ADS1119: Conversion time

Hi Michiaki,

See my responses below.

Best regards,

Bob B

Michiaki Tanii said:

Part Number: ADS1119

Hello team,

 

Could you give us your advice on the conversion time for each mode.

If data rate is configured at 20 sps with single conversion mode, actual conversion time at 50.01ms is between START command latched and DRDY is getting low ? [Bob] The short answer is yes.  In single-shot mode once the START command is latched it will take 50.01ms and the DRDY transitioning from high to low signals that the conversion result is now ready to be read from the ADS1119.  However, getting an exact timing is a little difficult as this will depend on the I2C SCL speed and when the START/SYNC command is actually latched (ACK).  Also, the internal oscillator can vary some from device to device and as much as +/-2% (+/- 1% typical).  A more accurate timing is given in section 8.3.6 of the ADS1119 datasheet as number of tCLK cycles.  So the Actual Conversion Time as shown in Table 4 can vary by as much as +/- 2%.

 

And the conversion data is average during the above time ? [Bob] The digital filter is an FIR filter and not an averaging filter.  The ADS1119 is an oversampling device where the input is sampled at 256kHz rate.  The process of using a Delta-Sigma modulator will push quantization noise into the higher frequencies and the digital filter acts as a low-pass filter.  The FIR characteristics at 20sps also include a notch filter for 50/60 Hz.

 

If customer configure continuous mode at 20 sps, 49.99ms is indicated moving average ? [Bob] There is not a moving average filter.  Each conversion stands on its own  and is single-cycle settling.  If you want a moving average filter, you would need to apply this externally in micro code by averaging the individual conversion results.  I should also add that if a step change in the input voltage takes place during the conversion period, then the conversion result will appear to average the result between the initial sampling period and the end of the sampling period, but the response is not really 'averaging' due to the FIR filtering method.

 

Thank you and best regards,

Michiaki