance Thermometers. A separate inspection/cleaning schedule could be established to manage this effect. This schedule could match the performance verification timing or be timed with other system maintenance due to the need to open the process barrier.
The initial verification period is set using the RTD manufacturer’s specifications and the user’s knowledge of the probes’ operating environment. After a number of verifications are complete, the user should have enough R0 data to predict actual accuracy deterioration rates. These calculated rates are much more appropriate for the user’s particular application than those provided by the RTD manufacturers and should be used to optimize the verification period.