Chopper amplifier vs. lock-in amplifier

[slugrustle]

It strikes me that a chopper amplifier (not chopper stabilized but proper, old school chopper amplifier) is either similar in concept to or one specific implementation of a lock-in amplifier.

Both have the concept of synchronous modulation and demodulation with amplification in between, followed by a low-pass filter (and perhaps more amplification). 

It seems the chopper amplifier only has the X output of a more full-featured lock-in amplifier, but otherwise the concepts are quite similar.  Am I missing some distinction here?

[KrudyZ]

Modern lock-in amplifiers use DSP filters to pull the signal from the noise at any frequency within their range.
Chopper amplifiers are trying to do almost the opposite. They chop at a frequency that is not supposed to be present in the signal and use the difference they detect to null the main amplifier, so it has low offset for the signals of interest.

[slugrustle]

If you consider the typical use case of an optical chopper with a lock-in amplifier, it's not that different conceptually.  And analog lock-in amplifiers are a thing.  That's a better comparison for the purposes of this topic.

[Kleinstein]

The chopper amplifier starts with an essentially DC signal at the input and includes the chopper as one of the first parts in the signal path. The modulation and demodulation are internal to the chopper amplifier and externally the signal is "DC" and the chopper signal is usually not accessible. Externally the chopper amplifier looks like a DC amplifier. The difference to a chopper stabilized amplifier is in the internals and the usually smaller (well less than the modulation) bandwitdth.

With an lock-in amplifier the chopper part is external, usually before the actual signal is generated (e.g. mechanical chopper in an optical path). This way one avoids a small essentially DC signal for the whole signal path. The Lock-in amplifier looks at an AC signal and one thus has amplitude and phase or in phase and quadrature part. In the simpler ones as amplitude and an adjustable phase shift to set upfront and not a measured phase.

They both use the same concept of phase sensitive rectification, but are still different things.  One can add an electrical chopper in front of a lockin amplifier to build a copper amplifier. However just a chopper amplifier can be simpler be generating it's own chopper signal and skipping the phase adjustment part. That phase adjustment part for an external reference signal can be quite some effort.
The chopper amplifier often uses rectangular modulation and demodulation, while a lock amplifier is usually considered to look at a sine and suppress the harmonics.

The copper amplifier part can also be done similar to the digital lock-in amplifiers. This is how most of the HP/KS DMMs do the auto zero mode: the switch the input before the main amplifier between zero and the signal and do the demodulation digital by taking the difference of 2 ADC readings.
A digital lock-in amplifier needs an extra ref. input to generate it's idealized reference signal.

[CurtisSeizert]

I think it is most helpful to consider them as members of the modulator/demodulator family, with the key differences being placement of the modulator and feedback. They both reject DC offsets between the modulator and demodulator, but the chopper is seeking to use DC input signals, so it places the modulator in front of the AC amplifier instead of in front of the excitation signal. The differences between the two just come out of this structural change based on the design goals in each case.