OPA2991: Poor linearity

Hi Robert,

The most likely issue is the OPA2991 incorporates a slew-boost circuit that takes a larger signal to engage. Essentially, the OPA2991 device offers the higher slew rate 21 V/µs when applying larger input differential voltages.

See figure below.

The OPA991 has slew boost circuitry to increase the slew rate of the device up to 21V/us; however, there must be a large signal applied to the inputs to fully turn on the slew boost circuitry. The natural slew rate of the device is significantly less than 21V/us which is why you are seeing the distorted output signal. A 150mVp step signal will start to engage the slew boost circuitry, but will not be large enough to obtain the full slew rate spec.  Please note that the slew rate measurement plot below is measured when applying a step signal (not a sinusoidal). See plot below.

I recommend taking a look at the application note below, which discusses this topic in some detail:

• Ramping Up on Slew Rate (ti.com)

Also, the TI Precision Lab Videos on Slew Rate for more information. Specifically Video #3 discusses small signal vs. large signal behavior.

The simplified TINA-TI/SPICE models are linear models that not accurately model the slew boost response at the smaller input signals.

The TLV9362 (10-MHz) is a good alternative for the OPA2991 when the slew rate becomes an issue in the application. The device provides a much higher slew rate at smaller input signals.

Thank you and Kind Regards,

Luis

Thanks Luis, your explenation looks like it explains the issues we have been seeing. I will search out an alternative opamp and will evaluate your suggested TLV9362 alternative for our application.