Ground Balance Theory

[Prospector_Al]

Hi Allan,

I am not in agreement with the argument re anomalous responses. Most GB methods at the moment work on weighting a sample of the decay curve to equal a sample at an earlier or later point on the same curve. Since the decay curve from one viscous material to another does not change (i.e. it is ideally always a 1/t^-1.00) law) then it does not matter if the matrix is not homogeneous. An included quartz rock, or a void, will only alter the amplitude and give no change in indication. Exactly the same as when the coil is bounced up and down to check for ground balance, the large changes in amplitude have no effect. In the real world there seems to be a small variation in the exponent. For four different materials I measured 1.03, 1.05, 1.06 and 1.07 as negative exponents. Hope to refine this later to measure at three decimal places. I have found with the above types of GB that small corrections are needed in the field when passing from say a red clay to a brown ironstone. It may be the result of a small change in the slope, or more likely the result of non-ideal TX pulses in a metal detector. With the foregoing viscosity measurements I make sure that the TX current pulse is as close to rectangular as possible.

Eric.

The GB systems for CW and PI systems are not analogous. When you bounce the coil up and down, you are looking for a gate position where a zero crossing occurs. Then, the portions above and below the base line cancel.

The nature of the ground signal is irrelevant. When the sampling gate is centered over a zero crossing you get a zero output, whether the signal is purely reactive or mixed with a viscous component. However, when the relative magnitudes of the reactive and resistive components change, that cause a phase shift of the received signal and the gate is no longer in the correct position.

I've read reports that in some areas, you only need to walk a few feet for the balance to go out of whack. This would not occur if the ground were homogeneous. But you're right, the ampltude variation does not matter--the portions above and below the base line change in synchrony.

At any rate, I think discussing these matters is important. It will lead to a better understanding of the problems and it may even lead to the Perfect Ground Balancing System,

Regards,

Allan

[dougAEGPF]

Hi Allan,

I am not in agreement with the argument re anomalous responses. Most GB methods at the moment work on weighting a sample of the decay curve to equal a sample at an earlier or later point on the same curve. Since the decay curve from one viscous material to another does not change (i.e. it is ideally always a 1/t^-1.00) law) then it does not matter if the matrix is not homogeneous. An included quartz rock, or a void, will only alter the amplitude and give no change in indication. Exactly the same as when the coil is bounced up and down to check for ground balance, the large changes in amplitude have no effect. In the real world there seems to be a small variation in the exponent. For four different materials I measured 1.03, 1.05, 1.06 and 1.07 as negative exponents. Hope to refine this later to measure at three decimal places. I have found with the above types of GB that small corrections are needed in the field when passing from say a red clay to a brown ironstone. It may be the result of a small change in the slope, or more likely the result of non-ideal TX pulses in a metal detector. With the foregoing viscosity measurements I make sure that the TX current pulse is as close to rectangular as possible.

Eric.

Paltoglou in US6586938B1 finds a substantial variation in the exponent (and pulse length dependence) over WA laterites and some ground in Victoria when using his detector.
dougAEGPF

[Ferric Toes]

The GB systems for CW and PI systems are not analogous. When you bounce the coil up and down, you are looking for a gate position where a zero crossing occurs. Then, the portions above and below the base line cancel.

The nature of the ground signal is irrelevant. When the sampling gate is centered over a zero crossing you get a zero output, whether the signal is purely reactive or mixed with a viscous component. However, when the relative magnitudes of the reactive and resistive components change, that cause a phase shift of the received signal and the gate is no longer in the correct position.

I've read reports that in some areas, you only need to walk a few feet for the balance to go out of whack. This would not occur if the ground were homogeneous. But you're right, the ampltude variation does not matter--the portions above and below the base line change in synchrony.

At any rate, I think discussing these matters is important. It will lead to a better understanding of the problems and it may even lead to the Perfect Ground Balancing System,

Regards,

Allan

Yes, my experience is just with PI where the reactive susceptibility response does not occur.

Eric.

[dougAEGPF]

This paper also finds that the exponent varies
EDDY CURENT METAL DETECTORS – PULSE VS. CW
Pavel Ripka* – Adam Lewis**
They found ii was from -1.1 to -1.3
srcid=ADGEESi91npvtBa6oWEQiWv6e224PNsFYLkBI_g_30VK HKAlWIzhr_hNSrK4XGjtW5nd6kOdTCNlGhmvWBuIrsp51X5Blb QgiO7vJYyW64PXEbHRn4akf7zT9scdvn5_SLMHHw5QtJs9&sig =AHIEtbQ38oGwJiCHqjAI79hi-348a_3boA
dougAEGPF

[Aziz]

Basic Math

Hi all,

be careful, when you provide formulas and parameters.

f(t, a) = 1/(t^(a)) = t^(-a)

for a = -1.0 ->
f(t, a=-1.0) = 1/t^-1.0 = t^(1.0) = t


Exponent must be positive to get the magnetic relaxation (induction) decay.

Play with the following site:
http://www.wolframalpha.com/input/?i...+for+0%3Ct%3C5

Cheers,
Aziz

[Aziz]

Basic Math II

f1(t): Exponential decay (Target eddy current induction response)
f2(t): Magnetic relaxation induction decay (for instance 1/t^1.03)

1. Which function does generally reach zero faster?
2. Affect on the frequency domain?

Aziz

[Aziz]

Frequency Domain

Hi all,

you guys are sure wondering, why I try to steer you into the other direction.
What's the purpose of changing into the frequency domain?
1. Better understanding of the matter
2. Prove of lossy ground balance method (infamous detection hole)
3. Prove of reduced detection range for large time constant (TC) targets (infamous detection hole)
(4. Exposing a patent-troll (thief) claiming to own low-, mid- and high-frequency range. )
(I hate to do that but I'll do it, when LabGreed forces me to do this. )

Cheers,
Aziz

[Ferric Toes]

Hi all,

be careful, when you provide formulas and parameters.

f(t, a) = 1/(t^(a)) = t^(-a)

for a = -1.0 ->
f(t, a=-1.0) = 1/t^-1.0 = t^(1.0) = t


Exponent must be positive to get the magnetic relaxation (induction) decay.

Play with the following site:
http://www.wolframalpha.com/input/?i...+for+0%3Ct%3C5

Cheers,
Aziz

You are correct. Will be more careful in future - no double negatives.

Eric.

[Ferric Toes]

Hi all,

you guys are sure wondering, why I try to steer you into the other direction.
What's the purpose of changing into the frequency domain?
1. Better understanding of the matter
2. Prove of lossy ground balance method (infamous detection hole)
3. Prove of reduced detection range for large time constant (TC) targets (infamous detection hole)
(4. Exposing a patent-troll (thief) claiming to own low-, mid- and high-frequency range. )
(I hate to do that but I'll do it, when LabGreed forces me to do this. )

Cheers,
Aziz

My turn now . Do you mean you mean PROOF? If you already have the evidence, then you have proof. If you don't have the evidence, then you can set up an experiment, or prove something mathematically. Proof is a noun, prove is a verb.

3) above. Do you mean reduced detection range as a fundamental limitation of PI, or only if GB is applied?

Eric

[Aziz]

My turn now . Do you mean you mean PROOF? If you already have the evidence, then you have proof. If you don't have the evidence, then you can set up an experiment, or prove something mathematically. Proof is a noun, prove is a verb.

3) above. Do you mean reduced detection range as a fundamental limitation of PI, or only if GB is applied?

Eric

Ooops!, I have meant proof of course. I'm not that perfect and smart.

Hi Eric,

the fact is very obvious. One even don't need the frequency domain to see its obviousness.
If you have two kind of data (desired, undesired effect) and they superimpose each other, which can't be distinguished, what's the result?
Loss of desired information.

Aziz

[Ferric Toes]

This paper also finds that the exponent varies
EDDY CURENT METAL DETECTORS – PULSE VS. CW
Pavel Ripka* – Adam Lewis**
They found ii was from -1.1 to -1.3
srcid=ADGEESi91npvtBa6oWEQiWv6e224PNsFYLkBI_g_30VK HKAlWIzhr_hNSrK4XGjtW5nd6kOdTCNlGhmvWBuIrsp51X5Blb QgiO7vJYyW64PXEbHRn4akf7zT9scdvn5_SLMHHw5QtJs9&sig =AHIEtbQ38oGwJiCHqjAI79hi-348a_3boA
dougAEGPF

Clues for the discrepancies can be seen in the paper "Magnetic susceptibility and viscosity of soils in a weak time varying field". It appears in Google but not a free download that I could find. I believe my method overcomes the discrepancies seen before, but I am not going into details on forums. Happy to post results though. One result in this morning - on a 10g sample of Oz ironstone the amplitude of the viscosity decay increased 10% for a 10degC increase in the sample temperature.

Eric.

[Aziz]

Clues for the discrepancies can be seen in the paper "Magnetic susceptibility and viscosity of soils in a weak time varying field". It appears in Google but not a free download that I could find. I believe my method overcomes the discrepancies seen before, but I am not going into details on forums. Happy to post results though. One result in this morning - on a 10g sample of Oz ironstone the amplitude of the viscosity decay increased 10% for a 10degC increase in the sample temperature.

Eric.

Free PDF is available here:
(read the instructions)
http://australianelectronicgoldprosp...varying-field/

Aziz

[Ferric Toes]

Ooops!, I have meant proof of course. I'm not that perfect and smart.

Hi Eric,

the fact is very obvious. One even don't need the frequency domain to see its obviousness.
If you have two kind of data (desired, undesired effect) and they superimpose each other, which can't be distinguished, what's the result?
Loss of desired information.

Aziz

Depends on the relative S/N. S is the decay from a large non-ferrous object with a TC of 1mS. N is the viscous ground signal which we consider as noise. If the sample delay is 100uS the ground signal is reduced by a factor of 10 compared to sampling at 10uS. You still have 90% of your object signal though. Proved this by plotting the decay of a US Silver $ against Red hill soil with the same starting amplitude at 10uS. This shows improvement of S/N at later times for long TC targets.

Eric.

[mikebg]

Here is a better operating link for paper mentioned by Doug in post #109
http://iris.elf.stuba.sk/jeeec/data/pdf/8s_106-43.pdf

[moodz]

the hot rocks mag vector will be aligned with the exciting mag field. The induced eddy current in target will oppose exciting vector .... works for me :-)