-1
\$\begingroup\$

Why do we need to test the features of an ADC like INL, DNL, offset, and error? And in addition AC specs like SNR, SINAD, THD, etc. These parameters are in the datasheet prepared by the manufacturer.

Aren't the values ​​we get from these techniques valid only in those conditions? For example, we get different values ​​at different temperatures. Why do we need ADC testing techniques, and is the accuracy of the result valid only for those conditions?

Note: I know these techniques. Please don't explain these techniques, just why we need these tests, as the manufacturer already gives these values ​​in the datasheet.

\$\endgroup\$
9
  • 1
    \$\begingroup\$ Can you explain what you mean? Is the question about why you need to test the parameters yourself, if manufacturer has testes them for you? In which case, you don't need to test if the parameters are already OK for you. \$\endgroup\$
    – Justme
    Commented Feb 2, 2023 at 18:03
  • 1
    \$\begingroup\$ I don't understand the question. You or someone is not happy with the specs notated in the ADC's datasheet and is telling you to test them? We need ADC testing techniques because the ADC designers would like to validate the performance they predicted via simulations in real silicon. \$\endgroup\$
    – Designalog
    Commented Feb 2, 2023 at 18:31
  • \$\begingroup\$ @Justme yes, of course the manufacturer gives the parameters. If I'm not mistaken, ADC testing techniques are being done better results. We are trying to find the most optimal value of the parameters given as maximum and minimum on datasheet. But I can't understand this. Doesn't getting different results in each different environment change my result? In other words, the AC and DC parameter coefficients obtained at 20 degrees will not be valid for 40 degrees. Then what is the purpose of the test? I hope I was able to explain. \$\endgroup\$
    – bb0667
    Commented Feb 2, 2023 at 20:15
  • \$\begingroup\$ @ErnestoG No, I'm not talking about any testing at the silicon level. I don't think the manufacturer's values are wrong. I just try to get better results by finding the most optimal deviation or tolerance within the given ranges for the parameters. But I think my coefficients or optimal values will change in any condition change. Then I can't understand the purpose of testing. I think the correct values will never be reached. So I'm trying to understand under what conditions these techniques should be used or which result has full validity. \$\endgroup\$
    – bb0667
    Commented Feb 2, 2023 at 20:19
  • \$\begingroup\$ @bb0667 so, for a given typical spec within a given set of conditions (vdd, temp, etc), you'd like to find whether you still get this spec number for a different set of conditions (that match better your use case), but you end up getting worse spec numbers than the datasheet? Is that the problem? \$\endgroup\$
    – Designalog
    Commented Feb 2, 2023 at 20:46

1 Answer 1

Reset to default
1
\$\begingroup\$

Why do we need to test the features of the ADC like INL,DNL, offset and error? In addition AC specs like SNR,SINAD,THD,etc.

We don't need to do any of these tests unless we are trying to claim a higher performance (for our finished equipment) than what would be generally implied from the data sheets.

Bear also in mind that external reference voltages may be used and these will degrade data sheet specifications and testing is sometimes the only way.

Also, with multiple channels, we might want ADCs with a similar performance and so, selecting several from a large batch may be a way to achieve this.

is the accuracy of the result valid only for those conditions?

Yes it is but, as with any test like this, it can be designed to replicate as much as it can of the real-world situation that the ADC is subject to.

\$\endgroup\$
4
  • \$\begingroup\$ First of all, thank you very much for the reply. I had overlooked that external power could affect ADC performance. I added one more to my reasons. (Although creating a high-accuracy reference voltage solves this problem) \$\endgroup\$
    – bb0667
    Commented Feb 2, 2023 at 20:31
  • \$\begingroup\$ My goal is to choose the most optimal value in the range given in the datasheet. I can make this value dynamic by calculating it over and over under all conditions, but this way can be very cumbersome. I am trying to find the most optimal value by performing tests under different possible conditions. But I am skeptical about what kind of reference I should act on the test results. I am aware that the outcome will change when circumstances change. But I'm trying to find the optimal result. \$\endgroup\$
    – bb0667
    Commented Feb 2, 2023 at 20:32
  • \$\begingroup\$ If we think that conditions will change, how can I follow a path to choose the most optimal value? Or am I talking about a confusing situation here? \$\endgroup\$
    – bb0667
    Commented Feb 2, 2023 at 20:32
  • \$\begingroup\$ I'm just answering the original question. If your goal is to get an improved performance, choose a very good ADC (or DAC) and test it multiple times under different temperature conditions and use the drift figures in the data sheets. If this isn't good enough then test many more times to determine the zero offset and gain drifts. Pay particular important to picking a reference voltage that is very, very good. \$\endgroup\$
    – Andy aka
    Commented Feb 2, 2023 at 22:09

Not the answer you're looking for? Browse other questions tagged or ask your own question.