Temperature References for Highest Accuracy Industrial Thermocouple Measurements
Reference Sensor Errors
Static errors from the Resistance Temperature Detector (RTD) are associated with how it was calibrat-ed, how well it tracks over the ambi-ent range and drift that may occur with aging. Dynamic errors stem from a lead/lag effect where the refer-ence is responding earlier or later than the various cold junctions. Tracking Errors: To a large extent cold junction temperature accuracy is determined by how closely the refer-ence matches or tracks its theoretical curve over the measurement range. It is not unusual to see specifications which list a reference tracking error of 0.02°C to 0.05°C per degree of shift from calibration temperature. Consequently, a reading taken 10?C from the calibration point can pro-duce up to a 0.5°C error. Proper calibration techniques will minimize this error. Instrumentation that uses a quadratic fit to three tem-perature/resistance calibration points for the reference RTD can reduce this error to 0.10°C over the entire 0 to 60°C range. Aging Errors: Reference RTD’s that are annealed and protected from deterioration due to moisture and vibration drift no more than 0.05°C per year. Figure 1. The maximum error (max. temperature difference between any two points) as a function of the magnitude of a step change in ambient tempera-ture where the UTR temperature was equal to the ambient temperature before the change. Figure 2. The relative magnitude of the maximum error as a function of elapsed time after the step change in ambient temperature. Figure 3. The distribution of error as a function of terminal location. Non-uniformity Errors in the UTR Plate Errors induced in the UTR plate are controlled by two parameters: thermal isolation from the environment, and thermal coupling within the plate. Since no significant power is dissipat-ed within the enclosure, static non-uniformity can only be produced by temperature gradients in the external environment. Dynamic errors occur due to changes in the ambient tem-perature. Static Non-uniformity: A thermal scatter error results when the cold junction terminals and their associated cold junction refer-ence device differ in temperature because static temperature gradients are imposed on the measurement assembly. Imbalance of the junction terminals can result if the UTR is exposed to non-uniform ambient conditions. To minimize, mount the UTR to a surface with consistent composition and in an area where ambient condi-tions will have the same effect on the entire UTR. Air flow past one end of the UTR, for example, will cause an imbalance. Dynamic Non-uniformity: In spite of the high thermal conduc-tance between terminals on the UTR, it is possible to develop small differences in temperature between terminals due to unbalanced rates of heating or cooling. The amount of unbalance in the relative heating or cooling rates at various points of the UTR is a func-tion of the total rate of heat-transfer between the UTR and ambient. Placing the UTR within an insulated enclosure, reduces the unbalance to very low levels. The principal thermal paths from ambient to the UTR include: 1.Conduction through the enclosure cover and insulation. 2.Conduction through the mounting panel. 3.Conduction through the input and output wiring. Terminals at the edges of a UTR are thermally coupled to ambient some-what more closely than the central terminals. The edges of the UTR have a ratio of surface area to ther-mal capacity which is greater than the central portion. As a result when a UTR operates in a varying ambient temperature, the ends of the UTR will tend to be slightly warmer than the center if the ambient is increasing and slightly cooler if the ambient is decreasing. The principle heat path is from the enclosure through the insulation. While the design of Kaye’s standard UTR’s have considered these error sources, UTR enclosures have been provided for particular installation requirements to optimize uniformity for specific applications. When start-ing the jet-engine test cell on a cold day, for example, one enclosure performs with a maximum error of less than 0.3?C when exposed to a 25?C rise in ambient temperature over 40 minutes. For applications where the UTR assembly is exposed to extreme ambi-ent conditions, Kaye’s water cooled UTR version provides temperature stability and uniformity.
Select the RTD You Need Kaye offers two NIST*-traceable RTD configurations for monitoring the UTR plate temperature: RTD-20 and RTD- 100. Each unit is provided with indi-vidual quadratic equations which gives temperature as a quadratic function of the unit’s output. *(National Institute of Standards and Technology.) The Pt 100Ω RTD 4-wire film resis-tance configuration (RTD-100) is designed for continuous vibration environments. It is housed with 4 screw terminals for excitation and measurement. Excitation is a current of 1mA rms or less. Easy to access and remove, the RTD-100 provides time-savings and convenience when you need to recalibrate the reference. The RTD-20 is a 4-wire bridge config-uration without terminals. Excitation is a voltage of 5 to 24 Volts. Two voltages of less than 100mV are moni-tored. The diagram to the right shows the monitored points. Whichever RTD you choose, the exceptional thermal characteristics of the UTR’s require that only one RTD be attached to each UTR plate to monitor its temperature. However, provisions have been made to mount two RTD elements on a plate when required by the customer. (See configurations listed on the back page.)
Multi-channel Ice Point Reference System Ice Point Temperature Reference Equipment The K170 Ice Point Reference per-forms ice point referencing for up to 75 thermocouples. The user wires external thermocouples to the unit’s input terminals which are in turn connected to matching internal TC’s that terminate to copper at the tem-perature of a thermoelectrically produced ice-water mixture. Thermo-couple grade copper wire is taken from ice to MIL style connectors for output. Individual pass thru shield connections can also be provided.
Discussion of Temperature Measurement Errors The degree of temperature measure-ment accuracy you can achieve using an ice point reference depends on the grade of T/C wire selected and the level of calibration you use. With each calibration step described below, you can successively improve temperature measurement accuracy.
Use premium-grade thermocouple wire for consistent results. While the difference in cost between standard and premium-grade wire may be significant, the accuracy you can attain with high quality wire is superior. Standard-grade Type N wire with a 0.75% limit of error, for example, will produce a 3°C error at 400°C. On the other hand, the premi-um grade contains a 0.4% or 1.6°C error at the same measured tempera-ture. The K170 uses only premium-grade wire. Even after calibration, the premium grade wire will yield better accura-cy due to its more homogeneous composition.
How to Meet Your Accuracy Needs. Assuming the use of premium-grade wire, you can employ the following guidelines to meet your measurement accuracy goals. The example below uses Type N wire. Method 1: At a gross level using the standard curve fit for a premium Type N T/C—no calibration per-formed—you can expect about a 1.6°C error at 400°C. Method 2: You obtain an order of magnitude improvement or about a 0.2?C error with a 3-point calibration of your external T/C’s, independent of the K170. The 0.1?C error from the calibration plus the 0.1°C from the K170 equal the total error of 0.2°C. (Assuming the K170 terminals to be at 25°C, the difference from ice point times the 0.4% wire error equals 0.1?C.) Method 3: Improve measurement error further by calibrating with the external thermocouples wired to the K170. This method reduces the error to about 0.1°C plus 0.4% of differ-ence between the terminal tempera-ture at calibration and at the time of taking actual data.
Thermocouple circuit of the K170 with external T/C wire connected to input terminals. Method 4: Calibrate the internal and external T/C’s separately when you need to reduce temperature errors to the lowest possible value. This method reduces the error inherent in the T/C characteristic differences between the internal and external wire, providing a total mea-surement error of 0.1?C. Perform this calibration by monitor-ing the temperature of a shorted terminal when the K170 has stabi-lized at each ambient calibration point. Kaye offers this calibration as a service.
Ordering Information