Brigadoon - An Occasional Corner on the Internet
Circuit for Ground Temperature Sensor
Designed by Mark Little
As discussed previously, this ground temperature sensor array uses the analogue temperature sensor, the LM335. Also discussed previously, there are three temperature sensors (10cm, 20cm and 50cm).

Individual Sensor Circuit and Calibrated Error
The diagrams above come from the SGS-Thomson data sheet for the LM335. They indicate that to operate and calibrate the temperature sensor to an error less than 0.5°C over the expected temperature range requires a load resistor (R1) and a 10k calibration potentiometer.

The value of R1 does not affect the accuracy of the temperature sensor as long as the sensor current is in the range of 450µA to 5mA. The value of R1 must be selected to keep the current drawn by the three sensors to a range that can be supplied via the cable as that cable has a resistance of its own.

As a rough calculation of R1, assume that the 10k calibration pot is not fitted, the V+ is +5V and the total resistance of the power supply cable is 1.0 ohms. With this configuration,setting R1 to 1.8k will result in the LM335 drawing about 1.1mA at 25°C which is approximately the minimum permitted sensor current. Of course, this is a rough calculation as it does not take into account the current drawn by the 10k pot and the fact that the current for three sensors is flowing through the supply cable but it does show that setting R1 to 1.8k means that the LM335 will have enough current available to function correctly at room temperature. At 100°C (very unlikely as a ground temperature), the output of the LM335 would be +3.7305V. This would mean that the supply current would be in the order of  700µA, about 1.6 times the minimum acceptable sensor supply current.

As these estimates show, the resistance of the cable supply V+ to a sensor would not affect the ability of the sensor to work. However, what has not been calculated yet is whether the voltage drop in the earth return wire will have a significant effect on the voltage measured at the Raspberry Pi.