More FBS Geekery

Yeasty Wrote:
-------------------------------------------------------
> Albert, there may not be a noticeable difference.
> I am merely speculating on what I see.
> My understanding is that the 25KHz produces
> harmonics that reach all the way down to ~1.5KHZ.
> Look again at the second picture in my first post.
> The spikes (harmonics) are signals at the
> frequency listed under the waveform.
> Pete


Harmonics are ALWAYS multiples of the fundamental frequency, NEVER a submultiple. Thus a 25KHz frequency CAN NOT produce harmonics
down to ~1.5KHz. Suggest you google Fourier series.

The early part of this thread has several errors of this type, but I think they've all been pointed out and we've moved on.

The other forum thread brought to my attention that there are a few Youtube video's, and suchlike, where people have measured the detectors frequency using a simple pick-up loop coil and a frequency meter. This is useful in as far as it will tell you which Channel is highest freq, and which is lowest. But as it's an average figure, measured over 1 second, it won't give you the breakdown of the two individual freqs.
As we've observed, they transmit two frequencies alternately, nominally 25KHz and 3.125KHz, each for the same time duration, 8 cycles of 25KHz, and 1 cycle of 3.125KHz. The exact relationship between a frequency meter average and the actual signal, can be calculated:

Average frequency = 9/16 x High frequency = 9/2 x Low frequency

Or, should you want to reverse this:

High frequency = 16/9 x Average frequency; Low frequency = 2/9 x Average frequency

The High freq is always exactly 8 times that of the Low freq.

This should allow you to check those readings in the video (and other online sources).
For example if someone says Explorer Ch11 measures 12.7KHz average.
We can work out: High freq = 22.56KHz; Low freq = 2.82KHz.

At the end of the day, all the detector can do is to measure the Admittance of the target. Some detectors do it at a single frequency, others do it at multiple frequencies.

In our everyday parlance we are always talking about the target being a high or low conductor, or conductance. But lets not forget that conductance is simply the Real part of the Admittance, a complex number.
The Imaginary part is the reactive part and one way it manifests itself is in the phase angle between the transmit field and the received target response. In the case of White's, the target is measured at up to
three frequencies. I suspect Minelab's FBS also evaluates the target at three frequencies although the three frequencies may be altered as part of the noise cancellation feature.

I'm intrigued as to what makes you think FBS uses three frequencies?

Pimento Wrote:
-------------------------------------------------------
> I'm intrigued as to what makes you think FBS uses
> three frequencies?


The frequencies are being sequentially transmitted and the receive channel needs to be gated coherently with the transmitted frequency,
then time is necessary to process the acquired signal. While this is happening, another frequency is transmitted and a second receive channel
is gated ON and a sample acquired, which also needs time to be processed. While this is happening, another frequency is transmitted and a third
channel is gated ON and a sample is acquired, which then needs to be processed.

Hopefully by then the first channel has finished and is now available to be used again. The alternative is to have more channels, which adds to cost
and requires even slower coil sweep speed. The coil has been moving all this time, granted its been moved slowly, as required by most Minelab FBS
machines, but how much slower do you want the coil sweep to be?

In the above, I --temporarily-- ignored the harmonics that were being transmitted along with the fundamental frequencies that triggered the coherent
sampling of the receive channel. So yes, the receive signal is the target's response to the superposition of the fundamental frequency and its harmonics.
However, a quick inspection of the Fourier series that mathematically describes the waveform in terms of the fundamental and harmonic frequencies shows
that the energy in the transmitted harmonics are dramatically less than the energy in the fundamental, so for all practical
purposes, they are not really contributing much to the target's Eddy current field and therefore to the target detection and analysis process.

______________________

Any married man should forget his mistakes - there's no point in two people remembering the same thing...



Edited 1 time(s). Last edit at 07/30/2015 03:54PM by Rudy.

I'm still unsure why you think there's 3 frequencies involved. This entire thread confirms there's just 2 frequencies, 25KHz and 3.125KHz. They are transmitted sequentially with their own demodulator, alternating every 320 microsecs (1 cycle of 3.125KHz). I can see that there will be a small delay arising from this sequential process, you can only perform the 'math' on out-of-date info. So, for example, at the end of the 1 cycle of 3.125KHz, the most up-to-date figure for the 25KHz reading will be 320 microsecs old. But I think all these delays are rather trivial, especially at the slower sweep speed that works best with these machines.
The slow sweep speed is most likely determined by the available processing power of the micro used in the design.

I'm pretty sure no harmonics of either of the two square-waves are used in any of the Sovereign-thru-CTX range. The 75KHz (3 x 25K) would be too high to be of much use, anyway. Using the 9.375KHz 3rd harmonic of the 3.125KHz would be possible, in principle, after all that's pretty much what Fisher do in their 5KHz/15KHz "CZ-series" machines. But according to reliable sources, like Carl Moreland (username "Geotech" on here), the clever stuff that the eTrac/CTX does is more down to using the complete half-cycle of the 3.125KHz signal as a PI signal. Not harmonic analysis.