I am using an AD 524 instrumentation amplifier (gain = 1) to detected potential drops across a resistor (200 Ω). As shown below, a current source (Keithley 6221) is used to send 1 mA current (DC or AC) through the resistor. Depending on whether one side of the resistor is grounded, the AD 524 output gives different values. I have checked that the potential drop across the resistor is accurate (V = IR = 1 mA*200Ω = 200 mV) using a multimeter, scope, and lock-in amplifier. It seems that the deviation of the op amp output (187 mV instead of 200 mV) occurs only when the current source is left floating. Does anyone have any ideas as to why this would be happening?