Post Go back to editing

AD7768-4FMCZ evaluation board input Signal Test

HI! Greeting for you!

I am using AD7768-4FMCZ evaluation board, and configure communication through SPI of MCU, MCU_GPIO reads DOUT0 data,

Based on this board, I have a question. AINx+/- of AD7768-4 supports differential signal or pseudo-differential signal input. The pre-stage circuit of AINx+/- on the evaluation board is a buffer circuit composed of ADA4896-2.

If I want to test the  evaluation board through a single-ended signal,

for example, the signal source IN is 0~1VDC, A0+_AMP is connected to IN, and A0-_AMP is connected to 0V. Of course, the signal source GND will share the ground with the evaluation board.

Can it be applied like this? Thank you very much for your reply in your busy schedule .

  • Hi,

    We will look into this. I will contact the product owner and get back to you.

    Regards,

    JC

  • Hi  ,

    Apologies for getting back to you just now. Yes, you should be able to use a pseudo differential signal with the signal source connected to A0+_AMP, and A0-_AMP connected to 0V.

    Thanks,
    Janine

  • Hi  JEstayo,

    thank you very much for your reply.

    Just as the single ended connection method above,this scheme uses a microcontroller to obtain the ADC voltage value ang publish the message through the USART.
    Test method: Voltage source outputs single terminal voltage as Analog Input , and multimeter measurement results for evaluation.
    Test Result: through the data transfered By USART , we Can see the data is basically stable at 0.01mV.
    Voltage source   /   ADC serial port data /  A0+_AMP voltage Tested By multimeter measurement(Fluke 15B+)
    1V                           / 992mV                       /  998mV
    1.5V                       /1472mV                      /  1498mV
    2V                         / 1984mV                     /  1998mV
    2.5V                     /2464.1mV                   / 2497mV
    3V                     /2976mV                       / 2997mV
    The evaluation of AD7768-4 serial port output results using universal representation values shows an error of approximately 33mV, which is not ideal compared to a 24-bit ADC.
    Currently, AD7768-4 is configured as SPI, ECO mode (low power consumption), with a extraction rate of X1024, and a Buffer Enable on the analog input port.
    What methods can be used to improve the measurement output accuracy

    I hope to receive your reply,Thank you.
  • Hi  ,

    Let's first check if the error is seen before the signal reached the ADC.
    Is the third column the input to the amplifier? Can you add the measurements for the output of the amplifier, there should be a test point labelled A0+.

    Can you measure as well the voltage in the negative terminal? Please measure A0-_AMP and A0-. This should be 0V, correct?

    Thanks,
    Janine

  • Hi  JEstayo, Thank you for your patience answer.

    And Now, Wired The A0-_AMP To GND ,and  A0-_AMP just  floating for The Experienment.VCOM is Set to half of the AVDD,2.5V .

    (By Fluke 15B+)

    Measured A0-_AMP test point is 2.616V, A0-_AMP test point is 0V,A0+   test point  is 2.615V, and  A0+   test point  is 0.013V.

    Clealy, The error might be generated by the negative input amplifier follower,dueto the attribute of Rail to Rail.

    As it exits, should we change  the method to calibate a Single Signal By  EVAL  AD7768-4FMCZ  ?

  • Hi  ,

    I was also able to get the same 2.616V in the A0+ test point when left floating. The VCOM is 2.5V but there is 116mV drop in R236. If the A0-amp is grounded, there is 13mV difference between A0-AMP and A0- test point.

    Although, the ADC is capable of accepting 0V, the amplifier has headroom requirements. Apologies for not being clearer on that. 

    To adjust the input signal by VCM (2.5V), the configuration below can be followed. R140, R238, R144, R130, R233, and R127 can be changed to 10kohms and the input signal will shift by 2.5V.

    Thanks,
    Janine