Alfred, I'm not exactly sure what you're asking. Are you asking how much of a practical signal is needed to see if there is a noticeable change in the ADC output of the device to ensure that it isn't noise? I'll try to explain it this way. We give noise tables in the datasheet on how much noise you can expect for different configurations of the device. Look at Tables 1 and 2 in the datasheet. For each setting, you can determine the RMS noise seen for any measurement. For the results, you would see something like Figure 3 showing the input-referred noise in the device. Normally, I would imagine that you would be able to practically detect a signal that is about 6x of the RMS noise value. The output noise of the measurement typically has a Gaussian distribution, and 6x of the RMS (or standard deviation) of the noise would be significantly outside the normal measurement. I'm not sure that's what you had asked. If the question is different than I've interpreted it, then post back. Joseph Wu
↧