Original question:
In Auto scan mode in ADS1158, if we are utilizing all 16 channel, and lets say I am using 8usec time delay, no chopping, and 4 averages, with this config, let's say I am getting around 800/900 samples per second data rate. Can we accurately capture a 5 kHz sine wave on one channel ? What parameter will tell us if can get accurate capture or not?
Hi Ranjita,
You can better understand how a delta-sigma ADC will sample a sinewave using the image below (from TI Precision Labs: https://training.ti.com/ti-precision-labs-adcs)
As you can from the right half of the image, a delta-sigma samples over the entire conversion period (other than the delay time at the beginning and some small overhead time at the end. You can learn more conversion latency in delta-sigma ADCs with this app note: https://www.ti.com/lit/sbaa535 ). If you are using the ADS1158 with 4 averages in Auto-scan mode, then each channel is sampling at ~15 kSPS. That means that the ADC will sample about 1/3 of the 5-kHz sinewave input period and then provide a single, "averaged" result. Then, you will scan the remaining 15 channels, and wrap back around to scan the sinewave input again. At this point, you would sample about 1/3 of the sinewave period, though it's highly unlikely you would begin and end sampling at the exact same point in the sinewave period compared to the previous conversion. Therefore, I am really not sure what useful information you could gain from this measurement.
In general, most AC-signal sampling is performed with either a SAR ADC or a higher speed, simultaneous sampling delta-sigma ADC such as the ADS1278. I would not recommend using the ADS1158 for this purpose.
-Bryan
Hi Ranjita,
You can refer to Table 3 in the ADS1158 datasheet for more information, which I have copied below.
As footnote #3 states, the "effective data rate for multiple channels... is divided by the number of active channels in a scan loop". To clarify this, it helps to think of this in terms of time, not samples per second. For example, if you are measuring all 16 channels when DRATE[1:0] = 10b, then each channel is being sampled at 15123 SPS. This is approximately equivalent to saying that it will take ~66us to sample each channel. Therefore, it takes 16*66us = 1056us to measure the entire loop.
Looking at just the first channel (or any individual channel), you get a pattern like this: sampling for 66us --> not sampling for 990us --> sampling for 66us --> not sampling for 990us --> etc. (note that "not sampling for 990us" refers to that specific channel. During this time the ADC is sampling the other 15 channels). So the effective data rate for that channel is equal to the data rate from Table 3 divided by the number of channels being sampled, which in this case is 15123 / 16 = 945 SPS (note that the inverse of 945 SPS = 1058us, or approximately the time I said it would take to the measure the entire loop in the previous paragraph. So it checks out).
However, in all cases, each channel is being sampled for the same amount of time (for a given data rate), regardless of how many channels you choose to sample in a loop. I will admit this is a bit confusing, so let me know if you have additional questions.
-Bryan
