What is your favorite most versatile op-amp?

[David Hess]

Noise can't be accounted for in software, so it's pretty important in my opinion.

Noise easy to null out by averaging in software. Though ADC sampling rate needs to be higher.

If flicker noise could be averaged away, then there would be no need for chopping.

"Hardware averaging" - putting a few devices in parallel?

Averaging by paralleling devices (parallel averaging?) or using larger ones is a prime way to lower flicker noise when chopping cannot be used.

What does not work effectively is averaging over time to remove flicker noise.

What I'm getting at, there are solutions to minimize noise & offset, but distortion /non linearity are different species. This is why I'd move to #1 in my priority list.

It is just application dependent.  Broadband noise, drift (essentially low frequency noise), non-linearity, and distortion (non-linearity in the AC domain) can all be improved in either the analog or digital domain.  "Predistortion" is a common digital technique for combating distortion in RF power amplifiers now.

Probably, DC accuracy may get better by increasing open loop gain using 2+ OPA as combine or composite whatever it's called. Anyway it's not so easy like paralleling for noise reduction.

Noise can usually be reduced to a very small subset of a properly designed circuit.  Lots of different things besides excess gain contribute to DC accuracy.

[MasterT]

///
It is just application dependent.  Broadband noise, drift (essentially low frequency noise), non-linearity, and distortion (non-linearity in the AC domain) can all be improved in either the analog or digital domain.  "Predistortion" is a common digital technique for combating distortion in RF power amplifiers now.

 I know how they do predistortion, was thinking to build ultra low thd generator ( <-120 dBc). Can't find an adc that could prove my algorithm is good. ADS127L01 comes in "hate to solder" package. And evaluation board costs a fortune.
 Though, it would be quite expensive method to correct cheap OPA non-linearity by introducing extra linear adc on chip, converting input to digital, pre-distort, than convert back to analog again by extra linear dac, don't you think?   

[maxwell3e10]

I think the question is not how to overcome op-amp limitations or to get to state-of-art performance in one particular spec. The question is how to get to 90% (in log space) of the state-of-art in many specs at once.

That is why I prefer FET-input op-amps. They have a little more voltage noise, but allow one to use larger input resistances and not worry about it. +/- 15V supplies also make it more versatile. Power dissipation is a different aspect, but usually I don't care about it.

[David Hess]

I know how they do predistortion, was thinking to build ultra low thd generator ( <-120 dBc). Can't find an adc that could prove my algorithm is good. ADS127L01 comes in "hate to solder" package. And evaluation board costs a fortune.

Though, it would be quite expensive method to correct cheap OPA non-linearity by introducing extra linear adc on chip, converting input to digital, pre-distort, than convert back to analog again by extra linear dac, don't you think?

It would be expensive and also futile because there are simple techniques to reduce distortion in differential input feedback amplifiers.  Once these are used to remove first (second?) order distortions, improvements through digital distortion correction would be futile because distortion can be below the noise floor.

These techniques are not often used simply because distortion is already usually better than needed and even less understood because they are easy to overlook in other designs.  They include:

1. The most obvious one is to use inverting instead of non-inverting stages to remove common mode effects.
2. Match the impedance and not just the resistance at differential inputs.
3. Bootstrap the supplies to increase common mode rejection.
4. Unload the outputs of integrated devices to remove thermal feedback.