Announcement

Collapse
No announcement yet.

SPULE32 Coil modelling software by AZIZ

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

  • SPULE32 Coil modelling software by AZIZ

    Aziz and myself have collaborated to bring you SPULE32 which is Aziz's coil modelling software.
    He has kindly provided the original source code ( C, C++, ASM ) which originally was built for windows and I ( with AI help ) have ported to Javascript.
    So to be clear Aziz did all the work and I am just a messenger.

    SPULE32 is now a single HTML file that you can open in a browser ( Chrome, Edge, Firefox ) and you can run the coil sims.
    I have added soil models and GPU acceleration ( selectable in settings - note you have to enable in browser settings also in some cases)

    Included are files with example models ( minelab.zip ) and results ( soil.zip ) built directly from minelab coil patents ( GPZ8000 ).

    BIG THANKYOU TO AZIZ FOR HIS WORK and GENEROSITY



    Click image for larger version

Name:	image.png
Views:	465
Size:	445.8 KB
ID:	454079​

  • #2
    Good job! The project is interesting. For people like me, it would help to have a step-by-step walkthrough: step 1 -> 2 -> 3… through to the result.

    The primary TX field itself isn’t the interesting part. What matters is the secondary field: the one produced by eddy currents in a conductive target, or the imbalance caused by non-conductive magnetic material such as ferrite or black sand. The result I’d most like to see from the simulation is the RX pickup from that secondary field or imbalance, in nV or pV.
    ​

    Comment


    • #3
      Hi all,

      I am very impressed, excited and happy.
      The coil software is taken over by Paul now (full right ownership).

      Paul did huge amount of the part in a very very short period. AI is even capable to handle big old messy projects.

      Big thank you to Paul.


      I am looking forward to see coil development progress.

      Aziz

      Comment


      • #4
        This is so cool. The coil app runs in my old Firefox on Windows 7. It’s fun to work with it.​
        The AI added soil model, GPU support, etc..
        And even vector potential functions (I never implemented this but found it in the source code of the java script).

        We have to ask the AI for a good documentation and making the PDF manual.
        No hurry. We can add this and even further features later.
        It's the beginning of the professional metal detector coil modelling and analysis tool.

        Well done Paul.
        ​
        Aziz

        Comment


        • #5
          Here is the latest version : Now includes the coil wizard buttona and multi targets .... just press it and then select defaults ( next next etc ).
          Wait for calculations then Press REPORT button to convert the results to a nice readable report.
          Attached below. ( unzip and open file with Chrome or similiar)
          Attached Files

          Comment


          • #6
            Hi Paul,

            seems to work well. I did not checked thoroughly yet however.
            One have to load the wirepath file seperately. But it works.

            Click image for larger version

Name:	Coil-with-Wire-Path-1.png
Views:	331
Size:	438.5 KB
ID:	454155

            Aziz

            Comment


            • #7
              Wow! Nice! The Coil Wizard is damn good.

              I never will regret the decision I made.

              Aziz

              Comment


              • #8
                Next version 1.00 with Coil Designer and Layout wizard. Includes nailboard print to print out the pattern 1:1 so you just have to stick the printout(s) to a sheet of plywood stick nails/pins and wind your coils.

                You can null the coil as well in the designer.

                Click image for larger version  Name:	image.png Views:	0 Size:	422.1 KB ID:	454167​

                Comment


                • #9
                  Hi all,
                  I am looking into the js-code:

                  Claude Code has modified and added so much, that maybe < 5% code fragments are left there from the original source. And the rate is steadily decreasing.
                  ML engineers will have difficult times to reverse engineer it.

                  Have all possible coil inventions been discovered and patented yet?
                  Can we or Claude invent the ultimate coil?

                  Anyway, I am very very impressed about the AI possibilities and what Paul makes out of it.
                  This is sure from outer-space. Sci-fi hyper dev.

                  Aziz

                  Comment


                  • #10
                    Originally posted by moodz View Post
                    Next version 1.00 with Coil Designer and Layout wizard. Includes nailboard print to print out the pattern 1:1 so you just have to stick the printout(s) to a sheet of plywood stick nails/pins and wind your coils.

                    You can null the coil as well in the designer. ​​​

                    ​​​
                    Great work guys! The html format makes it very compatible with most systems. I hesitate to comment early on because the features I don't see may be hidden or planned.

                    At first glance the geometry for the DOD looks incorrect (see attached) ? The DOD coil is also called a "Super-D" and now Nokta uses "DRC" for Dual Receiver Coil. The geometry is a centered tx coil shaped as a rectangle with two out of phase D coils on each side. Electrically equivalent to a OOO coil. This coil became well known when BC showed a picture inside of the GPZ7000 coil. As always it was discussed on Geotech long before. Recently I made three different sized DOD coils 8", 10" and a 12x15". It will be interesting to compare the real with the calculated. I may be wrong about dod, but I am not familiar with layout wizards round tx with inner fig8 rx?

                    Is there a way to input size, inductance and resistance to give the turns and wire gauge? In my case, the above target inductance's were about 300uh 0R6 for tx and 400uh 6R for the series rx coils. The question for the calculator would be number turns and gauge for each size 8", 10" and 12x15". Currently, inputting the turns and wire gauge is of no benefit. We could just build the coil with turns, gauge and size.

                    Another suggestion for layout wizard. On OOO and DOD coils the rx coils could be treated as mirror images. Lets say the rx coil on the right side is the master. As you adjust the size and shape of the master the left side rx coil would track the changes equally. It should track the null point the same on left and right.

                    This is a wonderful tool..

                    Thanks!

                    DOD as I know it
                    ​
                    Unknown geometry

                    ​ ​​ ​
                    Last edited by Altra; 11-01-2026, 09:43 PM.

                    Comment


                    • #11
                      Originally posted by Altra View Post

                      At first glance the geometry for the DOD looks incorrect (see attached) ? The DOD coil is also called a "Super-D" and now Nokta uses "DRC" for Dual Receiver Coil. The geometry is a centered tx coil shaped as a rectangle with two out of phase D coils on each side. Electrically equivalent to a OOO coil. This coil became well known when BC showed a picture inside of the GPZ7000 coil. As always it was discussed on Geotech long before. Recently I made three different sized DOD coils 8", 10" and a 12x15". It will be interesting to compare the real with the calculated. I may be wrong about dod, but I am not familiar with layout wizards round tx with inner fig8 rx?

                      Is there a way to input size, inductance and resistance to give the turns and wire gauge? In my case, the above target inductance's were about 300uh 0R6 for tx and 400uh 6R for the series rx coils. The question for the calculator would be number turns and gauge for each size 8", 10" and 12x15". Currently, inputting the turns and wire gauge is of no benefit. We could just build the coil with turns, gauge and size.

                      Another suggestion for layout wizard. On OOO and DOD coils the rx coils could be treated as mirror images. Lets say the rx coil on the right side is the master. As you adjust the size and shape of the master the left side rx coil would track the changes equally. It should track the null point the same on left and right.
                      Done ... Super D topology added // MirrorMod button in top left corner of coil designer default is ON // Enter target resistance and inductance of Rx and Tx coils in RHS of screen // also added Z axis editing to ( mouse drag ) to set coil vertical separation.

                      Minimum bend radius is added also ( default is 10 mm )

                      Remember to press the NULL button to balance the coils ( as coils may move when this is done ).

                      Click image for larger version

Name:	image.png
Views:	243
Size:	495.9 KB
ID:	454200​​
                      Attached Files

                      Comment


                      • #12
                        Well done Moodz, bravo Aziz—truly amazing work.
                        I’ve deleted all the Excel spreadsheets from my comp.​

                        Comment


                        • #13
                          Originally posted by moodz View Post

                          Done ... Super D topology added // MirrorMod button in top left corner of coil designer default is ON // Enter target resistance and inductance of Rx and Tx coils in RHS of screen // also added Z axis editing to ( mouse drag ) to set coil vertical separation.

                          Minimum bend radius is added also ( default is 10 mm )

                          Remember to press the NULL button to balance the coils ( as coils may move when this is done ).
                          ​​
                          Amazing..... Thanks

                          Comment


                          • #14
                            Nice. The coil designer generates its own wire path data.

                            Spule32 v1.1.0 by AZIZ and some anonymous hacker​


                            I think, Paul deserves his name on it.

                            My contribution sunk to < 1% now.
                            I see my contribution as a creative spark.

                            BTW, Spule is the word for coil (german). When I began the project under Win 3.1, it was a 16-bit version. Later, under XP, it became a 32-bit version.​
                            I never compiled or ported it into 64-bit version or later Windows versions.

                            Two important conditions were necessary:
                            1. AI technology
                            2. The right person ( = Paul)

                            Aziz

                            Comment


                            • #15
                              Hi all,

                              I am trying to find a good approximation formula for the simplified coil modelling using single turn coils with turnsfactor set to N turns count and wire diameter d set to an equivalent rectangular cross-section coil bundle with N turns count and edge length of a.

                              a=Bundle width, R=mean coil radius, d=coil wire diameter (for single turn), N = turns count

                              The formula f to be found & checked:
                              d = f(R, a, N)

                              Click image for larger version  Name:	Rectangular-Bundle-vs-Single-Turn-with-Turnsfactor-and-Diameter.png Views:	0 Size:	151.2 KB ID:	454225
                              Above the rectangular coil bundle (red) is simulated with the above single turn coil (green, thick wire). Both get in this example same inductance.
                              If the formula is correct, we can speed up calculations using single thick wire model to simulate bundle coils.

                              I have to make two more different examples to check the formula.
                              Aziz

                              PS: AI can't calculate right... The formulas given from AI are wrong...

                              Comment

                              Working...
                              X