7.3.4 Filtering

Remote junction temperature sensors are usually implemented in a noisy environment. Noise is most often created by fast digital signals that can corrupt measurements. The TMP461-SP device has a built-in, 65-kHz filter on the D+ and D– inputs to minimize the effects of noise. However, a bypass capacitor placed differentially across the inputs of the remote temperature sensor is recommended to make the application more robust against unwanted coupled signals. For this capacitor, select a value between 100 pF differential and 1 nF. Some applications attain better overall accuracy with additional series resistance. However, this increased accuracy is application-specific. When series resistance is added, the total value must not be greater than 1 kΩ. If filtering is required, suggested component values are 100 pF differential and 50 Ω on each input; exact values are application-specific.

Additionally, a digital filter is available for the remote temperature measurements to further reduce the effect of noise. This filter is programmable and has two levels when enabled. Level 1 performs a moving average of four consecutive samples. Level 1 filtering can be achieved by setting the digital filter control register (read address 24h, write address 24h) to 01h. Level 2 performs a moving average of eight consecutive samples. Level 2 filtering can be achieved by setting the digital filter control register (read address 24h, write address 24h) to 02h. The value stored in the remote temperature result register is the output of the digital filter, and is the value that the ALERT and THERM limits are compared to. The digital filter provides additional immunity to noise and spikes on the ALERT and THERM outputs. The filter responses to impulse and step inputs are shown in Figure 11 and Figure 12, respectively. The filter can be enabled or disabled by programming the desired levels in the digital filter register; see Table 5. The digital filter is disabled by default and on POR.

Figure 11. Filter Response to Impulse Inputs

Figure 12. Filter Response to Step Inputs