Taking a look at GROUND BALANCE on the bench

[bklein]

Useful?: http://www.st.com/st-web-ui/static/a...DM00075867.pdf

[mikebg]

. . .
I am very much in favour of exploring new ideas and new ways of doing things, but eventually you have to actually build something to prove whether it works or not. You cannot continue to push different methods or ideas as fact, when there is no concrete (pun intended) proof.

George, the Frequency domain is not new idea and new way of doing things. It is old powerful tool for analysis, design and measurement. Only it is used for design of radio equipment because the Time domain is not suitable. It is not suitable even for design of narrow band metal detectors. Then why should we use it for analysis, measurement and design of PI, TEM, triangular wave and other type wideband metal detectors?

Look above how two professional designers - Eric Foster and Thomas Breuer are compartmented in time domain. They even do not suggest how more useful information will be obtained from their diagrams if they are represented in Frequency domain. Every time function will be splited into two frequency functions which show what is happening and what needs to be done. Then they will see that gold and silver generate first order response and only conductive permeability generates second order response.

The "worlds best" BBS, FBS and FBS2 technologies of MINELAB are not new ideas and new ways.

http://www.minelab.com/emea/consumer/knowledge-base/minelab-technologies#FBS

They use Frequency domain, the old tool born about 200 years ago. Scientists of 19-th century also need several decades to understand that the most powerful tool is born. No need to prove whether MINELABs technologies work or not. Engineers need knowledge to understand how they work. They should read patent US4506225. It describes the discrete signal processing at WIDEBAND technology. Described is the modern technology despite the patent pending is written in 1981.

[PiTec]

Hi Eric,

Nice set of curves, but it is difficult to see any differences between magnetic soil and metal targets with a linear display. If you can, try and convert them to 1) log amplitude/linear time. 2) log amplitude/log time. Targets like a ring will have a single exponential and have a straight line slope in 1), while ground will still appear like an exponential. Ground in 2) will appear as straight line slope, while the ring will curve downwards at late times.

Below are the same decay curves with log/lin and log/log axes. Y= amplitude from 10 to 1000mV, X= time from 0 to 200µs. The problems I mentioned earlier are still visible:

• The curves for hematite and MV soil are not exactly the same (also easy to see in the lin/lin plot, there the red and blue curves are slightly different during the first 100µs). So my question was, how similar should they be? Also, they do not appear as a straight line slope in my plots. Maybe the TX fields I used were too strong and there is a distortion by saturation effects?
• First order (one TC) targets like the gold ring have a straight line slope in the log/lin plot, but as soon as there are more TCs involved (like the silver ring), the delay can become very similar to the viscosity decays.

If anyone wants to take a look at the data, here is the Excel file: decay_curves.rar

Thomas

[ODM]

Let's look how BW is doing the GB.

I'd like to see how BW is doing with GB in the field, as it's touted on this forum about as eagerly as the Jeohunter, with equally few amounts of actual neutral demonstrations - surely you don't want it to look similar?
We might also need a cure for big mouths, or enough pudding to fill them. As we have some quotes, yet another quote goes, "Proof of pudding is in the eating" - or this time, perhaps the proof of a pudding recipe.


On a more serious note, thanks everyone for the reference documents posted in this thread. Along with the coil configurations documented elsewhere in the site and forums, it helps clarify the issues of GB in both balanced coil configurations and others. No tool of analysis is the perfect tool for all situations either, if there was the world's best micrometer it would still be a hassle to use it as a scale. Occam's razor is also a decent tool - it's best applied with caution, of course, but it helps with analysis of "unoptimal" designs. The first assumption shouldn't be that the people investing a long while of genuine research into something have their facts wrong. Best regards!

[Aziz]

Who got it?

Let's look how BW is doing the GB.

Aziz,
curing the huge greed of the big evil company

[Ferric Toes]

Mickstv,
My hobby is design of TXs, RXs and antennas for QRP amateur radio. Just for that I'm posting here information for my hobby friends, who are ham designers. They do not need the circuit diagrams, PCBs and TV videoclips. They need theory that predict experimental results, they need analysis of block diagrams to design suitable circuit diagrams and they need algorithms to make suitable software.

There is a lot information in the WEB for builders like you. The builders like me need more knowledge because we are always dissatisfied with what is achieved. We allways think what to improve. Perhaps that supports our hobby still alive, although over 30 years all know that the HamRadio is a dying hobby.

As for PCBs, this forum is rich on worldclass experts for design and manifacturing. They can make this for you at highest professional level and manufacture PCBs of highest quality.

I can be useful to you only if you want to build a circuit diagram. I can show you the shortcomings of your diagram and how to improve it. Simply point me a circuit diagram.

However, I am sure that there are many on this forum, like myself, who are not so compartmented. I think I qualify as a researcher, developer, designer and builder. I have all those hats, and a few more, which I wear as appropriate.

Eric.

[mikebg]

So mikebg, do you have a working detector based on your many postings. Do you have any YouTube clips showing it's operation, if not we are all really interested to see what you have come up with, please post some YouTube clips.

Mickstv,
My hobby is design of TXs, RXs and antennas for QRP amateur radio. Just for that I'm posting here information for my hobby friends, who are ham designers. They do not need the circuit diagrams, PCBs and TV videoclips. They need theory that predict experimental results, they need analysis of block diagrams to design suitable circuit diagrams and they need algorithms to make suitable software.

There is a lot information in the WEB for builders like you. The builders like me need more knowledge because we are always dissatisfied with what is achieved. We allways think what to improve. Perhaps that supports our hobby still alive, although over 30 years all know that the HamRadio is a dying hobby.

As for PCBs, this forum is rich on worldclass experts for design and manifacturing. They can make this for you at highest professional level and manufacture PCBs of highest quality.

I can be useful to you only if you want to build a circuit diagram. I can show you the shortcomings of your diagram and how to improve it. Simply point me a circuit diagram.

[Ferric Toes]

Hi Eric,


This confirms what I have read in several documents about magnetic viscosity. However, among the many samples I tested there are some with slightly different decays. How similar should they be?

Below are some decay curves. Test setup is a small cylindrical mono coil, samples are in film containers inside the coil, standard PI waveform with 50µs on and 2µs off. Decay curves start approx. 5µs after the 2µs pulse. The TX coil decay curve was removed in the oscilloscope by subtraction, i.e. the plots show the pure decays of the samples. All initial amplitudes are equal so that the curves can be superimposed as in plots D) and F).

I actually made these to show how close the viscosity decay can be to that of metallic targets. It is very easy to get almost the same decay curve with two rings with different TCs by simply varying their distance and angle. This simulates the problem with the detection holes in GB systems that subtract a later sample from an earlier one.

Samples:
A) Hematite powder (alpha-Fe2O3, could also contain some amount of gamma-Fe2O3/maghemite):

http://www.kremer-pigmente.com/en/pigments/iron-oxide-pigments/iron-oxide-red-110-m--light-48100.html

The data sheet says the predominant particle size is 90nm

http://www.kremer-pigmente.com/media/files_public/48100e.pdf

Maybe I try to convert some into maghemite as Aziz suggested here:

http://www.geotech1.com/forums/showthread.php?14102-Let-s-made-a-PC-base-metal-detector-with-usb-interface-!!!/page59

B) Magnetic viscosity soil from my area (concentrated with a magnet)

C) A thin gold ring with one TC of 13.5µs

E) A thick silver ring with at least two TCs of approx. 15µs and 120µs

Thomas

[ATTACH]23599[/ATTACH]

Hi Thomas,

Nice set of curves, but it is difficult to see any differences between magnetic soil and metal targets with a linear display. If you can, try and convert them to 1) log amplitude/linear time. 2) log amplitude/log time. Targets like a ring will have a single exponential and have a straight line slope in 1), while ground will still appear like an exponential. Ground in 2) will appear as straight line slope, while the ring will curve downwards at late times.

Eric.

[Qiaozhi]

Your motto is:
I hate to think, I prefer to experiment.

Apart from being untrue (you only have to browse through the many forums to confirm that for yourself) it does lead to me to ask ... when are we going to see the results of your experiments? Or perhaps there aren't any?

I am very much in favour of exploring new ideas and new ways of doing things, but eventually you have to actually build something to prove whether it works or not. You cannot continue to push different methods or ideas as fact, when there is no concrete (pun intended) proof.

[mikebg]

I've heard much about vapour detectors, now you are proposing a thought detector. When you have finished thinking, imagine you are completely cancelling the ground with no loss in range and completely noise free, then feel the imaginary weight of that huge nugget. Think what that will do for your bank account.

No Mikhail, the trick is to think and experiment, and not necessarily in that order.

Eric.

This is absolutely true. I know two Bulgarian physicists working at CERN. They divide physics into two categories: theoretical physics and experimental physics. Theorists point to the experimentalists what to experiment. Then the experimental physicists spend huge money to satisfy curiosity of the theorists. Fortunately the experimentalists of metal detectors spend their time (and our time sometimes :-).

http://public.web.cern.ch/public/en/lhc/lhc-en.html

[Ferric Toes]

Only Aziz is not Betonkopf because he knows that we should make Fourier transform.

Für alle Betonköpfen (To all ConcreteHeads):

Your motto is:
I hate to think, I prefer to experiment.

For those who prefer to think, I will place suitable information in the WIDEBAND thread.

I've heard much about vapour detectors, now you are proposing a thought detector. When you have finished thinking, imagine you are completely cancelling the ground with no loss in range and completely noise free, then feel the imaginary weight of that huge nugget. Think what that will do for your bank account.

No Mikhail, the trick is to think and experiment, and not necessarily in that order.

Eric.

[Ferric Toes]

Eric,

pity that we don't see what happens in the first 10 µs after switch-off. This is the region, where the magnetic viscosity effects should be significantly visible (if any there).

Anyway. All reactive parts and the demodulation of the signals (in a PI circuit) are highly time dependent.
A function f(t). One have to get rid of the variable t.
This is the most constructive part I can contribute to the discussion.
Aziz

Thanks Aziz, that is a constructive reply.

Sorry if I was a bit sharp in my earlier reply but I had a hard day, and you do leave things hanging in the air sometimes.

I have two PI units that I use for TC and decay measurements; one that can sample as early as 1.0uS and the other starts at 5uS. What I see is the forward extension of the 1/t curve, and it doesn't appear to deviate. All samples are dry and non-conductive. At 1uS you could expect to see low conductivity signals (if there were any) that would be superimposed on the early part of the curve. The PI circuits I use for these tests are a bit different to what you would see in a commercial metal detector and I expect the results would be different. e.g. some earlier papers, and equipment that I designed, gave the slope of viscous ground as 1/t^-1.4, but now I am nearer to the theoretical t^-1.0. Some new gear I am building should be able to get still better accuracy. This is not entirely relevant to metal detection, but to other aspects of soil science.

One area that I have not investigated is what happens in the ON time with a balanced coil. For a truly rectangular TX pulse it should be the same of course, but added to the reactive signal which I think will be much bigger. For half sine and other shapes - who knows?

Eric.

[mickstv]

So mikebg, do you have a working detector based on your many postings. Do you have any YouTube clips showing it's operation, if not we are all really interested to see what you have come up with, please post some YouTube clips.

[mikebg]

Pity that we don't see what happens in the first 10 µs after switch-off. This is the region, where the magnetic viscosity effects should be significantly visible (if any there).

Anyway. All reactive parts and the demodulation of the signals (in a PI circuit) are highly time dependent.
A function f(t). One have to get rid of the variable t. ...

Only Aziz is not Betonkopf because he knows that we should make Fourier transform.

Für alle Betonköpfen (To all ConcreteHeads):

Now I know why you can't understand BROADBAND technologies of MINELAB and why since "insidetheweb" forum (over 12 years ago) you are wondering whether they are PI or not PI.

http://www.findmall.com/read.php?34,129198

The answer is that MINELABs machines are WIDEBAND because "broadband" is its synonym.
Patent US4506225 describes discrete signal processing at WIDEBAND technology. The classic PI technology is simply a specific case of the WIDEBAND technology, however with low efficiency and worst search head.

Frequency domain shows how to calculate target response because it is convolved with ground response, how to make excellent discrimination, how to eliminate AIR&GND signal.

http://www.geotech1.com/forums/showthread.php?19942-X-vs-R&p=163793#post163793

I'm waiting to see how you dear ConcreteHeads, using an oscilloscope and experimenting in time domain without Fourier transform will solve these problems. The only benefit of SPICE analysis in Time domain is that it shows how you waste time tumbling in the mist of convolution. Frequency domain lifts the mist.

Look at the block diagram to see that target response is distorted because it is convolved in Time domain with ground response. This block diagram is valid not only for Time domain. The difference is that there is multiplication in Frequenccy domain instead the math operation "convolution" in Time domain. Your oscilloscope can't make deconvolution.

Your motto is:
I hate to think, I prefer to experiment.

For those who prefer to think, I will place suitable information in the WIDEBAND thread.

[Carl-NC]

In my part of the ountry, North Carolina, the black sands as you call them are actually iron particles, many times I have gone panning and just pandown to the concentration of " black sand", take home the product and use a magnet to remove most of the "black sand".
The remainder is small sand particles ans small gold particlea on order of pin head size flatened out.

Just thought you might want to know.

Our mineral soil is ( I do not know for a fact) as strong as yours in the lower hemisphere, however we also have mineral rocks that are highly magnetic, some of these rocks are the size of grapefruit and when busted open contain powder of different colors, The locals and Indian tribes of North Carolina call them Paint rocks.

KM, the soil in NC (I lived there 21 yrs) doesn't come anywhere close to Oz. Even the reddest NC dirt doesn't have much viscosity effects. Oz detectorists sometimes refer to their soil as "corrugated iron." If you can get a hold of a sample, you'll appreciate NC conditions more.