Ground loop problem in op amp circuit

I am having problem with a few millivolts of ground loop in my single channel amplifier circuit when I ground the input (pin 1 on socket, - input of op amp). I am using INA121 instrumentation amplifier with LT1012 voltage follower to create virtual ground from power supply. Should the reference channel of the INA121 be isolated from the virtual ground, or could the cause be something else?

• Battery voltage? Commented Dec 10, 2019 at 8:27
• You may have an Induction loop current into 81M load, I=V/R e.g. 8mV/81MΩ = 100 pA . What frequency is the noise? 50Hz? amplitude? pure sine? Commented Dec 10, 2019 at 13:45
• Yes it is pure sine 60Hz output signal when input is grounded. thanks. Battery is 9V. Commented Dec 11, 2019 at 2:12

• There is no ground loop in your question since there are no current shown to cause noise. So we can assume this is induced line noise.

Common Design error.

• Mismatched source and Return impedances.
• The whole point of using an INA is to use the differential inputs and have balanced impedances from source to output.
• INA Impedance, Differential $$\10^{12}~Ω // 1 pF\$$

Your external to IC impedance on IN- = 81MΩ and IN+ =0 . Thus to see 1mV of noise all you need is $$\I=V/R=\dfrac{10^{-3}~V}{81 ~MΩ }= 12~ pA\$$

If you balance your sensor, wires and input impedance, if this is induced line noise, this should correct the problem.

• The Vbat/2 may need noise suppression cap.
• the cable may need to be twin ax shielded differential pair.
• you may need a Pi filter Balun = CM audio choke
• Thank you for your help. If I insert an identical high pass filter (with another C1 and R1) at the + input and connect to GND, should that do the trick? Is that what you meant by balancing the sensor and input impedance? Also, what did you mean by having balanced impedance from source to output; should I modify the output impedance of the amplifier? Commented Dec 11, 2019 at 8:23