Post Go back to editing

ADA4950 single-ended input with DC-offset to differential output AD9094

Category: Hardware
Product Number: ADA4950

Dear support,

I have some difficulties understanding the implementation of the ADA4950 in a single-ended to differential setup when the single-ended input has a DC-offset. 

I have a single ended input signal (relatively low frequency) which can be anything between 0 and 1.8V. This needs to be converted to a differential signal to be connected to the AD9094. 

I need a gain of 1...this is the circuit which I thought would work (see screenshot) but after some reading online I am a bit doubtful as people mention that a DC-offset on the input could cause issues on the output common mode voltage(?).

The VOCM pins are connected to the common mode output pin of the AD9094 ADC. Would this circuit convert the 0 to 1.8V single-ended input to a 1.8V differential peak-peak output voltage with a common mode voltage which is setup by the ADC? Or am I doing something wrong?

Thank you!

Top Replies


  • Hi MikeHuisma,

    Thank you for your interest in ADA4950 and AD9094.

     

    We want the inputs centered at 0V (purely AC) to achieve 1.8V differential
    peak-peak output voltage centered at VCM .
     

    See sample below for 1.8Vpp input with VCM centered at half of +VS.

     


    In your case, the input is 1.8Vpp shifted by 0.9V (0 to 1.8V).
    This will cause the output common mode issue you mentioned.
    Notice that the outputs are not centered at VCM.

     


     

    With respect to AD9094, you can't directly use the VCM = 1.43V due to the output
    voltage swing of ADA4950 being limited to -VS+1.2V. You'll see some clippings at the output if
    you try to do this on ADA4950 with single supply. See sample simulation below:

     



    Please see this link for some insights about differential drivers for driving ADCs.
    "Rules of the Road" for High-Speed Differential ADC Drivers | Analog Devices


    -AD9094 have internal VCM and is connected to the analog input pins.
    Keep in mind that if you want to use VCM externally,
    you need to disable this internal buffer.
    This is discussed in the input common mode section of the datasheet.
     

    -For your application, having AC coupling for both inputs and
    outputs of the differential amplifier might be necessary.
    -For the common mode voltage of ADA4950, you can set this at midlevel of VS+


    You may also refer to this article for additional details related to ADC front-end designs with amplifier drive:
    Wideband A/D Converter Front-End Design Considerations II: Amplifier-or Transformer Drive for the ADC? | Analog Devices

    Best regards,
    Peevee
     

  • Thank you for your very detailed response Peevee it is very appreciated!

    This makes more sense now. So you would recommend AC coupling both inputs, AC coupling the outputs, connect VOCM to the midlevel of VS+ (thus midlevel of 3.3V) and use the internal internal VCM to bias the AC coupled output of the ADA4950?

    Would you be able to share the LTSPICE simulation model?

    Also as a final request, could I send another screenshot of my schematic today just to be sure that everything is well? (as mistakes cannot be happening in this design)

  • Hi MikeHuisma,

    If it is OK not to use differential amplifier in your application, you can check the analog input configuration from the AD9094-1000EBZ (same board used for AD9694EBZ) and consider if BALUN input will work for your requirement.
    You can download the design files for the board here: EVALUATING THE AD9094 ANALOG-TO-DIGITAL CONVERTER [Analog Devices Wiki]

    You can get the spice model of ADA4950-2 here: ADA4950-2 Datasheet and Product Info | Analog Devices
    For the simulation I showed, I used the ADA4950-1: ADA4950-1 Datasheet and Product Info | Analog Devices

    Feel free to share your schematic again Slight smile

    If you want to discuss including the details of your project offline, you can send me an email.

    Best regards,
    Peevee

  • Hi Peevee,

    Thank you again for your response! I have updated the schematic according to your insight and now have the following:

    I have AC-coupled the input and routed it to INB+ (because when routing it to INA+ the output signal seemed to be half the input signal and INB+ amplifies it by 2). By having a 5V power supply with a VOCM of half this power supply (2.5V) the swing is well within the range of the amplifier.

    Then I AC-couple the output of the amplifier and provide ADC+ and ADC- to the input of the AD9094. As you suggested I will use the internal common mode biasing of the ADC.

    What do you think of this?

  • Hi Peevee,

    My apologies for asking this much of you but would you be able to take a look at the schematic I posted here yesterday today and have a quick look on how I've now implemented it?

  • Hi MikeHuisma,


    I have AC-coupled the input and routed it to INB+ (because when routing it to INA+ the output signal seemed to be half the input signal and INB+ amplifies it by 2). By having a 5V power supply with a VOCM of half this power supply (2.5V) the swing is well within the range of the amplifier.

    The AC coupling seems okay. About the signal being half, this is due to the resistor combination R5 to R10 before the ADC+/- .

    If you try to probe the OUT+/- pins you should see proper Gain = 1 if you're connected to INA+

    Best regards,
    Peevee