Post Go back to editing

AD9644 - driving DC-coupled signle ended analong signal


I have AD9644 evaluation board, which I need to use with DC-coupled analog signal.

At this moment I am not interested in ADC performance, so e.g. increased noise floor is not a problem.

The evaluation board is built with transformers/capacitors, blocking any DC component in analog signal.

I'd like to connect my analog DC-coupled signal straight to the ADC analog input , as single ended signal. There are two options:

a) Utilizing the AD8376 (U401) of the evaluation board. This requires changing C402/C403/C404/C406 => 0R, and activating active path (R303 => DNI, R309 => 0R etc; C305 => DNI, R311 => 0R etc), and populating anti-alias filters (between AD8376 and AD9644 (with DC capable topology).

And then feeding the single-ended signal to J301 & J303, and grounding J302/J304.

b) Feeding the single-ended signal straight to AD9644 VIN_POS_A/B and fixing the voltage of VIN_NEG_A/B.

This can be done by:

removing T302/T307

C305/C306/C311/C312 => 0R


The datasheet of AD8376 isn't precise enough in input => output transfer function. E.g. it doesn't tell how single-ended input to differential output transformation output common mode voltage is generated. A realistic functional SPICE model would be nice too.

Please provide some help so that I can decide which path to take, and what problems I might face.

BR, -Topi

  • Hi,

    The AD9644 is not designed to be driven from a single ended source, so option b is not recommended.  Option a above would require the AD8376 to provide a dc coupled single ended to to differential conversion at the input, and a differential output signal with 1.75Vp-p swing and 0.6V common mode.  I moved this post to the amplifier section of Engineering Zone where someone should be able to help you understand the options. 

    Best Regards,


  • Hello Topi,

    Here is the paragraph in the datasheet you should be focusing on. 

    Datasheet page 13. 

    "The outputs of the AD8376 are open collectors that need to be pulled up to the positive supply with 1 μH RF chokes. The differential outputs are biased to the positive supply and require ac-coupling capacitors, preferably 0.1 μF. Similarly, the input pins are at bias voltages of about 2 V above ground and should be ac-coupled as well. The ac-coupling capacitors and the RF chokes are the principle limitations for operation at low frequencies. "

    Since the AD8376 outputs are biased to positive supply and given your application is DC coupled, you would remove the ac couplinge capacitors and use a voltage divider to step down the voltage before it is inputed to the ADC.  This will result in substantial signal loss.  Additionally, you would replace the 1uH choke inductors with resistors to get to low frequencies.   

    What is your operating frequency?  The AD8376 may be on the ADC evaluation board but it may not be the best option for your DC coupled application.  I am not sure of the performance spec you're hoping to achieve and the operating frequency but the ADL5561/62/65 and ADL5566 are more DC coupled friendly amplifiers and they have very good linearity performance.  If your operating frequency is from DC to a couple of hundred MHz, I can also recommend other amplifiers from another group within Analog Devices.


  • Hi Qui,


    I did understand the open-collector output mode. But giving input voltages Vin+ and Vin- what shall be the output currents of Vout+ and Vout- ??  The best I could get from the datasheet is that positive and negative inputs are amplified separately to positive output and negative output, with very well matched gain. But then the "single-ended => differential" function isn't what I should be looking for.

    If the Vout+ is somehow dependent on Vin- (and vice versa), then this is not clearly stated on the datasheet (or I didn't find it). And from your answer I sitll cannot calculate (without trying with the actual device) what shall the output common mode voltage be in the case I replace output bias inductors with bias resistors (and remove all DC-blocking capacitors).

    2. My application requirement is down to DC, but other ADC performance is not critical at this moment. I just want to avoid doing another single-ended => differential circuit, if possible at all. The AD8376 just happens to be in the convenient place and I'd like to use it.

    If it is definitely impossible, then I shall build operational amplifier based adapter.

    BR, -Topi

  • Hello Topi,

    1. The open collector output mode implies that the outputs need to be pulled high to supply.  The AD8376 is a 5V device.  Therefore, to answer your question, the output common mode voltage will be 5V in the case where you replace the output bias inductors with bias resistors and remove all DC-blocking capacitors.  The AD9644 requires a common mode voltage of 0.9 V.  As a result of this, you will need to do something in between the AD8376 and the AD9644 to go from 5V common mode to 0.9V common mode while preserving your DC signal. 

    Attached are the S-parameter files for your simulation


  • Qui,

    This goes into academic now (I opted to use another full-differential operational amplifier, which has a suitable evaluation board available).

    I understood that AD8376 is a transconductance amplifier, thus there is Vin => Iout gain. And the open collector output constraints the output current to 0 <= Iout <= Imax; negative output current is not possible.

    If the open-collector transconductance output (i.e. current sink) is connected to biasing inductor; AND the signal is free from DC component; THEN the DC-component of the output VOLTAGE shall be biasing point (5 V in this case).

    But this is a feature of the biasing inductor only. It is not possible to build an open collector amplifier which works differently, when biased with an inductor.

    But when biased with resistor, the output common mode depends on the output current at Vin_diff=0. And also how the two paths interact together. But definitely the output common mode, with resistor biasing, can not be the biasing voltage (this could be possible only if the output current is zero, when the Vin_diff=0, which wouldn't work as there weren't any room to decrease output current below zero (as the output is open collector)).

    But anyway, I believe AD8376 is a great part when properly used.

    BR, -Topi

  • This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

    Thank you,
    EZ Admin