Measurements with the T3Ster equipment are essentially absolute voltage measurements, where data are measured and stored as differences from a reference voltage level. However, because of the nature of the temperature sensors, these are interpreted in most cases as differential temperature measurements.
The T3Ster Measurement Control & Evaluation software stores the measured curves in .raw and .rec files (these latter are processed for filtering electric transients. These file formats are described in the Appendix A of the User's Manual of the T3Ster Measurement Control & Evaluation Tool)
A sample recording is shown below in Table A
Table A: A .rec file saved by
the T3Ster software
For deriving real absolute temperatures, there are different ways. All work like measuring the temperature at one point with both your (relative) temperature sensor and an external thermometer, or measuring the temperature of a known source (cold-plate, thermostat) with your temperature sensor. Afterwards adding the measured relative temperatures to the reference yields the absolute values.
The T3Ster software displays relative temperature curves, fitted at their cool end.
In case of diode type sensors the temperature dependent parameter one really measures is the forward voltage change at a fixed operating current. One always assumes a linear approximation of this dependence while in reality this is an exponential curve. Although the linear approximation is correct, but its slope is very dependent on the operating point, i.e. the diode current. This means the diode should be calibrated at its operating current otherwise the measurement is incorrect. This process is automated when using T3Ster controlled thermostat option.
The T3Ster thermocouple pre-amplifiers are calibrated at 25 °C. Here the sensitivity is set to 5 mV/K and this value should not be changed on the gain potentiometer of the preamplifier.
In case the temperature range is large (tens of °C) and the absolute temperature value is not of special interest (i.e. knowing it within 1 centigrade is satisfactory), one can simply use the voltage output of the pre-amplifier divided by this 5 mV/K sensitivity as absolute value. (The precision of reading the relative temperature change is much higher, of course.)
Table B: Thermocouple non-linear characteristics
If one needs a much better absolute precision in a very large range, one has to consider the nonlinearity of the thermocouples into account. This causes quite large deviation.
Table B contains nominal values for European standard metal alloys. Deriving exact absolute temperature values measured by thermocouples values,
In case the temperature change is in the range that one accept as linear (for example between 10 and 40 °C as Table B suggests) one can do the thermocouple measurement in a linearized way correctly, just like for diode measurements. And this is the typical case at T3Ster measurements. Thus
An example on processing results is shown in Table A. This table originates from a .rec file of the T3Ster software.
The first channel recorded a diode measurement, the second and third channels were connected to thermocouples using thermocouple pre-amplifiers.
The items without decoration were in the original file. Items in boldface are added comments and calculated values. Boldface italics numbers are the sensitivity values and the starting temperature values that have to be changed according to the simple calibration described above.
After the row, where the transient starts, Column A is the time in microseconds, B, C and D are the A/D values read in bits, and F, G and H are the calculated temperature changes. You also may add the starting temperature values to the F, G and H columns, if needed.