Post Go back to editing

I am designing a circuit for driving a differential input ADC. As operational amplifier I use LTC 6254. My design works fine when simulating in LT Spice. However, when it comes to analyzing one of my outputs using a spectrum analyzer voltage level collaps

I simulated my circuit in LTSpice, as the spectrum analyzer I placed 50 ohm resistors. It works fine. 

  • Hi,

    A schematic would help us analyze what's wrong. What ADC are you using?


  • Hi Mustafa,

    For a simpler solution, use ADC drivers. To optimize the 65MSPS sampling rate of LTM9008-14, you can use either LTC6404-1 or AD8139. However, if you will only operate at 5 MHz (shown in your simulations), please use ADA4940 or LT1994 instead. It is much simpler to implement than using to op-amps to convert single-ended to differential signals.

    For the attenuation you observed in the output, do you have other components after the amplifier outputs? It seems to have created a voltage divider network with the analyzer. In 1 Mohm termination, it's not a big factor.


  • Hi,

    Let me be more clear.

    1) I have a single ended signal and I want to convert it to digital using LTM9008-14. Since  LTM9008-14  accepts differential input I use a driver circuit whose schematic is below.As operational amplifiers I used LTC6254. 50 ohm resistors represent Spectrum analyzer's input.

    When I simulate this circuit I get the result below, which is quite good enough for me.

    I simulate the same circuit, changing 50 ohm resistors with 1mega ohm resistors to simulate the input of the oscilloscope, nothing changes as seen below.

    2) I produced the circuit's pcb and then tested it by connecting the probe with one of the differential lines. When the oscilloscope's input impedance is set to 1 Megaohm everything seems fine as you can see below.

    3) However, when I set the input impedance as 50ohm voltage level collapses. 

    Such a result makes me think that the operational amplifier cannot supply the necessary current. This situation affects 2 things.

                a. Input impedance of the spectrum analyzer is 50 ohm. Since the voltage level collapses dramatically I see very weak signals. Harmonics cannot be seen, probably because they stay below noise floor. As a result I cannot analyze the quality of my circuit pproperly.

               b. I am not sure whether or not LTC6254 can supply the necessary current to drive LTM9008-14.

    Do you habe any idea about the situation?

    You can give advice for another operational amplifier with a really high GBP, or another single ended to differential converter.

    Thank you very much for your interest.

  • Yes my input signal is always below 5MHz.
    I have looked through the differential amplifiers you suggested  AD8139,    LTC6404-1,  ADA4940,  LT1994.
    One other requirement in my design is that the input source is a high impedance source and can only drive capacitive loads of ~15pF.
    Therefore I require the configuration I send in the above schematics. 
    1) Is it possible to drive the amplifiers you suggested in a similar way?
    I will check my circuit again about the 50 ohm load driving problem.
  • Hello Mustafa,

    Give me some time to work on the solution.


  • Thank you for confirming that. Allow me to look for the right amplifier for your application. Currently, LTC6254 is not fit to drive LTM9008-14. You should at least have an amplifier that has a THD of -72 dBc or lower at 5MHz, and 6254 could not have that. You could see it in your scope shots but not in the simulations because Harmonic Distortion is not modeled.

    Anyway, pleas be patient while I'm working on the solution.


  • Thanks for your interest. Moreover, I checked whether or  not there is something at the output of my circuit that causes a voltage divider effect for the 50 ohm load driving issue. I realized that the circuit is the same as the one I provided in schematics. 

  • Hi,

    You would at least need an amplifier with a THD of -72 dBc at fin= 5 MHz,  a much faster op amp would do so. Here are the possible choices: LTC6268/LTC6269, and ADA4857. Take note that ADA4857 runs at 5V and LTC6268/9 can run at 3.3 V. Also, reconsider your design. LTM9008-14 is specified to accept a differential input voltage of 2 Vpp. In your current schematic, you're running it at 4.8 Vp-p differential.


  • Hello,

    Forgive me If I am behaving too impatient. But deadline for my project approaches and I am a bit scared of not finding the the right amplifier. Is there anything you found so far? I hope there is something that can work for me. I am looking forward to hearing good news from you. 

    Best regards...

  • Hi,

    First of all thank you very much for your work and interest. I looked through the opamps you suggested. In datasheets their harmonic distortion performance seem quite pleasing. However before going with one of them I wanted to run a simulation using LT Spice with the same configuration I posted above. Here is the design.

    And you can find time domain result just below

    However, when I used the fft tool I see that 3rd harmonic is really high. You can find results below.

    I also simulated the same design with LTC6254. Here are the results

    Results seem better with LTC 6254. What these results may be about? What do you suggest me to suppress harmonics further?

    Moreover in the previous answer you said that my output is 4.8 Vp-p. But as you can see my differential output signal's amplitude is less than 2Vp-p.