Understanding Ground Balance

[Tinkerer]

The Ground Balance Enigma

Hi B^C,

Like you, I feel that a good ground balance is one of the most important parts of a metal detector.
I am considering options for the ground balance of the TINKERERS IB-PI.
Your advise, experience and knowledge would be much appreciated in the design of the ground balance.

Tinkerer

[simonbaker]

Everything that George Payne says in the article is correct.
As far as the TGSL is concerned, it responds to the rate-of-change of the sampled RX signal. If it was responding the DC level of the sampled signal, then it would be a non-motion detector, and you would need to occasionally adjust the threshold to account for DC drift. The GEB channel samples at the zero-crossing point. Which means that a change in amplitude can only be detected if the coil passes over a target which produces a phase-shift. Normal ground only causes a decrease in amplitude, but no phase-shift.

You are correct that nulling of the coils is not so important for the TGSL, I have already stated this numerous times, but it seems to fall on deaf ears, and several constructors are still trying to reduce the residual RX voltage to nV. The most important physical coil adjustment is the initial phase offset between TX and RX.

There is a nuance I still do not see your way. I don't see why the initial phase offset is critical. It seems to me it only establishes the steady DC voltage of the capacitor which integrates the synchronous detector output. Any target signal should still be additive based on its phase and magnitude. That's what I see from the equations describing the synchronous detector.

The George Payne aticle makes the assumption that ferrous targets cause the RX signal to phase-shift left, whereas non-ferrous targets cause the RX signal to phase-shift right. This is true for some detectors, but not for the TGSL. In the TGSL all signals cause a phase-shift to the left. (Some coil configurations can give phase-shifts to the right, and these will still work with the TGSL, but this is another story that I don't want to get into here.)

I think it is misleading to use the term "phase shift" for the TGSL detector, assuming "phase shift" means what you see if you look at the RX signal (which is composite of null signal and target signal) on an oscilloscope as a target is presented.
Yes, adding two signals with same frequency and different phase creates a composite signal that is "phase shifted" from either of the originals. The misleading part is due to the fact that the phase shift depends on the amplitude as well as the phases of the signals. If one is very big and the other small in amplitude, the phase shift of the composite is small relative to the big signal. Let's say we are trying to detect the small signal. It would seem that the big signal would "hide" the small signal by making its phase shift smaller. But the TGSL doesn't care how big the big signal is, even though it affects the apparent "phase shift" of the composite. The TGSL will see a voltage pulse from the small signal the same, regardless of the size of the bigger signal - so it does not measure "phase shift", rather it responds to magnitude and phase of each signal. Because it is a motion detector, the big constant signal is ignored. I can't see it otherwise unless someone shows some equations explaining it.

What GP is saying is that mineralized ground will keep pushing the RX signal to the left. In this case you need to offset the sample point in the GEB channel. Of course, if the ground mineralization changes rapidly you will get some false signals, so you then need to integrate the ground signal over time, and idealy provide some sort of automatic ground tracking. Detection of desirable targets can therefore be drastically reduced by the presence of highly minerlized soil.
The ground also causes a decrease in amplitude, which will likewise reduce the signal from any non-ferrous targets. Even though, in GP's article, the non-ferrous target causes a phase-shift to the right, and an increase in amplitude, it still has to overcome the amplitude reduction causes by the ground. Hence sensitivity is reduced.

I probably need different terminology to understand that idea. I know the TGSL rejects targets when the GEB channel and the DISC channel move in opposite directions, so ferrite makes us adjust the GEB phase detector to blank out ferrite, and that I think also weakens the good targets too.

I'll keep listening and trying to understand this better. I think it is valuable to keep discussing to develop good analogies and ways to understand it, since equations and computers often to give right answers that are hard to follow.

Cheers,
-SB

[Qiaozhi]

There is a nuance I still do not see your way. I don't see why the initial phase offset is critical. It seems to me it only establishes the steady DC voltage of the capacitor which integrates the synchronous detector output. Any target signal should still be additive based on its phase and magnitude. That's what I see from the equations describing the synchronous detector.

The assumption you are making is that the ground signal is a constant, so it doesn't matter if the output from the GEB sample gate is non-zero, as this signal goes into a discriminator that will only respond to changes in that signal. This would only be true if you could keep the coil at a constant height above the ground. Any non-mineralized ground causes the RX signal to decrease in amplitude as the coil gets closer to the ground. So any bobbing up and down of the coil will produce a change in the RX signal that will be produce false signals. Since non-mineralized ground does not produce a phase-shift, sampling at the zero-crossing point eliminates the ground effect.
The correct initial phase-shift is a requirement of the TGSL design. If you get the initial coil phasing incorrect, you cannot ground balance properly, and the discrimination "wraps round" the scale.

I think it is misleading to use the term "phase shift" for the TGSL detector, assuming "phase shift" means what you see if you look at the RX signal (which is composite of null signal and target signal) on an oscilloscope as a target is presented.
Yes, adding two signals with same frequency and different phase creates a composite signal that is "phase shifted" from either of the originals. The misleading part is due to the fact that the phase shift depends on the amplitude as well as the phases of the signals. If one is very big and the other small in amplitude, the phase shift of the composite is small relative to the big signal. Let's say we are trying to detect the small signal. It would seem that the big signal would "hide" the small signal by making its phase shift smaller. But the TGSL doesn't care how big the big signal is, even though it affects the apparent "phase shift" of the composite. The TGSL will see a voltage pulse from the small signal the same, regardless of the size of the bigger signal - so it does not measure "phase shift", rather it responds to magnitude and phase of each signal. Because it is a motion detector, the big constant signal is ignored. I can't see it otherwise unless someone shows some equations explaining it.

In reality this is not a problem. The TGSL is looking at the rate-of-change of the RX signal, and not at the absolute value. In theory the sensitivity could be increased somewhat if the nulling was closer to zero, but in practice the RX waveform then deviates from a pure sine wave. This causes problems with GEB sampling and hence causes falsing. That is - assuming you could get the TGSL GEB sample point in the correct position (at the zero-crossing point) which you cannot do if you completely null the coils.

I probably need different terminology to understand that idea. I know the TGSL rejects targets when the GEB channel and the DISC channel move in opposite directions, so ferrite makes us adjust the GEB phase detector to blank out ferrite, and that I think also weakens the good targets too.

I'll keep listening and trying to understand this better. I think it is valuable to keep discussing to develop good analogies and ways to understand it, since equations and computers often to give right answers that are hard to follow.

Cheers,
-SB

The GEB channel creates a positive signal for all metal targets, but not for non-mineralized ground. Whereas the DISC channel creates a positive signal for non-ferrous targets, and a negative signal for ferrous targets. Therefore, when both GEB and DISC are positive, there is an audio output. Using GEB does not weaken the good targets.

[nick_f]

Hi B^C,

Did you post about Minelab mods on http://goldprospecting.invisionplus....showtopic=3367
I am curious if the mods really work and if you started to do them.
Not that I have a M/L detector, just curious if such a major improvement is possible. You also mentioned some videos.

Regards,
Nicolae

[golfnut]

"The correct initial phase-shift is a requirement of the TGSL design. If you get the initial coil phasing incorrect, you cannot ground balance properly, and the discrimination "wraps round" the scale."

Often stated as with 20degrees -- Is this +/-20 So 40deg window. ??

I wnow my machines are non standard and I pay little attention to phasing to a standard - I guess SB and I are saying if you were 50deg off, then you require to slide both Disc and GEB gating by 50deg and it will work again.

Obviously you may need a larger pot to give the extra control range or the next size cap or somesuch, so your coil isnt scrap if 30deg.

Clearly if you want the result to be to the design you are forced to make it to the design.

S

[simonbaker]

"The correct initial phase-shift is a requirement of the TGSL design. If you get the initial coil phasing incorrect, you cannot ground balance properly, and the discrimination "wraps round" the scale."

Often stated as with 20degrees -- Is this +/-20 So 40deg window. ??

I wnow my machines are non standard and I pay little attention to phasing to a standard - I guess SB and I are saying if you were 50deg off, then you require to slide both Disc and GEB gating by 50deg and it will work again.

Obviously you may need a larger pot to give the extra control range or the next size cap or somesuch, so your coil isnt scrap if 30deg.

Clearly if you want the result to be to the design you are forced to make it to the design.

S

This is a subject I revisit many times. I am currently of opinion, similar to original thoughts, that "ground signal" has a true phase and is not simply a modulation of whatever null signal you have -- well, ferrite objects at least. This would imply null signal phase is not so crucial. However, null signal phase is potentially important for any effect that modulates it, so we don't want it in the "detection range" by principle. Davor points out that asymmetries in the half-wave synchronous detector can modulate the null signal -- extent of that is worth putting on the "to-test" list.

I have to add that I believe what causes "abnormal" null signal phases is due mostly to capacitive linking through the cable and shielding, etc -- which is why it isn't really affected by ferrite and most magnetic/conductive ground effects. I'm uncertain whether any "capacitive" effect to soil would amplitude-modulate the null signal "capacitive" components significantly.

I'm sure it's a good idea to get the null signal phase in the "normal" regions if possible for good measure!

I did a whole bunch of tests with different shields to show that GB settings are not affected by null phase when nulling with ferrite. But this stuff can be complicated so I should yield to conventional wisdom and not beat the horse. I just think many people have very usable MDs with some oddball null phases for most situations.

-SB

[sinclairuser]

I totally disagree with the sweeping statement about all detectors being tweaked for gold, there are more places in the world where this is actually counter productive than helpfull, the uk is one such place, in the uk as a whole detecting natural gold does not work as the deposits are so small a detector is useless.
The thing about ground balance that some dont get, is the sheer task of creating a workable GB that works everywhere, a lot of MD makers only want one model that sells everywhere, this is the real problem.
Take a look at this
building a workable GB just for the uk in a single detector with just these bedrock/soil combinations would be a task in itself.
Adding in the whole earth geology is frankly ridiculous, GB has to be more local otherwise its pointless.
ozzy soils or even californian for that matter are separate cases, and need to be addressed as such.

[Davor]

Practically all projects that I've seen schematics of have a GB regulation somewhere to about 60° max. From this span you must deduct some 20° typical coil shift, and you are left with some ~40°useful GB range. It is enough for most soils. You are in trouble only on salty terra rossa and beaches.

Perhaps the worse problem is the fact that practically all detectors have GB channel used as one of the two discrimination criteria, which is wrong in case of large GB excursions. When GB is cranked to it's max, say, due to the salty terra rossa, you are actually accepting counter-phase minerals as valid coloured metals, and you have a feeling that your rig is gone bananas, and that there is no way of proper ground balancing.

The solution would include a separate GB channel.