I wanted to know the difference between lockin detection vs FFT. It appears that lockin is a single point FFT. Whether lockin and FFT can be interchangeably used?
No.
An FFT uses a sine waveform without overtones as "correlation signal", thus the output is only related to a single frequency.
But the lock in uses a square wave as correlation signal. A square wave is a combination of various sine waves (--> fourier series). Fundamental frequecy, 3rd, 5th, 7th ... overtones
Thus the lock in also reacts on 3rd, 5th, 7th.. overtones.
Additionally a FFT doesn´t only give a "real" result, but a complex result.
Thus you can calculate amplitude as well as phase shift.
You may get some kind of complex result is you do the lock in with two "correlation signals":
1) the usual square wave
2) plus the 90° phase shifted square wave.
I need some more clarification. Can VCO not generate pure sine wave( like DDS chip)?
In that case I can multiply two pure signals without overtones and respond only to the fundamental? So FFT becomes synonymous with Lockin?
Also a basic question, what does it mean by "lockin"? What are we locking?
Certainly, I do not claim that both are identical. It appears that both the methods give what I want, Viz.. magnitude and phase. So what is the subtle difference between them?
There are different LIA approaches:
* sine or square wave reference signal
* one (with or without varaible phase shifter) or two reference signals
* different filter types of output filtering
So if your LIA is
* with two exactly 90° phase shifted sine input reference signals with exact same amplitude
* and the output filter is a simultaneously controlled integrator over exactly n fullwaves
then the result should be equal to a single frequency FFT output.
Absolutely. Nevertheless there are some applications where lock-in amplifier and FFT spectrum analyzer can be used both. Unfortunately garimella didn't tell his application.