• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Answers 2 answers
  • Subscribers 29 subscribers
  • Views 371 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

OPA1602: Understanding the claw curve in this datasheet

Michael Steffes
Guru 47485 points
Part Number: OPA1602

You may not be able to answer this from a 2011 introduction, but I will ask anyway for an upcoming article on available "linear" output current in audio amps - this is a really nice one from the OPA1602 Figure 29. 

This curve doesn't give you linear output current very well, but it very useful in any case. What I am saying right now is this is a typical process case designer full PDK simulation and that probably needs to guardband the corners for over process case by at least 20% down. Sound correct? 

  • 0
    TI__Mastermind 18405 points

    Michael,

    Generally when I collected these curves for precision op amps over the last 8 years while in my validation role I would collect a minimum of 30 units worth of data and then the curves you see are the average of the units for each Vout vs Iout line above. In general this was from one lot or sometimes more when they were available. As the rest of the family was released, different packages, singles, duals, quads, and lots we check the results against the original. If we saw any differences they were called out in the PDS. Hopefully this answers the question but this is how that curve came to be. 

  • 0 in reply to Chris Featherstone
    Guru 47485 points

    I am assuming you bounced these against designer sims, but thanks that makes sense. I will adjust my wording - 

  • 0 in reply to Chris Featherstone
    Guru 47485 points

    Well Chris, this look reasonable? 

    This curve is generated averaging a limited number of devices from a single process lot used for characterization. In theory, you should still guardband the breakpoints down by some amount for process variation – perhaps by 20%. 

  • 0 in reply to Michael Steffes
    TI__Mastermind 18405 points

    Michael,

    Yeah for the typical curves we take the average and we don't specify a min/max for claw curve. That being said however I like to think of the claw curve representing a few points of operation of the output stage which are tested in production and guard banded. These are voltage output swing to the rail and short circuit current Isc. Although Isc is generally specified as a typical spec in the table it is checked for abnormalities in production along with the output swing at different loads (2k, 10k for example). This is effectively checking the two extremes of the claw curve and a few points along the claw curve with the resistive loads. In the case of this part I see we have Rl = 2k for the output swing load condition.

    In addition we test Aol in production as well which exercises the output stage at a few different loads or output currents. These are most often specified with min/max and guard banded as well. 

  • 0 in reply to Chris Featherstone
    Guru 47485 points

    Oh yes I am aware of those Aol at swing limits test, in this part it is basically a no output current stress condition, way left on the claw curve I suppose, 

    I am trying to arrive at a min linear output current, not really in this test for this part,but thanks,