NRFX_SAADC_CONFIG_OVERSAMPLE relationship with ADC

Hi!

I am currently using the ADC to take two thermistor values from 2 AIN pins. Currently we are using the 10 bit ADC but I am looking to achieve 14 bits with oversampling in order to reach a threshold of precision. Currently we take the values from the ADC one at a time. After taking the values from the ADC, some maths occurs before a final single temperature value is produced. We require this final result once per second. 

While looking into updating the ADC to 14 bits I noticed that I will need to define an oversampling factor NRFX_SAADC_CONFIG_OVERSAMPLE. I was just wondering if someone can clarify the relationship between this factor and the final ADC resolution?

If I have the largest oversample factor, i.e. 256, does this mean I am now getting 14 bit resolution?

Thank you

Parents
  • Hi there,

    The ADC on the nRF52 has a effective number of bits of 10. Oversampling will take multiple samples on each channel, which will decrease the error and therefore increase the effective number of bits. You can still use a resolution of 14 bits without oversampling, but then the 4 LSB would be considered noise and wouldn't be any accurate. Making the effective number of bits 10.

    If you're using a resolution of 14 bits then I would suggest using a oversampling rate of at least 16x. 

    regards

    Jared 

Reply
  • Hi there,

    The ADC on the nRF52 has a effective number of bits of 10. Oversampling will take multiple samples on each channel, which will decrease the error and therefore increase the effective number of bits. You can still use a resolution of 14 bits without oversampling, but then the 4 LSB would be considered noise and wouldn't be any accurate. Making the effective number of bits 10.

    If you're using a resolution of 14 bits then I would suggest using a oversampling rate of at least 16x. 

    regards

    Jared 

Children
No Data
Related