ADS124S08-EXCEL-CALC-TOOL: Why the conversion period is not 50ms when Data rate setting (SPS) is 20?

Hi Kazuya-san,

The basic data rate is selected in the register settings in the DATARATE register (0x04).  However, there are several additional selections that can also affect the output rate.  One is the digital filter mode selected, and the other is whether global chop is enabled/disabled. 

When enabling global chop, the conversion time to completion of each measurement will include the time for measuring the input twice to average out the offset.  The total length of time will depend on the digital filter mode selected.  The low-latency filter (LL) is the fastest as the period for the first conversion will be very close to the datarate selected.  The sinc3 filter requires 3 complete internal conversion periods for the digital filter to fully settle before the results are valid.  So it is complicated as to conversion period, output rate, digital filter mode, etc..  I think that part of the misunderstanding is with what is considered a 'conversion' period with respect to when the output data is available.

From the graph you posted you will notice that the very first period where data is available takes 6 regular conversion periods (3 periods of sinc3 initial condition + 3 periods of sinc3 with inputs swapped).  The two measurements are averaged to remove the offset.  Subsequent output periods will be 1/2 of the initial output period (assuming no mux change).  The reason it is now 1/2 of the the initial period is each successive period uses the last conversion of the previous period.

The low-latency filter with global chop is much faster than the sinc3 filter.  Tables 18 and 19 in the ADS124S08 datasheet show the actual times for each use case and register settings for data rate. 

So I hope the previous discussion answers question 1.

For question 2, yes.  When 20sps is selected in the DATARATE register, the time for first valid data results depends on the digital filter and if global chop is enabled or disabled.

Regarding question 3, if global chop is enabled, you have an average of 2 measurements.  If the sinc3 digital filter is used, then you need 3 conversion periods per measurement. After the initial conversion cycle, there will be data from the first output period used in the next period and the inputs will be swapped for the next measurement.  So, as 20sps is approximately 50ms, then 3 periods is 150ms.  The throughput measurement rate (with no mux change) will be approximately 6.7sps output rate.

If the low-latency filter is used instead of sinc3, then for the selected 20sps the initial measurement takes about 113ms, and each succeeding measurement is 56.504ms which is about 17.7ms which is not quite 3 time faster than the sinc3 filter.

Best regards,

Bob B