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.

ADS8910B: ADC input driver from single ended to differential ADC positive input only and negative input grounded.

Part Number: ADS8910B
Other Parts Discussed in Thread: OPA2625, , OPA625, THS4551, ADS8329, ADS9110, OPA2328, OPA862, THP210, OPA320, THS4561, LM7705

Hi 

A quick question on the ADC input driver type for ADS8910B. In datasheet page 59, the typical application with differential input signal and opa2625 drivers. Unfortunately, my signal inputs are single ended (0V-4V roughly). In the design, I'm thinking to use one opa625 or similar unity gain amp to drive the 2.2 ohm resistor and 10nF cap to ADC positive input and simply ground the negative input to signal ground, instead of using the balanced two opa2625 input drivers.

Other than utilizing half of the ADC dynamic range (0V to Vref) and double LSB sizes, comparing with a single ended input ADC, is there any other concerns that could potentially impacting this ADS8910B performance?

I did see that on page 61, single ended can be applied to THS4551 inputs, but it seems like an over kill, the input signal are <300kHz and no need to generate Vocm if only drive the positive input.

Another reason to pick ADS8910B is the option to upgrade to ADS9110 with some layout change. I did not find other single ended input ADC with extended family (only ADS8329 with no upgrade options).

Any comments or suggestions are greatly appreciated.

Regards,

FZ

  • Hello Fubin,

    First, welcome to the TI E2E community!

    The ADS8910B requires a fully differential input drive; each input should be equal in magnitude and 180deg out of phase.  If the inputs are driven pseudo-diff, with the negative input connected to a fixed DC voltage (or ground), the ADC performance will degrade quite a bit.

    This requirement is specified in the input common mode requirements section of the datasheet:

    As you pointed out, you can use a fully-differential amplifier to convert from single-ended to differential, or you can use a 2x OpAmp configuration as well.  Below are a few analog circuits that discuss both of these approaches.

    https://www.ti.com/lit/an/sbaa265/sbaa265.pdf

    https://www.ti.com/lit/an/sbaa264a/sbaa264a.pdf

    For the dual opamp approach, the OPA862 implements this configuration including the required matching resistors.  You can also implement the suggested dual opamp circuit using the OPA2328 dual amplifier.

    Regarding the amplifier bandwidth, the amplifier must drive the switched-input capacitor of the ADC to full settling within the acquisition window, which is 300nsec when operating at 1MSPS.  This requires a minimum bandwidth of 20MHz for 18b settling, or 40MHz for the inverting amplifier stage.  Using the 130MHz THS4551 is not excessive in this case.

    If interested in more details on this subject, please refer to the TI Precision Labs - ADCs series more more detail.

    Regards,
    Keith Nicholas
    Precision ADC Applications

  • Hi Keith,

    Thank you for the detail clarification, I missed the fixed Vcm. I think OPA862 is a good option for convertion to differential drive.

    Also, any other FDA options? Thanks

  • Hello Fubin,

    Take a look at THP210.  This is a precision FDA, and has been fully tested driving the ADS8910B up to 800ksps.  If you need higher speed (especially if moving to ADS9110), then the previously mentioned THS4551 is a good choice.

    Regards,
    Keith

  • Thank you again . I will check out THP210. Two more questions:

    1) I am using TINA to simulate THS4551 device as the driver. This option (sbaa264a.pdf) should be better than OPA862 in terms of performance, right?

    2) I am having some issues with TINA simulating THS4551 while driving 10nF load, not sure if it's the limitation of the model. I downloaded the example (sbac188 from sbaa264a.pdf). Do you have a similar example for THS4551 and ADS8910B? If not, I can modify sbac188 and see. The input is 0V-4V. Vcc=5V, Vref=4.096V, Input signal can be shifted up to ~0.125V from gnd in previous stage.

    Thanks

  • Hello Fubin,

    1.  The OPA862 has very good noise and DC offset specifications.  However, the input common mode does not support voltages down to the negative supply rail, which will require a negative supply voltage to support a 0V to 4V input.  The THS4551 will not have this limitation, and a rail-to-rail input amplifier, such as OPA320 can be used that will support inputs down to the negative supply rail.

    2.  The standard circuit will not be stable with 10nF of load.  Please see the attached suggested circuit using the THS4551 and ADS8910B.

    Regards,
    Keith

    THS4551_ADS8910B.TSC

  • Thank you, Keith. The circuit showed great results. I noticed the -300mV on the negative rail.

    There is a -5V negative rail in the design. In this case, would it be a better solution using THS4561 (BW, Offset not as good), or OPA862 and THP210?, Using -5V replacing the -300mV supply.

    If using THS4561, the suggested circuit for THS4551 should be still good, right?

  • Hello Fubin,

    The -0.3V supply is only needed to ensure DC accuracy for input voltages less than 10mV, where the output of both the OPA320 and THS4551 must go all the way down to 0V.  If you can give up some accuracy for very low voltages near 0V, then you can eliminate the -0.3V supply and connect directly to ground.

    If you need full accuracy at 0V input signal, then some type of negative supply voltage is needed.  Yes, you can use the THS4561 and operate from +/-5V supplies in this case.  The same circuit values should work well for either amplifier.

    OPA862 can also be used, but in order to meet the common mode requirements and support 0V-4V input range, you would need to operate from a negative supply of -1V or less (-5V would work) and the positive supply voltage would need to be at least +5.3V.

    Regards,
    Keith

  • Thank you so much again for the detail answers and great suggestion. I will use THS4551 as the default and THS4561 as optional, or even add LM7705 if the board space is not an issue. No more questions.