Inexplicable OpAmp Behavior

[obuone]

Hello EEVblog folks.

I am working on a hobby project making programable DC load. In the design I am using MAX6070 4.096V Voltage reference and MAX5717 unbuffered DAC. Two MAX44246 dual- low offset opamps are used as current sense amp, voltage reference buffer, error amplifier and difference amplifier for remote voltage sensing. The voltage reference is buffered and fed into the REF input of the DAC. The DAC's output goes to the non-inverting output of the error amplifier though 10k resistor.

And here comes the issue. Everything works well if the input of the Opamp is below 1.1 V. However above that value the opamp starts to draw obscene amount of current from the input and the DAC output voltage crumbles. I know that the DAC is unbuffered with 2k ohm output impedance and can't drive an appreciable current, but this shouldn't matter. The MAX44246 has mere 1250pA input current by spec but i am measuring 100 uA of input current at DAC Full scale.

What the OPAMP is going on? Am I doing something fundamentally wrong? Do i have defective opamp or what. I have attached some images of the circuit and the measured input current and DAC output voltage as a function of code.

ps. I have already disconnected the DAC from the Opamp and checked its output. The DAC operates in spec.

Thank you!

[magic]

The analog part looks a little scary, are you sure it doesn't start to oscillate at some point?

[obuone]

Initially it oscillated. I had to remove C1 to stabilize it. That is one of the caps at the at the error amplifier. I also checked it with open loop by disconnecting the output of the amp from the gate of the transistor. I got the same result.

Even if it oscillated. How can that affect the input current of the amp?

[magic]

The reason I ask is because many opamps have differential clamping diodes between IN+/IN-. So one reason for excessive input current could be IN- being much lower than IN+ for some reason, like the whole circuit going out of whack due to oscillation.

(BTW, the IN- problem may be happening now after disconnection of the power transistor).

Connect it back and check if everything is alright with IN-.

[obuone]

Thank you. Wow i think you cracked this one wide open.

 The behavior looks kind of diode-y, but never occurred to me to check the other input as well.

Straight from the datasheet:
"The amplifier inputs are connected by internal back-to-back clamp diodes. In order to minimize noise in the input stage, current-limiting resistors are not used. If differential input voltages exceeding ±1V are applied, limit input current to 20mA."