• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 45 subscribers
  • Views 812 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.

AMC1200 plus ADC Combo

Oleksiy Pazyuk
Expert 2585 points

Other Parts Discussed in Thread: AMC1200, AMC1204

Hi.

I would like to estimate which maximum resolution is required from ADC to obtain maximum  accuracy at the output.

In the spec for AMC1200 is mentioned that output noise of the device equal to the 3.1mVrms.  Taking into account that the maximum voltage at the output of the device is ±2В  this give us that to obtain ±1LSB or better accuracy at the output of the ADC 10-bit resolution will be enough.

Am I correct?

  • 0
    TI__Genius 9975 points

    Hello Oleksiy Pazyuk,

    Assuming the following conditions apply:

    -In the absence of external noise in the system,

    - Using an ADC with Vref = 4 V (or whose full scale range is referred to 4 V),

    - Using an ADC whose bandwidth is matched to that of the AMC1200,

    - Using an ADC capable of delivering 10 effective bits.

    Then, the AMC1200 plus ADC will indeed provide an output where the LSB does not change due to the output noise of the AMC1200.

    Best regards,

    Jose

     

     

  • 0 in reply to Jose Duenas
    Expert 2585 points
    Hello Jose,

    Am I correct that we couldn't obtain resolution better than 10-bit with 1LSB accuracy by using AMC1200 + ADC?
  • 0 in reply to Oleksiy Pazyuk
    TI__Genius 9975 points

    Hello Oleksiy Pazyuk,

    To clarify terminology a little bit: the accuracy of a system depends on whether the system can be calibrated for gain and offset error, whether the non-linearities can be corrected and whether calibrations can be performed “on the fly” as the system temperature changes.

    Assuming your questions is still related more to the feasibility to deal with the AMC1200 output noise (and not with dealing with the device’s offtset, gain error and non-linearity), then the answer is: one can obtain better than 10-bit resolution provided one samples and process the output of the AMC1200 properly.

    For example, the following blog post talks about the kind of post processing that allows higher resolution: https://e2e.ti.com/blogs_/b/precisionhub/archive/2013/08/23/how-to-use-thermal-noise-to-your-advantage

    Evidently, the trade off is that data throughput is lower when sampling at a high rate and post-filtering.

    Assuming you do not use any of the post-processing, then the resolution limit is indeed 10.33 bits. Therefore a converter offering 10 effective bits (not to be confused with codeword width) should work fine.

    Hope this helps.

    Best regards,

    Jose

  • 0 in reply to Jose Duenas
    Expert 2585 points
    Hi, Jose,

    Thanks for your answer regarding AMC1200. But from general point of view do you have some figures from application about how far we could improve resolution by using post processing?
  • 0 in reply to Oleksiy Pazyuk
    TI__Genius 9975 points

    Hello Oleksiy Pazyuk,

    No, I do not have any measurement results or additional application information on improving the resolution of an AMC1200-based system by using post processing. Typically, systems that require better performance than that offered by the AMC1200 switch to an isolated delta-sigma modulator approach (AMC1304, AMC1305, etc.)

    Best regards,

    Jose

  • 0 in reply to Jose Duenas
    Expert 2585 points

    I guess if we talk about  improving resolution an there is no requirements for reinforce isolation it is better to use AMC1204. As I see from datasheet AMC1204 has slightly better SNR and as result ENOB for the same OSR.