ADS124S08: internal or external clock

Hi Hao,

These questions do not have a quick and easy answer in regards to accuracy comparisons or which method is better.  In most applications the internal oscillator will be good enough.  Across the full operating temperature range the ADS124S08 oscillator can vary by +/-1.5% over the nominal value of 4.096MHz.  Due to this variation there may be a need to have an oscillator that is more stable if there is a specific frequency that is desired to be rejected (such as 50/60 Hz power line-cycle rejection).

Not all external clocks are equal.  For example, an external oscillator might range in stability from 25 to 200 ppm.  An external clock output from a micro can be more like the stability of the ADS124S08 internal clock.  So it is possible to use an external clock that is better or maybe worse than the internal clock.

There may be some other reasons to use an external clock such as synchronizing multiple ADS124S08 devices.  Another reason may be to change the data rates or filter notches.  The data rate and filter notches are directly proportional to the clock frequency.

However, as I said at the beginning, most of the time the internal clock will work just fine in the application.

Best regards,

Bob B

Hi Bob,

You mentioned that＂Due to this variation there may be a need to have an oscillator that is more stable if there is a specific frequency that is desired to be rejected (such as 50/60 Hz power line-cycle rejection)＂. Do you mean that if we want to reject 50/60 Hz power line-cycle, a more stable oscillator can generate a more accurate frequency ?

Hi Hao,

The oscillator frequency stability may not have any influence whatsoever on the rejection of 50/60 Hz rejection.  In the datasheet see section 9.3.1.1 on page 37.  Only specific data rates allow for this rejection.  If operating at 100sps or faster, the rejection will not take place no matter the clock stability.

View the filter plots starting on page 38 of the ADS124S08 datasheet.  The notches can deviate as much as +/-1.5% worst case for the internal oscillator.  In particular look at Figure 59 on page 38 as an example which shows a detailed plot of 20sps and the rejection for 50/60 Hz.  In all likelihood, even with the internal oscillator deviation, the rejection will still be good.

I would suggest downloading the ADS124S08 calculator spreadsheet found here:

http://www.ti.com/tool/ads124s08-excel-calc-tool

If you go to the digital filter tab you can adjust the master clock frequency by 1.5% to show the digital filter response at various frequencies and given data rates.

Bob B