Unexpected Highly Asymmetric Slew Rate on Op-Amp

Question

I recently set up a circuit with an ADA4870 op-amp for driving an ultrasonic Piezo element. I chose this specific op-amp because it was advertised as having a slew rate of up to 2500 V/us (symmetric on rise and fall) on a 10-40V power supply. To test it I set it up with +30V for positive power, GND for negative power, and two 2.4k resistors from out to V- and V- to GND (for a 2x gain with feedback). I then fed in an Arduino digital signal set to be high for 5us and then low for 10us. The measured input waveform is a sharp step function, but the output from the op-amp looks like this:

The rise is very sharp as expected, but the fall is extremely slow at roughly 1 V/us. Normally a slow rise/fall time would make me suspect parasitic capacitance somewhere, but I can't see why that would slow down the fall time but not the rise. Does anybody have an idea of why this might be showing up?

Answer 1

Basically using a single supply, you must ensure the input stays in the Vcm spec range which is >2V or <Vcc-2V. otherwise the input is not operating in close loop linear range. (E.g. Vcc/2 is a good way to bias inputs if possible by AC coupling.)

Other fine details,

Ringing is most likely from excessive probe ground lead and for <=35ns Tr or >=20 MHz you disable the DSO video filter and reduce ground leads to 1cm or so with special probing methods. (Repeated often on this site)

To minimize stray C load current , minimize differential ground capacitance and match impedance from cable to load which is the root L/C ratio.

Answer 2

Suggestions

• build the circuit atop a single_sided copper_clad FR-4, so you have a Ground Plane to work with

• bypass the VDD using 0.1uF from VDD to Ground using 0.1 inch lead. Locate this capacitor at the OPAMP VDD pin.

• isolate your opamp VDD from the power supply, by inserting 10 ohms in series. This dampened any wire_ringing.

• phase shifts on the Vin- may cause ringing or oscillation; consider using 1Kohm or 470 ohm resistors, instead of 2.4K

• Bias your Vin+ to VDD/2 using 1Kohm and 1Kohm voltage divider.

• inject the MCU rectangular wave using 0.1 uF to VIN+; the TAU of 0.1uF and 500 ohms is 50 microseconds, so you may see some droop on the waveform