3.3.1 When to Use

The difference between the modes is in how quickly samples are acquired.

Burst average sampling is used best for a quick, short-term averaging operation. For extended periods, Averaging mode is a better choice.

However, it is possible to get the benefits of both Averaging modes by using a little bit of software. Rather than taking one sample for each point of the long-term average, use the burst-average to acquire a short-term average at each point. This has two benefits; averaging out short-term noise and obtaining a higher resolution from the ADC by oversampling the signal.

Oversampling a signal increases the effective resolution of the ADC by 1 bit for every log ⁡ 2 (# of samples). For instance, oversampling a signal by four will generate a result with two additional bits of resolution(1).

To compute the long-term average, take the sum of the results from the burst average and add them up in software. Then, divide (or bit-shift) the accumulation to get the long-term average.

Note:
  1. The extra bits of resolution will have more statistical noise than the standard bits of the ADC. However, averaging the accumulated result down will reduce the amount of noise present in the result. In other words, collecting 16 results and right-shifting by two will yield two additional bits with lower noise than accumulating eight results and shifting by 1.