• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 48 subscribers
  • Views 970 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Delta-sigma ADC

jbali khaled
Prodigy 45 points

why if we increase in delta-sigma ADC 's gain we lose ADC'S resolution ?????

  • 0
    TI__Mastermind 38745 points
    Hello Jbali,

    The resolution will always be the same, but the dynamic range will change based on the gain. Assuming the converter's reference voltage stays constant, the percentage of the full-scale range that is made up of noise will increase linearly with gain since the noise floor is also amplified by the PGA. I'd like to point out that this is not necessarily a limitation unique to delta-sigma ADCs, since any linear amplification stage will also amplify whatever noise appears at the input. Is this the phenomenon to which you were referring?

    Regards,
    Brian Pisani
  • 0 in reply to Brian Pisani
    Prodigy 45 points
    so if we increase the PGA 's gain the noise too but the resolution will be the same??
  • 0 in reply to jbali khaled
    TI__Guru** 100885 points
    Jbali,


    Let me explain it this way:

    Generally the resolution is a function of the amount of noise in the ADC readout. As you increase the gain of the ADC, the same noise for a smaller input range give you less resolution.

    As an example, lets say that you have a 24 bit converter that has an input range +/-2.048V. Let's also say that the last four bits are noise.

    This would give a least significant bit size of 2.048*2/2^24 = 0.244uV. However since the last four bits are noise, the noise is 3.91uV of noise. This can be derived from either 20 bits of resolution coming from 2.048*2/2^20 or by taking the LSB size and multiplying by 16 (from 2^4).

    If you apply gain, you change the input range, but you don't necessariy improve your noise. If you have a gain of 16, then the input range becomes +/-2.048/16 or +/-0.128V. If you have the same amount of noise, this becomes 3.91uV of noise on an input range of 0.256V.

    This is a noise of about 1 part in 65000. This becomes the equivalent of 16 bits of or resolution (coming from log(0.256/.00000391)/log(2)). In the previous example, it was about 1 part in 1000000, which was about 20 bits of resolution.


    Joseph Wu