Multi-Frequency

[Aziz]

Paul99 - it is becoming clear to me that you are a reincarnation of Urban Fox, who was banned previously (along with Doug) by Carl. Continue like this and you'll be heading for the banned list once again.

Aziz - You've been warned before. Clean up your act or you will also taking a sabbatical.

We are very lenient here, and don't want to waste our time deleting abusive posts, so please take this is a your LAST WARNING.

Ok sir, yes sir.

But it would be fair, if you allow rejoining Doug as well. At least as long as Paul99 is allowed.

Aziz

[golfnut]

"The best way to avoid this is to shunt this coil with resistor"

Sure, but I dont have the machine any longer...

S

[Paul99]

Paul99 - it is becoming clear to me that you are a reincarnation of Urban Fox, who was banned previously (along with Doug) by Carl. Continue like this and you'll be heading for the banned list once again.

Aziz - You've been warned before. Clean up your act or you will also taking a sabbatical.

We are very lenient here, and don't want to waste our time deleting abusive posts, so please take this is a your LAST WARNING.

Qiaozhi, do you think you might be so kind as to remove Aziz's post #36 from the thread, as it should be obvious that it contains content designed to damage Minelab's very good reputation.

[Aziz]

Qiaozhi, do you think you might be so kind as to remove Aziz's post #36 from the thread, as it should be obvious that it contains content designed to damage Minelab's very good reputation.

Agreed.

[Aziz]

Hi all,

most detectors have simple demodulators due to cheap & easy implementations. Even when they mention dozens of frequencies sent to the ground, they process very few frequencies. Some of them even use single frequency out of two or three (not at the same time).

My laptop project is demodulating the true multi-frequency TX response and I get hundreds of frequencies demodulated (depending of the FFT size). This isn't necessary of course as a two frequency demodulation works very fine as long as you transmit all the TX energy on that two frequencies. But if you have a wideband TX and you process only few frequencies, you are wasting TX energy. That's not required.

Now the interesting stuff on the two-frequency demodulation:
The reactive (X) response dominates in the low frequency region.
The resistive (R) response (=target response) dominates in the high frequency region.
You can remove the reactive response in the high frequency region (& do a ground balance):
S(high) = high frequency demodulated signal
S(low) = low frequency demodulated signal

R(high, residual) = S(high) - a*S(low), where
a = linear combination factor
R(high) = ground balanced target response

As the high frequency region has low response energy (short period of time) you have to transmit more high frequency radiants (waves) (seen in the scope above). That's all. All simple math. And simple GB.

You don't need dozens, hundreds, billions of frequencies. Two frequency demodulation works just fine.

Aziz

[Aziz]

Urbanfox,

no matter what you do, Minelabs reputation has been already destroyed.
Even I cannot do more to destroy it further.
Aziz

[Qiaozhi]

Urbanfox,

no matter what you do, Minelabs reputation has been already destroyed.
Even I cannot do more to destroy it further.
Aziz

OK - let's leave it there.
Any further Minelab bashing will simply be deleted, no questions asked.

The topic of this thread is Multi-Frequency, so please stick to it from now on. Discussing Minelab technology is perfectly acceptable, but denigrating the company is not. Also, no more personally abusive comments or you know what's going to happen.

Post #36 has been removed.

[Skippy]

GolfNut: your scope traces show the voltage on the transmit coil. Using a pickup coil just creates a transformer, your pickup coil is the secondary, it's voltage mimics the primary voltage. The Current is completely different, it's the integral of the voltage (V = -L.dI/dt). To see that, you need a hall effect probe, or a (small) series resistor as a 'ammeter'. You will see the result is a triangle-type ramping waveform. Of course the search-coils receive-coil is a differential transducer, so what you get back from the target is not obvious.

[Carl-NC]

I have a small amount of additional for the exp2. Using a small pick up coil to visualise the mag field from the Tx coil on a scope..
(pick up coil resonates at over 200kHz so ring is not from pick up coil) You see additional decaying signals after Tx step edge.
This displayed voltage mimics the current in the Tx coil, and its resultant mag field.

Sorry, but this is completely wrong. The actual voltage drive doesn't have all this ringing, it is a product of your pick-up coil. And Skippy is right, this doesn't represent the coil current or the magnetic field. Use a current probe to directly measure the current, and it quickly becomes obvious what is going on.

[golfnut]

Sure, the ring is from my probe.

I am confused at having to flush the idea of the two coils (probe and Tx coil head) in close proximity not coupling magnetically. I thought thats how transformers worked.
Especially when the source has a good E field screen, the only energy left coming out is mag H. This is why a hall sensor would also produce output? Ho hum, bow to greater knowledge.


The Tx drive was 20v pp incidentally.

S

[Davor]

A transformer-like behaviour is what you have ...with superimposed ringing. Ideal pickup would be resonance-free, but in reality a dampened one works fine.

A transformer with a core provides near 100% coupling, so a secondary winding sees a low parallel impedance of a primary, hence it is heavily dampened. Lightly coupled transformers tend to ring.

Even with all this ringing it is possible to discern what is what. Thanks.

[Davor]

For anyone who may have lost track, Davor was referring to Mr. C's writeup quoted in post #7 by golfnut.

What Mr. C (as in I suppose "Bruce Candy") was quoted as saying was from an engineering perspective neither gibberish nor inaccurate, even if it does happen to have been written to support the Minelab marketing effort. I do the same for FTP-Fisher, and I suppose that Carl does the same for White's. It's not like nobody here understands this stuff.

Davor, you got rather badly hammered in this thread and for very good reason. For your sake as well as the sake of those on this forum who endure your posts, I hope that this time the pounding hurt bad enough to make you think. If you're clueless as to what your next thought should be, here's a virtual Chinese Fortune Cookie for you to open:

<open>

"Better to learn from people who are smarter than you are, than from those who know less."

--Dave J.

Dave, I have immense respect for your work, but occasionally your comments seem as if you didn't read the critical parts. I understand that you use multi frequency technology, and I like to believe I know why, yet the quoted Mr. C passage doesn't say it. It also abuses the multi frequency buzz to push PI-like technology as a multifrequency. It resides exclusively in uncorroborated phrases, e.g. marketing. You never stroke me as a person proud of such practice.

In a mentioned passage a distinction between rectangular and square wave spectra is also meaningless in context. Duty cycle affects spectrum minima, and for square wave those are at even harmonics. While these are the facts, the spectra of the excitation signals are not discussed any further, only periods, as this is truly a PI-like technology with silly timing scheme. Hence the "spectra" are just borrowing a "buzz" for the rest of the torture.

At two places it uses conditional sentences of a sort, invoking a sense of a promise. "These can be used to more accurately..." translates to "If these are used, the more accurate...", and "it is possible to extract a better assessment..." translates to "If some unsaid approach is applied it would lead to a better assessment ...". It would be fine if those conditions are resolved in some past or future, but they are not. Conditional sentence type 1 and type 2 are mutually contradicting, and the result is unclear at best:
- type 1: It is possible and also very likely that the condition will be fulfilled.
- type 2: It is possible but very unlikely, that the condition will be fulfilled.

Hence we'll neither learn if different frequency R signals are ever used to determine time constants, nor we'll ever learn if Fe targets discrimination works well, or it sucks big time.

This may as well be the best ever metal detecting technology in the world, but I just fail to notice any grounds for such conclusion in text. To me it is obscured with marketing.

I'd be happy to learn some more. I have no problem receiving a lecture or two and I hope you'll be generous this time.

[Davor]

First, sorry for necroposting here, but opening a new topic with same old title made even less sense. Without all the trolls pressure this may even work this time.
Last time I took the marketing gibberish with flaming passion, but perhaps it is time to look at this more calmly.
I can even admit that I understand the BBS/FBS approach, the time sequential frequency approach, but I never tolerate marketing Stierscheiße.

Speaking of FBS/BBS, I find it odd that the market leader let the key patent expire in 2014 without a successor, which makes this technology a fair game for enthusiasts and beyond. Even their page on patents is void of any fundamental technology for multifrequency rigs.

Fun part is that a whole detector could fit into a single micro, with very little of anything else aside. Even an analogue solution may be reasonably simple, with proper design even simpler than a classic VLF. Any takers?

[WM6]

Any takers?

Takers awaiting up to 40% more from you.

[Davor]

I wish that myself. 40% more money and 40% more free time.

Long ago at a Zagreb Faculty of Mechanical Engineering and Naval Architecture there was a quote carved in a bench, something like: "I wanted to live without toil, and now I'm toiling without life" - that sums it.