Eliminating Sensor Errors in Loop Calibrations

Calibrating a loop is more than just 4 mA to 20 mA

Significant performance improvement can realized by optimizing the loop calibration measurement system to better accommodate the unique characteristics of the temperature sensing element. All temperature probes and their sensing elements are unique, with variations in materials, construction and usage, or exposure to different environments. This uniqueness continues throughout the useful life of the sensor, in the form of drift due to mechanical shock and vibration or to contamination of the materials when exposed to the material they are measuring. Only through periodic verification can these differences and changes be accommodated, improving total measurement performance.

Temperature plays an important role in many industrial and commercial processes. Examples range from sterilization in pharmaceutical companies, metal heat-treatment to ensure optimal strength in aerospace applications, temperature verification in a cold storage warehouse, and atmospheric and oceanographic research. In all temperature measurement applications, the sensor strongly affects the results; unfortunately, many measurements are made without optimizing the system to get the best performance from the temperature transducer.

The majority of process temperature measurements are performed using a sensing element connected to a transmitter. Figure 1 shows a diagram of a common configuration.

Figure 1. Diagram of a typical process temperature measurement system

In many applications, it is common to verify the elements of the measurement system separately, but in doing so, significant improvements made possible by considering the system as a whole are ignored. One of the main reasons the elements are verified or calibrated separately is that it is often considered to be more efficient. Verifying the meas­urement component is done simply and quickly with an electronic thermocouple (TC) or resistance temperature detector (RTD) simulator. This approach does not verify the performance of the associated temperature probe, and assumes all probes are identical and closely follow some standard. In practice, no two probes are identical; they all vary from the ideal standard, and over time and usage their characteristics change. Understanding how probes vary from the ideal will allow you to optimize the measurement system to achieve the best performance

Figure 2.

Figure 3.

Rosemount Inc. uses the example provided in Table 1 for information on the possible performance improvement of their Model 644H Smart Temperature Transmitter. To achieve this performance improvement, the Rosemount 644H is given information (Callendar Van Dusen Coefficients) that allows it to correct for the unique performance of the temperature sensing element, in this case a standard IEC751 Pt100 sensor.

Dry-wells and microbaths are good choices for verifying the performance of temperature probes and other related sensors. But they do not have the capability to calibrate the transmitter’s output or readout and, by themselves, do not allow the entire measurement loop to be optimized. A heat source, combined with an intelligent electronic process calibrator that is capable of calibrating the transmitter and readout, is required if the above performance improvement is to be realized and maintained.

Total system accuracy calculated using RSS statistical method Table 1

By combining the automating and documenting capabilities of the Fluke 744 Documenting Process Calibrator with Hart Scientific’s intelligent and stable family of field dry-wells and micro-baths, you have the capability to test the entire loop. This combination of equipment allows you to easily verify the characteristics of the temperature sensor and measurement electronics. Using this information, the entire loop can be adjusted to optimize system measurement performance. Below are some examples of how to optimize the performance of your measurement system using these instruments.

The Fluke 744 is connected to a Hart Scientific dry-well or microbath by way of a serial RS-232 interface cable. Version 2.3 or greater firmware for the 744 is required. The firmware version is displayed briefly on the display of the 744 during power-up. If you do not have the required firmware, contact Transcat for information regarding an upgrade. The serial cable may also be obtained from Transcat. The heat source is connected to the 744 pressure port and is accessed by the 744 TC/RTD source key. Due to the length of these tests, it is recommended that a fully charged battery or battery eliminator for the 744 be used. A diagram of the connection of this equipment is pictured in Figure 4.

Figure 4. Connecting a Fluke 744 to a Hart Scientific dry-well

In many process applications, the instrumentation of choice for temperature measurements is a transmitter that accepts the output from the temperature sensor and drives a 4-20 mA signal back to the PLC, DCS or indicator. This example describes one method for verifying performance and offers to optimize this measurement to improve performance.

To perform this test, the RTD sensor is removed from the process and inserted in to the dry-block calibrator. The mA connections from the transmit­ter are connected directly to the 744 Documenting Process Calibrator (see Figure 5). In most applications, this solution provides adequate performance. But if your application includes a uniquely shaped sensor, you might want to consider the use of a micro-bath. If increased heat source accuracy is needed, the use of a reference thermom­eter combined with the 744’s User-Entered Values feature can be used. See application note 1263925 for more information on 744 User-Entered Values.

Once connections are made, you are ready to acquire trans­mitter configuration (if you have a transmitter with HART communications), set the test parameters, and configure the calibrator for mA measurement and dry-well control as the sourcing parameter.

Pressing the HART key on the 744 allows the calibrator to acquire the transmitter configuration from a transmitter with HART communication capability. Following is a sample of this acquired configuration information.

Pressing the HART key on the 744 again presents the following screen with several options for configuring the calibrator to the correct parameters for this test. For the purposes of this example, we’ll use the transmitter configured to output a 4-20mA signal; therefore the correct configuration of the 744 is to measure mA and source temperature via the dry-well.

Pressing the AS FOUND soft key on the 744 provides access to parameters needed to configure an automated test. Below is a typical definition that will test the measurement system from 50°C to 150°C sourcing temperatures using a dry-well in ascending order.

After the test has been defined, the Fluke 744 will take over and run the test recording the sourced temperature, measured output of the transmitter, in mA. At the end of the test, the results will be displayed on the screen, allowing the test technician to evaluate the results and take corrective action if needed. Following is an example of the results.

One method of optimizing this system to minimize error is to shift the URV or LRV of the transmitter to the values measured by the 744. With a transmitter with HART capabilities, this is easily done via the 744, by simply entering new values in the HART SETUP screen below. With an analog transmitter, you will need to mechanically adjust the Zero and Span adjustments when sourcing the appropriate temperature curve values. The 744 has a convenient menu key that allows you to easily set the correct value on the dry-well with a single button press.

Figure 5.

Calibrating and adjusting measurement systems using calibration constants

Another method of reducing uncertainty and optimizing temperature measurement sys­tems is to carefully characterize the temperature sensor, calculate correction coefficients and load these correction coefficients into the measurement equipment. This is the method used in the Rosemount 644H example on the previous page. This method does a better job of reducing the error in the measurement system that comes from the sensor. But it requires transmitters that have a correction or linearization algorithm that can accommodate the sensor. For example, Platinum RTDs typically use the Callendar-Van Dusen (CVD) equation for linearizing the sensor’s output. A characterized sensor will provide unique CVD coefficients that can be input into the transmitter, allowing its conversion algorithm to more closely match the unique characteristics of the sensor.

The Fluke 744 connected with a dry-well can help to collect the necessary information to characterize the sensor, but additional software and resources will be needed to take this data and generate new CVD constants. Examples of the required software include Hart Scientific’s TableWare or MET/TEMP II. Other software that could be used are Mathcad, Mathematica, Maple or Excel, but these packages require a considerable knowledge of the equations used to linearize these sensors and the ability to use these packages to perform ing fitting of the collected data.

The method of characterizing a probe is similar to the procedure above, but instead of measuring the output of the transmitter, the output of the sensor is connected directly to the 744. An example of data collected by a 744 on a temperature sensor is shown below.

Figure 6.

Data like this can be entered into Hart Scientific’s software using the screens in Figure 6 and then unique CVD constants calculated for that probe.

These coefficients can then be entered into a suitable measurement device that allows its linearizations to match the characteristics of the probe.

Summary

Using a dry-well in combination with a process calibrator allows measurement systems to be verified and adjusted to optimize measurement performance. By verifying the entire measurement system, unique characteristics of the sensing element can be com­bined with the measurement electronics to minimize measurement error. This can result in a significant reduction in measurement errors. The Fluke 744 Documenting Process Calibrator, combined with a Hart Scientific dry-well makes this process faster and easier.

Content for this article was provided by Fluke.
Fluke. Keeping your world up and running.