• State Resolved
  • Locked Locked
  • Replies 1 reply
  • Subscribers 79 subscribers
  • Views 431 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.

General ADC questions

Nikhil Bapat
Intellectual 860 points


Hi,

While talking to a customer about developing inclinometer based on MS432P401M he brought up some interesting points which I could not answer right away (I am basically mechanical engineer)

1. When talking about ADCs the mentioned resolution is 14 bit but do we actually get that much resolution or do we lose the LSB? The datasheets for MCUs of some competing companies mention the  No Code Loss ADC resolution but TI does not do that.

2. What are the drawbacks of ADC oversampling? Can I oversample it just enough to make up for the loss of LSB?

  • 0
    TI__Guru 61235 points

    Nikhil,

    You are talking about missing codes, which means that there is a conversion result that the ADC simply does not produce.  It could happen anywhere in the codes.  For example, take a 14-bit ADC, which gives conversion results from 0x0000 to 0x3FFF.  Any of these could be a missing code, say 0x0400 doesn't appear if you do a very fine voltage sweep across full range.  0x0400 would be called a missing code.  Oversampling has nothing to do with correcting this.  Oversampling adds least-significant bits on to the results via what is essentially an averaging method of multiple conversions.  It will not really fix a missing code issues.

    Missing codes manifest from the DNL error spec of the ADC (differential non-linearity error).  DNL is the worst-case deviation over all codes of the distance from one code to the next code from the ideal (the ideal is 1 LSB).  So, if you have a minimum DNL GREATER than -1, you must have a missing code.  Similarly, if you have minimum DNL GREATER THAN -1, you are guaranteed no missing codes.  The weirdness comes in if you have minimum DNL = -1.  I suspect there is something about the way ADCs are tested such that getting DNL of +/-1 is a common spec.  Anyway, if you have DNL = -1, you may or may not have no missing codes.  What vendors do is state in the datasheet "No missing codes guaranteed".  This is because their test tolerances are such that they are able to determine there are no missing codes.  I'm not sure of the details on this.

    For the MSP432P401M, it seems TI took a slightly different approach to conveying the "No missing codes."  If you look at the datasheet slas826d, Table 5-28, it shows the DNL of -0.99 min, +1 max.  That -0.99 is a subtle way of saying no missing codes guaranteed!

    Regards,

    David

**Attention** This is a public forum