• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Subscribers 46 subscribers
  • Views 674 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.

OpAmp really needed on ADS?

Bruno Saraiva
Guru 13040 points
Other Parts Discussed in Thread: ADS1222, REF5025, REF5050, OPA2335

Hi there,

I'm designing a circuit which uses an ADS1222 to perform 16-bit ADC conversions out of an inclinometer (SCA103T for what matters).

The circuitry includes two reference voltage IC's: REF5025 as the analog reference for the ADS, and REF5050 as the voltage input for the sensor.

I've seen a lot of examples using OpAmps in two different stages of such circuits:

- Between the sensor output and the ADC input;

- Between the voltage references output and the inputs of those two components.

As per the datasheets, the REF50xx is able of supplying 10amp. The sensor requirement is 4amp. Would I really need a buffer before it on that situation?

The power requirement for the ADS1222 is much lower than that, and again is seems that the REF should be easily able to supply the required juice.

Do we gain something by using the buffers? Or are we rather adding complexity, errors, offsets, etc? If the OpAmps are a benefical addition, is my initial choice OPA2335 a good option?

Thanks!

  • 0
    TI__Guru* 94335 points
    Bruno,


    My thought would be that it's probably ok to go without the buffer. I'll talk about what goes into the reasoning.

    The reference input to the ADC is really a charged capacitance sampling of the reference. This is described on page 10 of the datasheet. If you look at the equation, it's equivalent to about 500kOhms from a capacitance of 16pF sampling of about 16us (based on a clock frequency of 2MHz).

    To get a proper sampling of the reference, you need the reference to be able to drive the reference capacitance and to settle to the final value in the sample period of 16us. If it doesn't settle to the final value, it can be seen as a gain error. Imagine that the reference is 2.5V and only settles to about 2.49V each time (during the 16us sampling period) then you'd have a gain error of 0.4%.

    This can be aided with some extra capacitance on the reference input. If you add an external 0.1uF cap, then the 16pF referene cap is primarily filled from the 0.1uF cap as a ballast of charge that's available. It's discussed in the datasheet in the same section.

    To be honest, it's not too much to ask for the reference to drive the 16pF of capacitance, especially with some extra cap as holding up the charge. Going back to the reference, it looks like the reference settles in less than 16us (based on plots in the datasheet), and it's not a large current load to supply. I'd say buffering the reference isn't necessary. I know that I've used a REF1004 reference in the past without a buffer together with the ADS1222. Still, I'd use the REF5025, because it's lower in noise.


    Joseph Wu
  • 0 in reply to Joseph Wu
    Guru 13040 points
    Joseph,
    Thanks so much for the detailed reply. I will give this a try on the first design, the external cap was surely on the picture. I will just double check the calculations as per your suggestion and see how safe it is expected to be.
    I believe that the same apply for buffering the analog signals from the sensor into the ADC?
    Regards,
    Bruno
  • 0 in reply to Bruno Saraiva
    TI__Guru* 94335 points
    Bruno,


    Inputs behave in the same way. However, the ADS1222 has an optional input buffer that can be used to help drive the inputs and increase the input impedance. Note that you limit the input range of the device using the buffer (to about 50mV above the negative supply and 1.5V below the positive supply).


    Joseph Wu