Announcement

Collapse
No announcement yet.

A Single Coil TEM field metal detector with simultaneous transmit and recieve.

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

  • #31
    Originally posted by moodz View Post

    whether in simulations or real circuits ... resistance is does not affect current sources ... we just dissipate more power.
    A 1.0000 amp current source is a 1.0000 amp current source through 1 ohm or 100 ohms.
    This is undoubtedly true, but the higher the resistance of L1, the smaller current will fluctuate in L2

    Comment


    • #32
      Hi Paul, looks interesting. Any chance you can post a YouTube with it running and showing us how it reacts on the oscilloscope to different objects like ferrite core, aluminium, lead. I guess on the raw input, unless you have a preamp strapped to it now and for a complete cycle, thanks.

      Comment


      • #33
        Originally posted by mickstv View Post
        Hi Paul, looks interesting. Any chance you can post a YouTube with it running and showing us how it reacts on the oscilloscope to different objects like ferrite core, aluminium, lead. I guess on the raw input, unless you have a preamp strapped to it now and for a complete cycle, thanks.
        looking at one of the mosfet gate drives and the raw RX signal.

        Blue = no target.
        Orange = Iron
        Green = aluminium

        No amplification. Pulse rate = 10 khz. ( Scope says 20 khz but there are two magnetic transitions per cycle )
        Even though the mag pulses are bipolar the RX signal come out as unipolar ... which is nice.

        I cant really post a youtube as there is a factor of demonstrating features in the public domain that are the subject of patents.


        Click image for larger version  Name:	image.png Views:	0 Size:	1.64 MB ID:	448119​

        Comment


        • #34
          Thanks for that. I think I still have the circuit built from when I was playing with the bipolar alternate TX that you posted. So just add a few extra bits and see what happens....

          Comment


          • #35
            Hi all,

            I for one would still run a concentric co-planar IB coil for this transmitter.

            This is the reason for:
            Two channel processing. TX is the main target channel (including ground noise).
            A small RX coil with a TX bucking coil is the ground channel. It detects 1000 times more ground noise as the coil coupling coefficient k will be affected by the ground mineralization.
            This ground channel will be used to clean the TX main channel (removing ground noise).
            The TX channel could be integrated over a very large window time (single integrating window).

            We don't need much inductance for the RX coil. RX and TX bucking coil could be really small size too. The TX coil will still pick up deep gold.

            Now go for it!

            Comment


            • #36
              Two months ago I was involved in experiments for the Volvo car industry with a friend who is working on a patent for wireless energy transfer between robots. And I wasn't committed to the detectors. But I was encouraged by this thread using a mono coil for transmission and reception with discrimination decided to repeat my experiment on the GT for which the Altra detector is fully functional with a DD coil works perfectly in the field. So I can't put an oscillogram against it! But I used the GT as a platform and changed the input part and some other things on the basic schematic and attached my mono coil which is wound in a very specific way, nothing like that I haven't noticed anyone mention anywhere. The result is on the video, in the next few days I will set aside to turn off the double tone when the iron is crossed and maybe replace a few other things so that in a few days I could get a fully usable detector with a mono coil and bipolar transmission that has excellent discrimination and ground rejection on one channel.https://youtube.com/shorts/iLerw0GsR...Xi6ea7g0cDw9OG
              Attached Files

              Comment


              • #37
                Originally posted by moodz View Post
                No amplification. Pulse rate = 10 khz. ( Scope says 20 khz but there are two magnetic transitions per cycle )
                I consider the detection edges when quoting the pulse rate. In this case, it's really an effective 20kHz pulse rate.

                Even though the mag pulses are bipolar the RX signal come out as unipolar ... which is nice.
                Does this mean we need to re-introduce EFE samples? Swing a big magnet and see what happens.

                Comment


                • #38
                  Originally posted by Carl-NC View Post


                  Does this mean we need to re-introduce EFE samples? Swing a big magnet and see what happens.
                  I would say no ... the coil is reversed on each edge so the EFE induced voltage is too.
                  Waving a powerful magnet seems to confirm.

                  Comment


                  • #39
                    Originally posted by moodz View Post
                    Before any "big reveal" ... you should all read up on this Minelab patent : https://ipsearch.ipaustralia.gov.au/patents/2023204592

                    Minelab are trying to lockup bipolar pulse induction till 2043.
                    What is interesting if you follow this document trail you will see at that site that Nokta has raised an objection to that patent since it would appear that Minelab is possibly trying to patent the idea of a controlled constant current bipolar transmit scheme.
                    Checkout the correspondence paper trail ( the lawyers are finding the gold here )
                    It is every companies right to protect their own IP however reading the patent seems to indicate they are patenting the method of producing bipolar constant current transmit scheme ( eg using feedback loops, current sensing, controlled high voltage and low voltage connections across a coil at various timings ...etc etc adapted and arranged. )
                    It is probable that the latest GPZ8000 model utilises this method. However it does require a separate recieve coil hence the horribly complex donut coil required on the GPZ8K.

                    Food for thought ...
                    Nokta objection to the patent is 24 pages and i am told is being backed financially by the Turkish Gov. Nokta have lodged their own patent for their new PI detector the Magnetar 9000 which is ZVT based.

                    Comment


                    • #40
                      After thinking about it some more, that's what I expected. EFE is common-moded on the entire waveform, so if it causes the +peak to increase to +peak+Δv then the -peak should also increase to -peak+Δv. Reflected to the load resistor, you should see a +Δv after the rising edge and -Δv after the falling edge. When averaging multiple sample, EFE disappears.

                      Comment


                      • #41
                        Btw,

                        the coil need not to be IB type. Mono coil as TX, a smaller GB coil placed in the center position (seperate RX) with a few turns should do the GB channel for GB info. The GB coil would detect during the high dI/dt the GB info. There wont be high voltage on the GB coil.
                        GB response is 1000 x more than target response. Enough for GB without losing much target response.

                        Comment


                        • #42
                          Originally posted by Aziz View Post
                          Btw,

                          the coil need not to be IB type. Mono coil as TX, a smaller GB coil placed in the center position (seperate RX) with a few turns should do the GB channel for GB info. The GB coil would detect during the high dI/dt the GB info. There wont be high voltage on the GB coil.
                          GB response is 1000 x more than target response. Enough for GB without losing much target response.
                          Huh ? You already have a coil .... "the monocoil" .... we dont need another one.
                          There is no point in me ( and the AI ) spending all this brain power and tokens to simplify the coil down to one simple coil and now you want to add another one ????

                          The GB is very simple ... you get it from the signal .. ( the rx signal is Self detection + Ground signal + Target ( sometimes).
                          We balance out Self Detection and Ground and you are left with Target.

                          Just to prove the point I tweaked the earlier code I use on the E class transmit ( using an IB coil in that case ) ... and now we get the discriminating monocoil TEM with GB

                          What we are looking at below is channel 1 and 2 ( virtual DSP channels in the software ) showing target responses and virtual channel 3 is the GB balance channel but also once balanced responds with Fe / Non Fe info. ( +ve for non iron ... -ve for iron ... no response for ground )
                          The response are not as strong ( its a cancellation channel after all ) in the GB channel but we only need the sign of the signal to disc the target.

                          There are 3 non iron targets then two iron targets.

                          This is not a simulation ... its straight off the test platform.



                          Click image for larger version

Name:	image.png
Views:	405
Size:	141.6 KB
ID:	448216​

                          Comment


                          • #43
                            Hi Paul,

                            anyway. You gonna do with mono coil. This is nice, if ground noise isn't an issue.

                            Aziz

                            Comment


                            • #44
                              Originally posted by moodz View Post
                              Before any "big reveal" ... you should all read up on this Minelab patent : https://ipsearch.ipaustralia.gov.au/patents/2023204592

                              Minelab are trying to lockup bipolar pulse induction till 2043.
                              What is interesting if you follow this document trail you will see at that site that Nokta has raised an objection to that patent since it would appear that Minelab is possibly trying to patent the idea of a controlled constant current bipolar transmit scheme.
                              Checkout the correspondence paper trail ( the lawyers are finding the gold here )
                              It is every companies right to protect their own IP however reading the patent seems to indicate they are patenting the method of producing bipolar constant current transmit scheme ( eg using feedback loops, current sensing, controlled high voltage and low voltage connections across a coil at various timings ...etc etc adapted and arranged. )
                              It is probable that the latest GPZ8000 model utilises this method. However it does require a separate recieve coil hence the horribly complex donut coil required on the GPZ8K.

                              Food for thought ...
                              What the examiner thinks of this ML patent.
                              This document is an examination report for a patent application related to a metal detection method involving specific current control techniques, highlighting issues with support, clarity, novelty, and inventive step.

                              Overview of the Examination
                              • The report assesses the patent application filed by Madderns Pty Ltd for Minelab Electronics Pty. Limited, focusing on a method for detecting targets using a metal detector that involves controlling transmitted electrical current through inductive windings.
                              • The examiner considers whether the application meets legal requirements for support, clarity, novelty, and inventive step, citing relevant prior art and literature.

                              Key Issues Identified
                              • Support and Disclosure: The claim defining the method lacks support because it omits essential features necessary for the invention’s operation, such as the measurement of the slope of the current during the first non-zero constant current period and the transitions between current states. The description indicates that these features are critical for achieving the claimed benefits, but they are not explicitly included in the claim, leading to inconsistency and insufficient support.
                              • Clarity: The claim’s preamble states it is a method to detect a target using a metal detector, but the rest of the application describes the device’s operation and design, causing ambiguity. Additionally, the term "transition time" is unclear, as it could refer to a single point or an interval, but the examiner interprets it as a point in time, which affects understanding.
                              • Novelty and Inventive Step: The application’s claims are not novel or inventive over prior art, particularly document D1, which discloses similar methods involving transmitting a sequence of magnetic fields with periods of rapid change and constant current, controlled by negative feedback loops. D1 explicitly describes measuring current slopes, controlling transition timing, and using feedback loops, all of which are features of the claimed invention. The use of multi-period sequences is considered obvious, based on common general knowledge, and thus does not contribute to inventive step.

                              Prior Art and Literature
                              • The examiner references multiple patents and non-patent literature, including prior patents from Minelab and other sources, which disclose similar methods involving negative feedback, current control, and specific signal sequences.
                              • The search strategy involved extensive keyword and classification searches across patent databases and online resources, confirming that the disclosed features are well-known in the field.

                              Legal and Procedural Notes
                              • The application does not meet the requirements under the Patents Act 1990, specifically lacking support, clarity, novelty, and inventive step.
                              • The applicant has 12 months from the report date to address objections; failure to do so will result in lapse of the application.
                              • The next continuation fee is due in July 2027.

                              In summary, the examiner concludes that the claimed invention is not sufficiently supported, lacks clarity in key terms, and is not novel or inventive over existing prior art, primarily D1, which discloses similar current control methods in metal detectors.
                              He concludes: that the application, as filed, does not meet the statutory requirements for patentability!









                              ​





                              ​

                              Comment


                              • #45
                                Lots of work on this lately ... The TX and the RX required some tricksy circuit wrangling but getting really good results now. I am using a 45 cm mono coil 300uh and 0.5 ohms basketweave. Full damping in 0.9 microseconds. Flattop pulses are better than 20 uA tilt factor at +/- 1.1 Amps and 750 volts flyback...the circuit is built with 1200 volt MOSFETs but I didn't go there yet ( tx supply is 3.3 volts .... 5 volts would be nearly 1200 volts peak ). The TX can be any frequency from 4 to 100 khz.

                                The front end amplifier uses a compound structure to achieve a gain of 1000x using 3X LME49990 .. ( opamp may be obsolete but hard to beat 0.9 nv noise ).

                                The current "tilt" at 32 Khz is approx 1 mv over 13 us seconds on the 300 uh coil ... this equates to 22 microamps so its pretty flat which is fairly good for a 1 amp pulse. ( dont get too excited though .. over one second the tilt is around 3 amps )
                                The tilt is important because if we have too much its going to limit the gain of the frontend before it overloads. You cant realistically achieve 0 tilt ... you just need to manage it so it does not overload the frontend amp.
                                In this case I have managed to put a 1000x amp on what is essentially a PI and in the past I have never put more than 30x - 100x on the preamp an sampling under 4 us on a largish coil.

                                Below is the rx signal before and after gain ( x1000 ) in relation to the gate drive.
                                The ADC wants something in the range of 3 volts PP so 1000 gain works here and we are sampling less than 4 us even though the coil is 45 cm diameter and relatively large.
                                ... and here is the important bit ... a mono coil that transmits and receives simultaneously is inherantly balanced ... think of two coils occupying exactly the same space but connected so they cancel except one of the "virtual" coils is used for transmit and the other for receive .. both at the same time. Because it is inherantly balanced ..... a ferrite near the coil causes an equal effect in both "virtual" coils and because it cant get "between" the coils ( they are the same coil ) it cant unbalance it. A large ferrite even if placed on the coil will distort the magnet field and change the inductance of the coil but it will produce NO target response because the coil balance is not disturbed. ( in this method ).

                                Another way to think about it is to consider a coil carrying a constant current if you place a ferrite there does the current change ?

                                Answer = no.

                                So all the ferrous "response" will be in the transitions .... and it there is an affect here and we can measure it ... thats another story.

                                Theres some maths there ...


                                ​ Click image for larger version  Name:	PXL_20261116_085305292[1].jpg Views:	0 Size:	610.9 KB ID:	454568
                                ​ Click image for larger version  Name:	PXL_20261113_063023611[1].jpg Views:	0 Size:	717.4 KB ID:	454571
                                Click image for larger version  Name:	PXL_20261116_085329107[1].jpg Views:	0 Size:	551.3 KB ID:	454569 ​
                                Last edited by moodz; 11-16-2026, 10:08 AM. Reason: typo

                                Comment

                                Working...
                                X