Taking a look at GROUND BALANCE on the bench

[mikebg]

... In practice, it's pointless to use Fourier transforms and other methods to design a metal detector until the target signal has been isolated. Even then, one must keep in mind that "signatures" based on data analyses are valid only for targets of known dimensions and materials, such as coins. Gold nuggets have irregular paths for the eddy currents and they are usually alloyed with other metals, which means that their conductivity is not predictable...
Prospector_Al

Allan,
All for nonferrous conductivity is predictable in Frequency domain, because we can use normalized locus.
Look that a wideband metal detector can show if the target is as ring or not.

[Prospector_Al]

Are you both (Aziz and mikebg) selectively reading the posts in this thread?
If not, then how did you miss the following post by Carl?



mikebg - By the way, no answer at all from the Scientific American archive site. Perhaps it's a dead link, in the same way as Aziz's "holly [sic] grail of MD".

------------------------------------------
Note: The Latin adverb sic ("thus"; in full: sic erat scriptum, "thus was it written") added immediately after a quoted word or phrase (or a longer piece of text), indicates that the quotation has been transcribed exactly as found in the original source, complete with any erroneous spelling or other nonstandard presentation.
------------------------------------------

There is another source of information that is not mentioned often: Many very clever ideas using sophisticated means of data analyses have been patented. I have read over a hundred of these patents. They looked interesting, but not practical. What many armchair theorists seem to overlook is that the signal from a target is inextricably intermingled with the ground signal. In practice, it's pointless to use Fourier transforms and other methods to design a metal detector until the target signal has been isolated. Even then, one must keep in mind that "signatures" based on data analyses are valid only for targets of known dimensions and materials, such as coins. Gold nuggets have irregular paths for the eddy currents and they are usually alloyed with other metals, which means that their conductivity is not predictable.

It would take a computer like "Watson" to sort out what's what. Not exactly appropriate for a hand-held metal detector.

It may be that many clever ideas are not pursued because theyv'e been tried already and found wanting...

I second Qiaozhi's suggestion that if you you want to impress the Forum readers with your knowledge, then design a metal detector using your ideas and show us the results--at least some wave forms on a CRO screen...

Prospector_Al

[Midas]

Here is a brain question for you .... imagine I have a clock with infinite precision and a pulse generator .....now at a certain precise time the sig gen emits a single pulse .... all the data in the universe can be encoded into that time value by choosing the correct time value to emit the pulse ... irrespective of the pulse width amplitude and phase in FD.
If you cant understand this ... you are not as smart as I am

Of course even theoretically there is no such thing as infinite precision thanks to the uncertainy principal. Also its not actually a question smarty pants.

[moodz]

The problem with FFT is that it is designed for "continuous waveforms" ... FFT has no time dimension .... for instance it is not hard to show that the frequency energy of a PI pulse changes with time across a single pulse decay. FFT does not work here because it has no time dimension .... ouput is amplitude and phase ... no indication of whether the data was gathered over years or attoseconds ... Unfortunately ( for FFT ) the PI pulse time dimension is extremely important ... FFT throws this data out .... I can achieve near perfect "right angle damping" using a couple of dollars worth of components ... your FFT system has some time to go till the same result can be achieved numerically ....( mainly related to how fast and accurate your ADC/CPU is == $$$ ).

Here is a brain question for you .... imagine I have a clock with infinite precision and a pulse generator .....now at a certain precise time the sig gen emits a single pulse .... all the data in the universe can be encoded into that time value by choosing the correct time value to emit the pulse ... irrespective of the pulse width amplitude and phase in FD.
If you cant understand this ... you are not as smart as I am

[Carl-NC]

This all reminds me of C++ vs Lisp, Unix vs DOS (OK, that one's easy), VI vs WYSIWYG, and Ford vs Chevy.

Tellyawhat Mike, if you want to convince us then start a new project from scratch. Maybe a very simple PI. Use FD and go step-by-step through the design/analysis process. I'm listening!

[Altra]

Aziz, Do not expect answer.
The FFT appears in the 60ties of the 20th century. Because of its power, it quickly begins to apply. Do not expect the experimenters in this forum soon to apply it. Your question is too modern for them.

Five years ago I asked the experimenters in this forum for something that is much more easy and enough old because it matches the level of human knowledge in the 19th century.
I came 5 years ago on this forum with a question that no answer to this day. See my first post on the forum and read what desperate answers received:

http://www.geotech1.com/forums/showthread.php?14005-Targets-frequency-response&p=67949#post67949

Forum participants did not know that to design even a NARROWBAND metal detector they need WIDEBAND information how to select operating frequeincy.
Since after years no one has shown that he uses Frequency spectrum for measurement of targets, I decided to show how to make it:

http://www.geotech1.com/forums/showthread.php?20163-WIDEBAND-(BROADBAND)-TECHNOLOGY&p=167869#post167869

I hope that someone experimenter will start to use broad band spectrum of Frequency domain. Then he will understand that he need FFT.

See post 6 in this thread.

http://www.geotech1.com/forums/showthread.php?20023-VLF-vs-PI-SHOOTOUT-PROJECT

I am doing a 256 point dft with a dspic33 to desplay the frequency in real time. The ad5933 chip on the little green board
is doing a 1024 point dft on each frequency generating real and imaginary data for each frequency. You are wrong about people on this forum not
understanding frequency domain. It's not always the best tool for real world solutions, maybe in a text book.

[Carl-NC]

Guess extreme power consumption is a part of "mine is bigger" psychology that may sell better the rigs that suck more batteries, or that suck altogether - I don't know.

There is truth to that... I am told on a regular basis that if TDI had a much bigger battery pack, it would sell better.

It is odd that a technology that radiates virtually zero energy to the surrounding space often wastes so incredible much, while simultaneously is practised in areas far, far away from a nearest power plug. It is absurd.

We like generating heat.

[Qiaozhi]

has anyone made the frequency domain (FFT) response of various targets (ferrite, magnetite, maghemite, iron, copper, etc.) on an induction balanced (IB) coil configuration and a wideband continious wave transmitter?
This is somewhat enlighting and reveals everything and the holly grail of MD.

Aziz

Aziz, Do not expect answer.
The FFT appears in the 60ties of the 20th century. Because of its power, it quickly begins to apply. Do not expect the experimenters in this forum soon to apply it. Your question is too modern for them.

Are you both (Aziz and mikebg) selectively reading the posts in this thread?
If not, then how did you miss the following post by Carl?

In the context of my current project, which is a hybrid VLF + PI, it's very easy for me to think about, and even simulate, this circuit in the TD. So I got to thinking about what it would look like in the FD and, holy smokes, it would be a huge spectral mess. I think if I had used FD I would still be in theory-land, instead of field testing.

mikebg - By the way, no answer at all from the Scientific American archive site. Perhaps it's a dead link, in the same way as Aziz's "holly [sic] grail of MD".

------------------------------------------
Note: The Latin adverb sic ("thus"; in full: sic erat scriptum, "thus was it written") added immediately after a quoted word or phrase (or a longer piece of text), indicates that the quotation has been transcribed exactly as found in the original source, complete with any erroneous spelling or other nonstandard presentation.
------------------------------------------

[mikebg]

Was ist das FFT?

Hi all,

has anyone made the frequency domain (FFT) response of various targets (ferrite, magnetite, maghemite, iron, copper, etc.) on an induction balanced (IB) coil configuration and a wideband continious wave transmitter?
This is somewhat enlighting and reveals everything and the holly grail of MD.

Aziz

Aziz, Do not expect answer.
The FFT appears in the 60ties of the 20th century. Because of its power, it quickly begins to apply. Do not expect the experimenters in this forum soon to apply it. Your question is too modern for them.

Five years ago I asked the experimenters in this forum for something that is much more easy and enough old because it matches the level of human knowledge in the 19th century.
I came 5 years ago on this forum with a question that no answer to this day. See my first post on the forum and read what desperate answers received:

http://www.geotech1.com/forums/showthread.php?14005-Targets-frequency-response&p=67949#post67949

Forum participants did not know that to design even a NARROWBAND metal detector they need WIDEBAND information how to select operating frequeincy.
Since after years no one has shown that he uses Frequency spectrum for measurement of targets, I decided to show how to make it:

http://www.geotech1.com/forums/showthread.php?20163-WIDEBAND-(BROADBAND)-TECHNOLOGY&p=167869#post167869

I hope that someone experimenter will start to use broad band spectrum of Frequency domain. Then he will understand that he need FFT.

[Davor]

I have a "Super BBS" transmitter design that solves this problem, but it's also a power hog.

There are multi-resonant networks that may serve you as diplexers, and simultaneously feed a Tx coil with near equal signal level on several frequencies. They'd require multiple voltage drives, but at respective frequencies would perform very efficiently. Guess extreme power consumption is a part of "mine is bigger" psychology that may sell better the rigs that suck more batteries, or that suck altogether - I don't know.

It is odd that a technology that radiates virtually zero energy to the surrounding space often wastes so incredible much, while simultaneously is practised in areas far, far away from a nearest power plug. It is absurd.

[Carl-NC]

I think if detectors can be used for 8 - 10hrs continous then recharge serious users are not bothered. Prosumer users are after performance, Power consumption is not a huge seller. Speaking from a consumer view.

True, but I've also shot myself in the foot by designing a water case that only holds 4-AA batteries. Ergo I'm starting out with half the watt-hours, and need all the help I can get.

[golfnut]

I think if detectors can be used for 8 - 10hrs continous then recharge serious users are not bothered. Prosumer users are after performance, Power consumption is not a huge seller. Speaking from a consumer view.

Most of these puppies should be using Li rechargables - they are light and store masses of energy and not that expensive. Model aero's use them by the bucket load. Either Lipo or LiFe.

S

[Carl-NC]

I'm curious, what motive is behind merging CW and PI? IMHO BBS is a waste of energy, same as most of the PI in the market today, but it has a marketing grand standing. Is it just for the show, because you can do that, or something more tangible?

The way I'm combining CW & PI results in better power efficiency, a GB that adds target depth rather than reduces, and the ability to ID iron.

You are right, BBS is somewhat of an inefficient power hog, but besides marketing grand standing it is widely considered to be a top choice for detecting deep silver. However, it is rather poor at detecting low conductors... I have a "Super BBS" transmitter design that solves this problem, but it's also a power hog.

[Aziz]

Hi all,

has anyone made the frequency domain (FFT) response of various targets (ferrite, magnetite, maghemite, iron, copper, etc.) on an induction balanced (IB) coil configuration and a wideband continious wave transmitter?
This is somewhat enlighting and reveals everything and the holly grail of MD.

Aziz

[Davor]

I'm curious, what motive is behind merging CW and PI? IMHO BBS is a waste of energy, same as most of the PI in the market today, but it has a marketing grand standing. Is it just for the show, because you can do that, or something more tangible?