Announcement

Collapse
No announcement yet.

SPULE32 Coil modelling software by AZIZ

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

  • #16
    Hi all,

    after making 5 test case models, it showed, that I have to add the wire diameter w used in the bundle coil. With parameter a and w the compactness of the bundle coil is described.
    It is generally possible to find a good approximation formula for single loop turn coil emulation.

    But 5 test cases are not enough. I also have to make different coil shape models. Different sizes, different compactness, .. it goes into dozens of test cases...

    This is the approximation formula for 5 test cases I made. The AI can find the best approximation formula.

    Click image for larger version  Name:	Formula-1.png Views:	0 Size:	6.8 KB ID:	454242
    d, a, R, w: units mm

    Nice.
    Aziz

    Comment


    • #17
      And here is the big science challenge for you all.
      Find a better approximation formulas for air-core coils.

      The Key Individuals Behind the Approximation Formulas
      • Harold A. Wheeler (1899–1996): An American radio engineer who formulated the highly popular Wheeler's Formulas. Published in the 1920s and 1980s, his empirical approximations for single-layer, multi-layer, and flat pancake coils are universally used by engineers for quick, 1% accuracy calculations. [1, 2]
      • ​Frederick W. Grover (1876–1973): A physics professor and U.S. Bureau of Standards consultant who dedicated his life to calculating the inductance of complex geometric shapes. His seminal book, Inductance Calculations, provides the definitive collection of high-precision handbook tables and geometric approximation formulas​.
      • Ludvig Lorenz (1829–1891): A Danish physicist who developed the precise mathematical Lorenz Formula (1879) for single-layer cylindrical coils. His work represents the transition from basic theoretical approximations to highly rigorous geometric solutions utilizing complete elliptic integrals.​
      • Frederick E. Terman (1900–1982): Often called the "father of Silicon Valley," this Stanford professor compiled and simplified numerous practical radio frequency (RF) inductor approximation methods inside his famous Radio Engineers' Handbook.​
      • H. M. Greenhouse: Formulated the widely used Greenhouse Method (1974), which breaks down the planar spiral inductors on integrated circuits (ICs) and PCBs into individual straight segments, calculating their self and mutual inductances individually.​
      • <Put your name here >
      This is possible now.

      Comment


      • #18
        Interesting work, guys.

        I have my own coil modelling project 'on the back burner' , just no time available to spend on it. I was always intending to model each loop as a single turn - the business of modelling actual multi-turn bundles just seems to overly complicate things. [ I do have some mathematical additions to George's 'coil calculator' program, but that's another project to finalise ....]

        I'll have a think about your 'puzzles', I may have some pointers / ideas.

        Comment


        • #19
          Aziz .. modelling coils perfectly is old hat .. they had to do it for CERN and more recently for the various fusion projects that use very tricky magnetic containment.

          The actual method is based on Maxwells equations from first principles and the practical means is done by MOM method of moments.
          ( to avoid having to calculate for all of free space ).

          Skin effect, magnetic field, induction, resistance, capacitance can be derived from DC to daylight ( its not just an expression the math is actually valid from DC to optical frequencies ).

          What is amazing is it actually includes radiated energy (ie energy being emitted or radiated ( AKA antenna ).

          This applies to calculations not just for coils ... but any conductor / any shape. For instance calculating the eddy current patterns in a silver dollar.

          Click image for larger version

Name:	image.png
Views:	177
Size:	10.2 KB
ID:	454257

          But Claude did point out that if you are within 1% using approximation formulas .... its a waste of good math for metal detectors.​

          Comment


          • #20
            Aziz .. modelling coils perfectly is old hat .. they had to do it for CERN and more recently for the various fusion projects that use very tricky magnetic containment.

            The actual method is based on Maxwells equations from first principles and the practical means is done by MOM method of moments.
            ( to avoid having to calculate for all of free space ).

            Skin effect, magnetic field, induction, resistance, capacitance can be derived from DC to daylight ( its not just an expression the math is actually valid from DC to optical frequencies ).

            What is amazing is it actually includes radiated energy (ie energy being emitted or radiated ( AKA antenna ).

            This applies to calculations not just for coils ... but any conductor / any shape. For instance calculating the eddy current patterns in a silver dollar.

            Click image for larger version

Name:	image.png
Views:	179
Size:	10.2 KB
ID:	454259

            But Claude did point out that if you are within 1% using approximation formulas .... its a waste of good math for metal detectors.​

            Comment


            • #21
              ???? I posted once and it spammed the post. I told you stop using PHP or youll go blind.

              Comment


              • #22
                Hi Paul,

                yep. I agree with you. But good math wouldn't be wasted, if AC modelling (skin, proximity effect) would be taken into account.
                This would make things a lot not trivial (far beyond my scope).

                Anyway, single wire loop coil with turns factor set is quite inaccurate.That was the reason, why I have started modelling the coil bundle with each multiple loop turn in the later coil simulations.

                We should really work on yet missing real coil inventions before ML does.

                Cheers

                Comment


                • #23
                  Hi all,

                  I had a good idea during vacuuming at home to solve at least the skin effects in our coil software (AC behaviour).
                  It would require x times more processing overhead for x discrete wire diameter divisions of the round wire model. So a model which starts with thin wires up to the end wire diameter in x discrete steps (wire diameter will be modulated).

                  The each x calculated B-fields must be linear combined to meet the current density distribution model of the wire.
                  Since the Biot-Savart Law with Ampere's Law is accurate enough for round wires, it would perfectly work to make x discrete calculations with x discrete wires with different wire diameters.

                  4 discrete divisions would be enough I think. Or adjustable for more.

                  There we have it. The changes to the software would be relative small I think.

                  Well, I have to think of it further. It's a cheap idea only.
                  Aziz

                  Comment


                  • #24
                    Hi all,

                    the idea of post #23 won't work.

                    Aziz

                    Comment


                    • #25
                      ... hmm I think modelling the coil in SPULE ( it even models the sheilding now -- testing ) then going straight to 3D print in is the next BIG thing.

                      Throw away your aquadag ... . The sheild is printed in conductive polymer as part of the body :
                      It can be a solid / a layer down to .08mm on top of another non conductive part or a faraday optimised mesh.
                      Easy wind designs with modelling and printed sheilding means you can have that coil in a day or too.

                      workflow :

                      1. Select design or create your own in SPULE.
                      2. Model design and optimise.
                      3. Push the "COILCAD" button to create the 3D assembly
                      4. Print on 3D printer.
                      5. Wind the coils onto the "EASYWIND" former
                      6. Assemble cable / fine adjust as necessary.
                      7. Test


                      This is what can be done now.

                      Next project idea: The wire is layed out by the 3D printer too. So you just print the whole coil ... housing, sheilds and wiring in one job.

                      Click image for larger version  Name:	image.png Views:	0 Size:	1.20 MB ID:	454480​

                      Click image for larger version  Name:	image.png Views:	0 Size:	641.3 KB ID:	454479​
                      Last edited by moodz; 11-13-2026, 02:05 AM.

                      Comment


                      • #26
                        ​​​​​
                        Last edited by moodz; 11-13-2026, 02:03 AM.

                        Comment


                        • #27

                          Comment


                          • #28
                            ???

                            Comment


                            • #29
                              Hi all,

                              it is not too late to find some interesting and novel coil concepts and publish it here to avoid patenting by the market leaders.

                              Some brain storming ideas to look for:
                              - low ground noise generation
                              - low EMI noise induction
                              - deep & shallow target detection (combining large and small concentric coils)
                              - fast coils (low capacitance -> basket weave coils)
                              - anti-interference coils
                              - mini & portable ground loop concepts (extreme depth)
                              - two-box systems
                              - gradiometer coils
                              ...

                              Aziz

                              Comment


                              • #30
                                Hi all,

                                well, I have checked one interesting coil configuration for a PI coil on heavy ground mineralization. It ended about 10 % more detection depth on heavy ground mineralization with AI analysis.

                                The proof of concept must be checked with real coil simulations however. It is an optimization process of all possible parameters and there will be a sweet spot for real gain.

                                The coil concept makes the whole big picture difficult to see and no one looked at it thorougly.

                                It will take some time ..
                                Aziz

                                Comment

                                Working...
                                X