• State TI Thinks Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 47 subscribers
  • Views 503 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.

DAC8820: Error budget of DAC

gary mella
Intellectual 440 points
Part Number: DAC8820

How to check the temp drift performances of DAC8820. This is required to evaluate error budget. What is accuracy difference of simulation vs real world(assuming conditions are identical).

Can I also validate the specs mentioned in datasheet  using TINA. Do you have any readymade testbench for the same

  • 0
    TI__Guru 53773 points

    Hi Gary,

    The datasheet specification for drift (really only full-scale error is specified) is a value based on testing on actual devices.  There are some additional guard-banding for these limits based on simulated process shift, test accuracy, and test repeatability.  So generally you see better temperature drift performance on real devices.  

    If you are looking evaluating the device across the whole temperature span, then you can estimate 2 ppm/°c × 125°C, or about 250 ppm or 0.025% error increase.

    The TINA model is useful for validating some of the specifications, like gain error, offset error, and glitch. Other aspects like drift and linearity are a bit harder to simulate in TINA, so I would base your budget on the PDS specification.  

    We do not have a testbench for this device, but the model is in TINA by default.

    Thanks,

    Paul

  • 0 in reply to Paul_Frost
    Intellectual 440 points

    Hi

    Can you tell me the procedure to evaluate gain and offset error over temperature in TINA. Would it be possible to measure INL and DNL? BTW how do I simulate the DAC as no digital models ( such as counters) exist. I do not wish to use pattern generator

  • 0 in reply to gary mella
    TI__Mastermind 46290 points

    Gary,

    The TINA model really doesn't give many options other than using the pattern generator or using fixed input voltages to look at performance near the end-points. Generally speaking the worst case error occurs at full-scale code. For a practical system you will need to include the output buffer in TINA. The models have options built-in to modify the ambient operating temperature.

    INL and DNL is not included in the model. For a complete error calculation including INL/DNL you would need to more rely on the datasheet.

    We are revamping our SPICE model approach to include more of these parameters as well as making them more friendly to code sweeps without the use of the pattern generator.

    Meanwhile, replies will be delayed for the rest of this week on account of the Christmas holiday in the US. Thanks for your patience.