Synchrounous signal extraction technique name

I've seen it called any number of things, including things not even repeatable here. So we probably have licence to call it pretty much anything we want. How about if I call it "Fred Bloggs" from now on, and you call it "Tim Wescott".

After four beers I'm always incoherent.

your colleagues are correct. there are formulars that work much faster than the conventional math used in basic DFT/IDFT (inverse....)_ FFT is commonly used.. (Fast fourier Transforms)/ if you want to see more code like this. search google for files. fourier.pas,FFT.pas, FFT.C, FFT.CPP etc..

you will find lots of examples.

In my line of work, we extract the magnitude and phase of a response signal when the excitation signal is known. The excitation is of the form A*sin(wt) and the response is of the form B*sin(wt+theta) + noise. The response is sampled 16 times over a full period. Let's denote the sampled response as R[0] through R[15]. Then the inphase (real) and quadrature (imag) components are extracted by:

real = sum(i=0 to 15) of [R[i]*cos(2*pi()*i/16)] imag = sum(i=0 to 15) of [R[i]*sin(2*pi()*i/16)]

From this the magnitude and phase of the response signal is determined.

My question is what is this technique called? I'm sure there is a common name for it. Some of my colleagues refer to it as a DFT (Discrete Fourier Transform). Since only a single frequency component is being extracted, I don't feel this is proper nomenclature. I'm looking for a more correct term to use to refer to this single frequency extraction technique. Does anyone have a good name for this?

Thanks,

On Fri, 27 May 2005 23:14:28 GMT, "Steven K. Moore" wroth:

How does that technique differ from the traditional "lock-in amplifier"?

Jim

This is sampled first and numerically processed (aka DSP). I think of the lock-in amp approach as analog I & Q multipliers and lowpass filters to obtain real and imag components.

Your computation is the very definition of DFT evaluated at Fs/16, Fs=sampling frequency =w =2*pi*Fs/16 in the case. DFT is a function that evaluates at 0, Fs/16, 2*Fs/16,..., 15*Fs/16. So you could term this "DFT evaluated at the resolution frequency."

Nah- it's called "coherent detection" generally because there is coherence between his sampling frequency and the signal frequency.

Is it incoherent or noncoherent detection for you?

I'd vote for "The Moore Detector". ;-)

Cheers! Rich

If you wish, it's still a lock-in amp.

Thats just silly. Use "Tim Bloggs" and "Fred Wescott".

Cheers DaFTerry

On Sat, 28 May 2005 00:13:18 GMT, "Steve" wroth:

Almost any analog technique can be implemented digitally by sampling followed by digital processing. I still think your device could ne called a lock-in amplifier. Maybe a digital lock-in amplifier?

Jim

Thanks, but I'm quite fami,iar with FFT algorithms (My master thesis developed a parallel processing spectrum analyzer based on Cooley-Tukey FFT algorithm.) The need is not a full spectrum decomposition, but merely the extraction of a single frequency. Sort of a DSP version of a lock-in amplifier.

It doesn't seem that a definitive name has bubbled to the top. I've heard:

• DFT evaluated at the resolution frequency

• I/Q demodulation

• coherent detection

• Fred Bloggs :)

• Tim Wescott :)

• "The Moore Detector". ;-)

Thanks for all your input. Steve Moore

On Sat, 28 May 2005 18:28:51 GMT, "Steve" wroth:

If Win Hill says it's a lock-in amp, then it's a lock-in amp.

Jim

My most humble apologies. When I went back to collect all the answers suggested, I failed to record the response from one of my most respected voices in this newsgroup. Win, I beg your forgiveness.

Steve Moore

Why? Where's the amplifier?

John

The "amplifier" part of the name is tradition, little more. BTW, most modern lock-in amps are DSP based these days.

signal

term

anyone

This guy does only 4 samples per cycle, which results in serious calculation minimisation, and calls it Quadrature Direct Fourier Transform (QDFT). see

Gerhard van den Berg