Extracting X and R components from received signal

[nick_f]

Hi All,
I was reading from Bruce Candy's "Metal Detector Basics and Theory" document and I find it very infiormative for general public.
Does anybody know what circuits are able to extract X and R from the signal?
For a sinewave received signal Rx, I think if we could calculate sin(Rx) and cos(Rx), that would give us X and R (I don't know what circuits can do this operation anyway).
But for Pi detectors, where we have a multitude of frequencies in a square signal, what can be done to obtain X and R?

Regards,
Nicolae

[Carl-NC]

For a sine wave, you use quadrature synchronous demodulators.

[mikebg]

DSP

Nicolae, the signal in PI is not square. The DSP for nonsinusoidal signals is described in patent US4,506,225. See FIGs 9 and 10 and read how they are calculated.

[nick_f]

Hi Carl and Mike,
Thanks for the useful information. I was wrong about Pi Detectors.
I just found this paragraph in Bruce Candy's document:

"A transmit voltage (e.g. –6 volts) is applied to a transmit coil which produces a transmitted magnetic field. This magnetic field is suddenly turned off (this is when the spark occurs in the ignition system - analogy with the ignition system of cars).

After the magnetic field is stopped a period occurs when the metal detector measures a receive signal produced by a magnetic field from the environment. During this receiving period, there are no X components, only R, because X only responds to the transmitted field which has been stopped. These R signals come from 2 sources:
• The eddy currents in a metal target dying down and thus their generated magnetic fi eld decaying in strength.
• The magnetisation of the mineralised soil, caused by the magnetising transmitted field, decaying in strength.
Fortunately, most mineralised soils demagnetize in a predictable way following the magnetizing transmitted fi eld, so it is possible to subtract this predictable signal from the received signal, thus only detecting metal target signals, which as commented earlier, may have any decay rate from very fast to slow"

Regards,
Nicolae

[mikebg]

Hi Nicolae, and Bruce Candy was and is wrong about PI metal detectors.

Do you want to call Pulse Induction the excitation method described in the aforementioned patent? It is simply Nonsinusoidal Induction.
Unlike Bruce Candy, I would explain in two other ways the operation of metal detectors. The first way is in time domain, which seems logical to a scientist used to make mathematical analysis with calculus, that is, solving differential equations.
The second way is convenient for a designer who prefers to make calculations in the frequency domain because two complex mathematical operations called Convolution and Deconvolution are replaced by multiplication and division. So the formulas in frequency domain are simple enough, so except for the calculation, they can be used for analysis. In the frequency domain variable time is replaced with another dimension - periods for a unit of time.
I will explain operation of both types of metal detectors in the frequency domain because it is visual understandable. For this purpose it is sufficient you to explore the figures in thread "Targets Frequency Response", but it is better to read from beginning interesting posts there. I will post my explanation in that thread.
Recently the REMI group analyses in time domain the eddy currents in a coin. Instead Frequency Response function, in time domain is used Impulse Response, Step-Down Response or Step-Up Response. These things, discussed in above patent, are explained in Wikipedia.
See you in mentioned thread.

[nick_f]

Hi Mike,

Thanks for feedback. I was reading your various posts, they are interesting. At the moment they are too advanced for me, but I'll keep reading.

Regards,
Nicolae