OPA657 / OPA 657: problems with simple testing

Aaron,

Two problems with your configuration:

1) High speed amplifiers shouldn't be used on breadboards; the parasitic capacitance you'll be getting from the breadboard will easily cause your amplifier to oscillate, as shown; it's output is swinging from +-5V, the rails.

2) The OPA657 is not a unity gain stable amplifier.  It's a decompensated amplifier, best used at a gain above 10V/V.  This is likely why your amplifier is showing oscillation in a buffer configuration.  Minimum gain configuration we show is around 7V/V, but the amplifier is marginal at best;  bandwidth charts show sigificant peaking at this low of a gain.

Put the part on a PCB with a proper high speed layout in a gain configuration of at least 10V/V and you shouldn't have any problems with it oscillating.

When an OPA 657 oscillates from +/- 5V (its rails), is it essentially broken then?

I implemented your suggestions:

1) increased the non-inverting gain to ~50
2) mounted the OPA 657 onto a PCB board via a 8-SOIC-to-DIP adaptor

It did not oscillate for about 2 minutes (gave a DC output, but it was the wrong value). After that, it started oscillating like it did on the breadboard, and is always oscillating now...

Aaron,

If a part oscillates, it doesn't necessarily mean that you have damaged the part, however when op-amps oscillate they tend to self-heat and may end up damaging themselves.

Now that you've fixed the gain issue, here's a few other things you can try:

1) Replacing the part with a new one (if you haven't already)

2) Replacing Rf in the circuit to the recommended value, 453Ohms.  Rf=20k might be causing issues, as this is much higher than the values tested in the datasheet.

3) Try placing a small resistor (50 Ohms or so) in series with the probe output to shield it from the 20pF of capacitance that the oscilloscope adds to the load.

If none of that works, I'd think that your layout or component selection may be causing issues.  I'd recommend getting one of the EVM's with a known good high speed layout to do testing on, and then work your way out from there.  You can order one of the SOIC EVM's for free and populate it yourself here:

https://estore.ti.com/DEM-OPA-SO-1A-FREE-DEM-OPA-SO-1A-P833.aspx

Hope that helps.

Thanks so much, Jeffrey. i see that Rf is the resistor from output to inverting (-) input; what is R_L (100 Ohms on the datasheet)?

Aaron

Aaron,

Rl is the load resistor.  As long as your load resistor isn't too small, you shouldn't really have much issue with it outputting properly.  You currently have your load at 1MOhm||20pF (the impedance of the oscilloscope), so the output swing shouldn't be an issue.