Selection of the right thermometer
Accuracy
Many thermometers with resistance thermometers offer ppm, ohms, and/or temperature specifications. The conversion from ohms or ppm to temperature depends on the thermometer used. For a 100Ω probe at 0°C, 0.001Ω (1mΩ) equals 0.0025°C or 2.5mK. 1ppm is also equivalent to 0.1mΩ or 0.25mK. You also need to pay attention to whether the technical indicator is "reading" or "range". For example, "1ppm reading" is 0.1mΩ at 100Ω, while "1ppm range" is 0.4mΩ when the full scale is 400Ω. The difference is huge!
When checking accuracy specifications, remember that reading uncertainty has a small impact on the total uncertainty of the calibration system, and it does not always make economic sense to purchase the lowest uncertainty thermometer. The "Bridge-Super Resistance Thermometer" analysis method is a good example. A 0.1-ppm bridge costs over $40,000, while a 1-ppm super resistance thermometer costs less than $20,000. Looking back at the total system uncertainty, it is clear that a bridge can only improve performance by a small amount - 0.000006°C in this case - but at a very high cost.
Measurement error
When making high-accuracy resistance measurements, ensure that the thermometer is able to eliminate thermal EMF errors caused by dissimilar metal connections in the measurement system. A common technique for eliminating thermal EMF errors is to use a switched DC or low frequency AC current source.
resolution
Be careful with this indicator. Some thermometer manufacturers confuse resolution with accuracy. A resolution of 0.001°C does not mean an accuracy of 0.001°C. Generally speaking, a thermometer with an accuracy of 0.001°C should have a resolution of at least 0.001°C. Display resolution is very important when detecting small temperature changes - for example, when monitoring the solidification curve of a fixed-point vessel, or when checking the stability of a calibration bath.
linearity
Most thermometer manufacturers provide accuracy specifications at one temperature (usually 0°C). This is useful, but you are often measuring a wide range of temperatures, so it is important to know how accurate the thermometer is over its operating range. If a thermometer is very linear, its accuracy specifications will be the same throughout its entire temperature range. However, all thermometers have some degree of nonlinearity and are not completely linear. Make sure the manufacturer provides accuracy specifications over the operating range or linearity specifications that you use when calculating uncertainty.
stability
Since measurements are made over a wide range of environmental conditions and over various lengths of time, reading stability is very important. Make sure to check the temperature coefficient and long-term stability specifications. Make sure that changes in environmental conditions do not affect the accuracy of the thermometer. Reputable manufacturers all provide temperature coefficient indicators. Long-term stability metrics are sometimes combined with accuracy metrics-for example, "1ppm, 1 year" or "0.01°C, 90 days." Calibrating every 90 days is difficult, so a 1-year indicator is calculated and used in the uncertainty analysis. Be wary of providers that offer "0 drift" indicators. Every thermometer has at least one drift component.
calibration
Some thermometers have technical specifications that "do not require recalibration." However, according to the latest version of ISO guidelines, all measuring equipment needs to be calibrated. Some thermometers are easier to recalibrate than other devices. Use a thermometer that can be calibrated through its front panel without special software. Some older thermometers store calibration data in EPROM memory and use custom software for programming. This means that the thermometer must be sent to the manufacturer for recalibration - perhaps abroad! Since recalibration is time-consuming and expensive, avoid using thermometers that still use a manual potentiometer for adjustment. Most DC thermometers are calibrated using a set of highly stable DC standard resistors. Calibrating an AC thermometer or bridge is more complex and requires a reference inductive voltage divider and precision AC standard resistors.
Traceability
Measurement traceability is another concept. With good DC resistance standards, traceability of DC thermometers is very simple. Traceability of AC thermometers and bridges is even more complex. Many countries still do not have established AC resistance traceability. Many other countries with traceable AC standards rely on AC resistors calibrated through thermometers or bridges that are ten times more precise, significantly increasing the measurement uncertainty of the bridge itself.
