Announcement

Collapse
No announcement yet.

THE PERFECT GROUND BALANCE

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • THE PERFECT GROUND BALANCE

    A very important factor for deeper detection depth is the ground balance.

    An ideal ground balance would mean that the detector only sees the targets.

    It would not see the minerals in the ground. This means that one could lift the coil off the ground 6 inches or put the coil right on the ground and the output would stay 0. And this should work every where, even in the worst mineralized soils.

    It would not see the earth's magnetic field. The Earth is surrounded by a magnetic field of about 0.2 to 0.5 Gauss. Near the equator, the field lines are near parallel to the surface of the ground.
    In higher latitudes, the field lines are nearer 45 degrees. When we move our coil through these field lines, we generate a signal. The signal amplitude is proportional to the amount of field lines we cross and the speed at which we cross the lines.

    Large coils at high speed = high amplitude signal.

    What is the usual amplitude of the signal? About 1uV is the number I find in texts. Is this right? what are the parameters that produce this number?

    There are other factors.

    The general magnetic field lines go in the direction North-South. However, the local field lined close to the ground, may be different.
    Close to the ground, the field lines may be at 90 degrees to the ground. They also may vary a lot in angle as well as field strength, within 1 or 2 meters distance. Patches of magnetic minerals and hot rocks can do that.

    A perfect ground balance would have to compensate for all that, so that the output is not affected and stays at 0.

    As AZIZ states, the perfect ground balance is the "Holy Grail" of metal detecting.

    How can we find the perfect ground balance?

    If the combined brain power of the forum works on the solution, we can find a solution.

    Let's go for it!!!!!!

    Suggestions?????

    Should we start with looking at a few existing ground balance schemes?

    Tinkerer

  • #2
    Originally posted by Tinkerer View Post

    As AZIZ states, the perfect ground balance is the "Holy Grail" of metal detecting.

    How can we find the perfect ground balance?
    Probably by building detector that even do not need a ground balance.

    There was approach in MD story very near to this request.

    Comment


    • #3
      I really like Whites half sine approach. Being able to isolate the permeability component is a big step towards reducing the problem to something tractable. I believe the effect on resonant frequency of the LC tank is also going to be directly related to the permeability so it'll help with that too.
      I have two questions:
      Can a variation of this method be done with unsymmetrical TEM wave form? Is it as simple as scaling of the second ground balance sample to compensate for the higher dI/dt of the falling section of the waveform, or is the relationship between the two samples actually dependent on the specific ground conditions. ie. unpredictable.
      My second question is, if it is possible does that actually put us in the clear as far patent violation goes?

      Comment


      • #4
        Originally posted by Midas View Post
        I really like Whites half sine approach. Being able to isolate the permeability component is a big step towards reducing the problem to something tractable. I believe the effect on resonant frequency of the LC tank is also going to be directly related to the permeability so it'll help with that too.
        I have two questions:
        Can a variation of this method be done with unsymmetrical TEM wave form? Is it as simple as scaling of the second ground balance sample to compensate for the higher dI/dt of the falling section of the waveform, or is the relationship between the two samples actually dependent on the specific ground conditions. ie. unpredictable.
        My second question is, if it is possible does that actually put us in the clear as far patent violation goes?
        Patent violation? Are we building or offering for sale a commercial product? Or are we just investigating old and new methods of ground balance for academic interest.

        For academic interest, we can do anything we want. We can copy any patented method. We can build any style and method as long as it is only to be used for experimenting and testing.

        You like the recent White's patent ground balance method? Can you describe to us how exactly it works? Then we can see if it fits on the TEM method too.

        Specific ground conditions are easy. The problem is with the changing conditions. The rate of change of the conditions and the unpredictability of this rate of change.

        Tinkerer

        Comment


        • #5
          I think you posted a correct question. You can't distinguish a hot rock from a target, unless of course you discriminate it as such.

          As I see from prior art, discrimination is typically realised as a "channel" that triggers a kind of alarm, using a threshold to numb a detector down below a trigger happy operation. And with no quantitative relationship with a target size. And plenty of delay. There is a huge range in Rx signal levels encompassed by a simple "beep", and frankly GB seem of lesser importance with target signal detection being a few orders of magnitude too deaf.

          Next, I've seen some of the old designs that deal with GB at Rx coil. That would be as if notching it out at the Rx entry. This approach will give you extra room for amplification. I think it is the way to go.

          As I seen it in White's patent they sample for GB at current maximum, a tad to the left and right. Both samples being on opposite sides of a zero voltage, when balanced give you a voltage zero crossing moment. So a crucial detail for ground sampling is a voltage zero crossing. TEM signal, being very quick at voltage zero crossing seem a bit unpractical.

          I have some ideas, but not mature yet. Some of them deal with doubly balanced coils - something I have no experience at all.

          Comment


          • #6
            Davor,
            thanks for the feedback. What is a doubly balanced coil?

            With the TEM method and an IB coil, we have an excellent opportunity for FE definition. Most hot rocks and extreme ground response is due to the magnetic and/or FE content.

            What if we start with that? Define the FE or reactive response. If we can control that, we should be able to control a large part of the ground response.

            If we go in this direction, we might just as well treat each one of the ground related responses individually. Once we have each individual part of the ground response under control, know how to control them, we can then see how to lump the corrections again.

            Tinkerer

            Comment


            • #7
              Originally posted by Tinkerer View Post
              What is a doubly balanced coil?
              Well. that's what I'm trying to find out. Having NO previous experience in MD is a challenge, don't you think?
              Anyway, my view of GB is that for many reasons proximity of earth screws up most of the balancing techniques, ergo ground signal. Most probably the most of it is cancelled by an appropriate screen or a balanced Rx frontend. In any case, unbalanced Rx is a Hyde Park for every capacitively induced feedback to freely express themselves. Besides, even +inside -outside overlapping balance schemes get a blow in ground proximity. Everything else, including hot rocks is a free range for detections of any kind.

              Shortly, my guess is that we'll get much better off if ground proximity is nulled by some devious double balancing scheme where neither Tx see any of the Rx coils, nor Rx see any of the Tx coils. My first guess is - four leaf clover design. If both Tx coils are arranged in counter phase at x axis, and both Rx coils are arranged in counter phase at y axis you get a double balanced coil design. Neither of the coils of Rx will see Tx pair, and even if a whole shebang is somewhat tilted the other coil will partly compensate.
              Funny thing the 8 configuration - it is a part of a torus, a structure that ignores influences outside of it. Put another one rotated 90° and you get perfectly orthogonal structures. With balanced input and output - earth proximity should be eliminated, ergo a first approximation of earth signal is eliminated.

              Comment


              • #8
                Since you have no previous experience with MD design and there are probably many other readers in the same situation, I will try to explain the most common problems involved with ground balance.

                Hopefully somebody pitches in and expands or corrects my feeble explanations.

                The ground is a large target. It responds to the TX pulse. Mild ground gives a low amplitude response. The response of hot ground can be like a junk of iron.

                When we swing the coil, the distance to the ground is not even. Most of the time, on flat ground, at the center of the swing, the coil's distance to the ground is about 2". At the end of the swing, it is maybe 6". Changing the distance to any target, changes the amplitude of the response.

                The motion filter is also affected by this swing. Setting the motion filter for a certain sweep speed, for example the standard sweep speed of 1m/s reduces the perceived response for changes in the ground that are slower than that.
                However, at the end of the swing, there is a dead time, where the coil moves hardly at all. The motion filter can then not cope.
                There is also a time delay between the target response and the audio output. This can make that the problem of the sweep speed appear with a delay, more towards the center of the sweep.

                We generally use HP filters and and integrators (LP). If we have a high PPS or pulse repetition rate, we can integrate many signals and still end up with only about 100ms delay. With a coil diameter of 30cm, it puts the target position way out of the center of the coil.
                With a 1 meter coil, it would be quite good, specially since large coils are swept slower.

                I believe NOKTA, advertise a detector that has an accelerometer incorporated to help fixing the problem of sweep speed X motion filter.

                Enough for now. I hope some of the expert detectorists will add their feedback here.

                Tinkerer

                Comment


                • #9
                  I think everything you described fits perfectly to my picture of MDs, their use and motion filters. Thank you.

                  I'm still convinced that the most of the ground problems are related to proximity of a conductive material (ground) rather than the ground's magnetic response. Otherwise the motion detectors would cancel out real targets as well, yet they hopefully just iron out the slow variations due to near/far problem when waving a coil against the ground, and eventual large masses in the ground.

                  Comment


                  • #10
                    Originally posted by Tinkerer View Post
                    Patent violation? Are we building or offering for sale a commercial product? Or are we just investigating old and new methods of ground balance for academic interest.

                    For academic interest, we can do anything we want. We can copy any patented method. We can build any style and method as long as it is only to be used for experimenting and testing.

                    You like the recent White's patent ground balance method? Can you describe to us how exactly it works? Then we can see if it fits on the TEM method too.

                    Specific ground conditions are easy. The problem is with the changing conditions. The rate of change of the conditions and the unpredictability of this rate of change.

                    Tinkerer
                    Its not neccessarily true that you have the right to construct a patented device if its not for commercial use. According to the fountain of all knowledge, wiki:

                    'The rights conveyed by a patent vary country-by-country. For example, in the United States, a patent covers research, except "purely philosophical" inquiry. A U.S. patent is infringed by any "making" of the invention, even a making that goes toward development of a new invention—which may itself become subject of a patent.'

                    Though its probably fair to say given the cost of enforcing a patent most campanies wouldn't bother pursuing individuals unless they represent a significant commercial threat.
                    Never-the-less I still think a good open source design should be free from patent violations. Especially if its going to be discussed\promoted on a forum run by someoone working for the company that owns the patent...

                    As far as describing how it all works in detail your probably better off reading the whole patent (which really is very well written, far clearer than most) than listenening to my half-assed attempt at an explanation. But I think the key improvements is firstly a method that can be used to isolate the permabilty component. Secondly the discovery that the ground remenant component can be to some extent predicted by using the permabilty component in that it doesn't change too quickly from one piece of ground to the next.

                    I'm sure Carl could explain it a lot better although he is probably constrained to some extent.

                    Comment


                    • #11
                      There is nothing that Carl can do to diminish value of the patent. It is solid, well written, obvious, and at long last original. You can't infringe a patent by talking about it, even by explaining it - no matter how precisely. Such activity could only improve its future by awe and envy.

                      As for the patent use, it is OK to reproduce patented technology for philosophical, amusement and experimental use, where experimental is a bit tricky because the same expression holds two distant ends of meaning. Something may be a principle under experimentation (no infringement) and also a tool for performing experiments (infringement). So it is a bit slippery, and does not go nicely with (commercial) copycats.

                      So far as my activities here go, I'm not into copying of anything. Producing a spice model or two is deeply in philosophical grounds, and is here only for exploring the broad scope of MD technology and my amusement. Heck - I'm laboratory-less at the moment... dangerous as a toothless taipan snake.

                      Now, back to the ground balance. I think there are 3 mechanisms collectively encompassed by this notion:
                      1. design flaws (that affect detection) due to the ground proximity, metallic decorations etc.;
                      2. coil inductance and its variation due to replacing one search head with the other, ground proximity, bumps and bruises;
                      3. ground permeability influence on targets, e.g. tau change.


                      The very point of permeability compensation is establishing a time reference for subsequent samples in case the very reference point is prone to floating, e.g. a free running sine oscillator. GB in White's patent takes care of LC tank time constant as it is prone to changes in case of replacing a Tx coil, and in lesser degree to ground proximity. This enables their apparatus to timely trigger a sine current shortening, thus launching a PI pulse at a precise moment (bravo!), however it has a minute influence on target discrimination because PI pulse discontinuity is a perfect time reference per se.

                      The way I see it, target tau IS influenced by ground permeability, BUT it does not change with a manic detectorist waving wildly with a search coil. Whatever a mechanism of permeability is involved, be it ferrous dust, high mineralisation, rain, ground waters ... it affects all targets more or less uniformly by relative shortening or prolonging their tau. So in extreme cases "channel" discrimination type detectors are in trouble because the absolute ground correction will not shift their time constant - it is a fixed delay.

                      I think I know how it can be modeled in spice.

                      Shortly, ground balance is a principle to blame for various imperfections, but not the ground influence on target tau.

                      Even shorter, fix your design first, cry about GB later.

                      Comment


                      • #12
                        Ground balance consists of resolving the biggest target ... The ground ...mitigating this signal to leave the rest of the targets .. If there are any.

                        PS patents with poorly drawn technical diagrams ( ie hand drawn ) are unprofessional ... If any of my staff handed in a patent like that they would fail their annual review.

                        moodz
                        Last edited by moodz; 03-03-2012, 10:26 AM. Reason: Typo

                        Comment


                        • #13
                          Earth field compensation

                          One part of the ground problem is the Earth's magnetic field.

                          As we move the coil across the Earth's field, we generate a current in the coil. A large coil generates a stronger current because it cuts through more field lines.

                          Extreme signal amplification exacerbates the problem.

                          This problem is not felt everywhere. Near the equator, over mild ground it does not exist. It appears over hot, magnetic ground, because the magnetic ground changes the angle of incidence of the field lines more to the vertical.

                          The same happens over magnetic hot rocks.

                          The faster the movement of the coil through the field, the stronger is the current that is generated. A quick tilt of the coil makes the coil cut through perpendicular field lines fast and therefore generates a strong signal.

                          If we were to move the coil slowly, in a uniform direction, the current would be a steady offset that could be easily compensated. The Earth field problem would still appear over magnetic hot spots.

                          Magnetic hot spots also cause a reactive signal response to the TX pulse.

                          What would happen if we were to phase shift the reactive response 180 degrees to the Earth field response, scale and sum it to the Earth field response?

                          Any suggestions of how to fix the Earth's field problem??????

                          Tinkerer

                          Comment


                          • #14
                            Originally posted by Tinkerer View Post
                            Any suggestions of how to fix the Earth's field problem??????
                            Bipolar pulsing.

                            Comment


                            • #15
                              Originally posted by moodz View Post
                              Ground balance consists of resolving the biggest target ... The ground ...
                              I'm afraid it doesn't. Ground balance is merely a reference point for applying coherent multiplication. With IB VLF it is basically a 90° local oscillator signal. Nothing more and nothing less.
                              Originally posted by Tinkerer View Post
                              One part of the ground problem is the Earth's magnetic field.

                              As we move the coil across the Earth's field, we generate a current in the coil. A large coil generates a stronger current because it cuts through more field lines.
                              Coil pickup due to the swinging of a search coil, and phase shift due to the ground proximity, a.k.a. GB are completely different animals. While induction due to Earth magnetism and swinging falls into a ~10Hz range, VLF is, well, somewhat higher than that. I can imagine poor designs that might produce such effect that they trigger a phase change just by Earth magnetism induction, but it is not that likely. Ground proximity produces much stronger effects:
                              1. change of resonance due to inductance change, hence phase change
                              2. change of amplitude with possible phase change due to PWM

                              One possible way of PWM mitigation could be a symmetric design that cancels even harmonics, or forced oscillation controlled by crystal or something else. Using such approach and 4xf base oscillator you don't have to extract GB, you make it with a couple of flip flops.
                              Originally posted by Carl-NC View Post
                              Bipolar pulsing.
                              Right on. That enables both symmetry and an accurate zero crossing, which is a perfect way of extracting GB phase.

                              I made some drawings that may be amended in case I missed something or if some ideas can be presented more straightforward.

                              First row presents an VLF exciter, IB coil with partially balanced out Tx signal, a target response, and a composite of target + IB reduced Tx signal. Whole shashlik goes into a Rx.

                              Second row shows a concept of coherent multiplication (mixing) of Rx signal with GB normalised signal. As mixing usually happens with analog switches the amplitude is 1. Due to the troubled design of a typical detector, a perfect 90° GB is spoiled by some phase error that is propagated further to other "channels" of a typical discriminating Rx, but I didn't go that far - I was happy with GB. The resulting signal is actually a slowly changing DC whose amplitude is an ideal one, + and - the error. I hope the idea of error contributions are clear enough.

                              A third row shows what happens when there are no real targets and there is an error in induction balance of a Rx coil. Due to the phase error we get some small resultant signal that disables us from going any deeper in weak signal exploration.

                              Conclusions:
                              • if you have a perfect GB you can live with some error in IB
                              • if you have a perfect IB you can live with some error in GB
                              • with unchanged GB phase error and unchanged IB error you can't go any deeper because your dynamic range remains the same - regardless of extra juice you pour into your Tx. This problem is indicated by a vector described as residual IB x GB error.

                              Please note how target angle is related to the target tau
                              Attached Files
                              Last edited by Davor; 03-03-2012, 07:29 PM. Reason: amended a picture

                              Comment

                              Working...
                              X