OP, inverting or non-inverting?

As usual, it depends. If you just want to get the most bandwidth out of your amplifier, opt for non-inverting - an inverting amplifer with a gain of -1 has a "noise gain"of 2, and thus half the bandwidth of a non-inverting amplifier with a gain of one.

This advantage becomes negligible at high gains.

If you are worried about distortion, opt for inverting - the amplifier input then stays at an almost constant voltage independent of the excursions of the input signal, thus pretty much eliminating any moduation of the base-width with changing collector-to-base voltage (Early effect). FET-input amplifiers are subject to a similar sort of problem.

Jim Williams gives "Williams' Law": Always invert, except when you can't.

Keeping the inputs nearly still helps with all sorts of bias-related problems, such as common mode range limitations and startup problems. The gain of an op amp isn't exactly constant over its common mode range, so that noninverting amps tend to have more distortion.

It's also quite common for an op amp to have an output voltage swing that's wider than the input common mode range--which means that stage 1 is liable to wake up with its output outside stage 2's CMR, so that stage 2's input stage is turned off. This can lead to latching problems, especially if you have a circuit like a PLL, where there's feedback from stage 2 to stage 1. (That one bit me many moons ago when I first started using the late lamented MC34084 quad FET op amp.)

Inverting amplifiers are more or less immune to these sorts of troubles.

Cheers,

Phil Hobbs

Good advice for logic designers too.

Inverting configurations using an op-amp are invariably noisier than non-inverting.

Graham