I am trying to make 10Mohm of input impedance into a ADC (CS5530). I am planing to use differential input buffer to ADC so I am trying to use TI's document as reference. I tried the one differential in differential out single supply and dual supply. But I am not getting decent signal out of buffer (I am using sinus signal).

Basically I am trying them without ADC putting signal in and look them out with scope. I can't see sine signal out.

I am confused about GND levels I mean I should buffer GND level of signal generator and if it is big from other GND then opamp may broken or may stuck on rail limit. Circuit must consider voltage difference between 2 inputs.

Any suggestion is appreciated about getting 10M input impedance.

Schematic is as follows:

dif in dif out single supply

dif in dif ou dual supply

ADC: planing to use CS5530
opamp : LT1013
circuit built on breadboard


  • 1
    \$\begingroup\$ I'd love to understand which of the two schematic you are using, to see a photo of your breadboard and to receive some explanation on how you are feeding and reading the signal. \$\endgroup\$ Jul 15, 2014 at 13:48
  • \$\begingroup\$ @VladimirCravero I editted question it is my breadboard. I tried both schematics. the picture is from dual supply design. right side the bnc is signal generator input output is two wires that is connected to scope \$\endgroup\$
    – erkan
    Jul 16, 2014 at 8:29

1 Answer 1


In the dual supply schematic, there's no reference for the input voltage. Just as a for-instance, if there's no reference to ground on the input side, then you could be sending in 1000V+1*sin(x) which would result in the op-amps being railed for the entire time. Once it's got a good reference, it will be just sin(x) and shouldn't be railed so long as the resistors are chosen well.

The single supply also has this issue. The input is AC coupled which is fine, but the only place current to go is into or out of the output of the op-amp. If you've already railed it, then it won't be able to move from the rail.

You'll want both circuits to have some kind of path to ground so that when the signal is 0V input, it will level out to 0V output.

RT should be split into two resistors with GND sent to the middle of it. CM on the single supply should likely be connected to ground probably through a resistor.

  • \$\begingroup\$ Thank you for your answer. Single supply circuit i already connected CM to ground. I will try splitting RT and make voltage divider to GND see if it works. But still if i use GNd then it is like it wont be diferential. I break out some multimeters and i see com is connected to circuit ground so if you connect voltage to com then it become GND level for multimeter. (or i've seen that wrong). Maybe i should change it like that. I am concerned MCU and GSM module will effect from these GND changes \$\endgroup\$
    – erkan
    Jul 16, 2014 at 8:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.