Announcement

Collapse
No announcement yet.

Coil improvements

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

  • #31
    There is other information about coils that I am still looking for:
    Ergonomics – this involves size, shape and weight of the coils.
    ·How can we optimize the weight of the coils?
    ·What shape and size of coil gives the most depth?
    ·What size and shape of coil gives the most efficient ground coverage?
    What is the best “all round” coil, the coil that gives the best compromise between all of the above?
    Tinkerer



    Tinkerer,

    ·How can we optimize the weight of the coils?
    Use the thinnest and lightest wire at a modelate to high TX PPS frequency.
    Fill coil housing with foam to keep the coil rigid but light, with no movement inside the housing that might cause false signals.

    ·What shape and size of coil gives the most depth?
    This depends on what you are looking for. Eric Foster made a chart that shows the relationship of target size to coil radius and suggests ways to optimize a coil size to detect a particular size target. Coins are pretty common targets so the rule of thumb is to use a coil that is in the 10" to 12" range for these common targets. PI machines are pretty sensitive to coil size when trying to operate a large coil at low delays. Eric Foster mentioned many times the difficulty in getting a coil larger than about 11" to operate at 10us. The capacitance of the coil shield seems to be the controlling factor. If I were to attempt making a coil in the 14" range operate at low delays, I would use it with a shaft mounted control box to minimize the coax capacitance. I would use all the techniques that I have shared to make a fast coil like using AWG 30 Teflon wire, Teflon spiral wrap or polyethylene spiral wrap, a low capacitance shield like Scotch 24, and a short coax 30" to 34" long.
    ·What size and shape of coil gives the most efficient ground coverage?
    I do mostly beach hunting so this tip is related to my beach experience. Those who hunt for gold nuggets may want to share their experience.
    To cove a lot of ground try a coil that is 18" long by about 6" wide. Hays Electronics make one that is about this size. The rule of thumb seems to be: make the coil width about 1/3 the length. When you wind a coil that is a circle and compress it into an oblong shape, the inductance is reduced between about 5% and 10% . One of my 10.5" coil is 270uH in a round circle. When I squeeze it into a 13" by 5" oval it is 250uH and when I squeeze a little more to be 4" wide it is 240uH. This means that if you plan on making a oval coil, make it about 10% higher inductance so it will be your desired inductance in it's final oblong shape.
    What is the best “all round” coil, the coil that gives the best compromise between all of the above?
    Eric Foster mentiond that an 11" coil is the best compromise for general purpose use.

    Here is a coil making idea that I have kicking around in my head. Maybe some forum members can improve upon it. This should make a rigid, waterproof and relatively light coil for beach hunting.

    Bend a .5" ID PVC plastic pipe into a 12" ID circle with a heat gun and by filling the pipe with sand, capping the ends to keep the pipe from kinking.

    Insert into the bent .5" ID pipe a flexible pipe 3/8" OD X .25" ID with Scotch 24 around this like a long sock. Solder a ground lead to the Scotch 24.
    Insert and glue a T connection. Use a guitar string as a long needle and feed about 20 turns of Teflon insulated wire through the inner shielded .25" ID pipe, one turn at a time at the T connection opening. This will probably take two people to keep the wire from kinking or getting tangled. When finished, spray expandable foam into the inner .25" pipe to secure the wires.

    I hope this stimulates some tinkering.

    bbsailor

    Comment


    • #32
      Originally posted by hobbes_lives View Post
      That's the part I'm curious about. I'm curious of how much eddy current exists in that direction, given that it is 90 degrees to the coil's lines of flux. There is likely some, due to not being at exactly 90 degrees, but there shouldn't be much. Perhaps an experiment is in order...



      I'm quite familiar with a PI's response to open vs closed loops. It's the same reason why a closed hoop earring can be detected quite deep, whereas an open hoop can only be detected as very shallow depths. It's all about closing the loop for eddy currents to flow, thus lengthening the decay time of the received signal.

      Now, I propose a second part to the experiment. Turn the same closed loop of wire 90 degrees to the face of the coil and see what happens then. What's the PI's response in this case? (I know you already know the answer.)
      Kyle,

      I see by one of your previous posts that you are using Scotch 24 on a mono coil. What is the minimum delay that you are able to sample? If you seek small gold, then obtaining delays below 10us will improve the response to these tiny gold targets. My fastest PI works at 7.5us and some shield material that is not detected at 10us has a response at 7.5us. Some of my coils are being used by people who are in gold territory operating in the 5us to 6us range. I suspect that at these very low delays, the coils will be very sensitive to small gold. These very low delay coils are also very sensitive to themselves. That is why stranded wire or Litz wire is recommended at these very low delays. Also, shield material that is good at 10us might be not so good at lower delays.

      I researched and tried a lot of shield material and came upon Scotch 24 as a good shield at low delays and shared that discovery with members on this forum. I am constantly refining my coil making techniques and understanding of what makes a good coil.

      I suspect that if you used Scotch 24 on a mono coil and it were to touch around the cross section diameter of the wire bundle, that it would not create any problems at about 10 to 12us. However, my suggestion to leave a gap is to accomodate those who are trying to squeeze the fastest speed out of a coil below 10us. Granted, the shield runs 90 degrees to the coil and should have minimal interaction. But, being so close to the coil and having any amount of eddy currents generated on the shield or within the coil wire could be sensed by high gain circuits at very low delays.

      The coil is only one component of the potential speed of a PI. Once the coil has the least amount of capacitance and minimal potential eddy current design using all the techniques discussed here, your attention needs to shift to the TX and RX circuits. The first candidate is looking for a low COSS (output capacitance) MOSFET which contributes to the value of the damping resistor (Rd). Less capacitance means a higher damping resistor value and potentially faster coil discharge TC.

      The coil discharge TC is governed by Rd which is in parallel with the input resistor (Rin) while the discharge voltage is above .7V on the first amplifier stage on a mono coil. You want the coil discharge TC to be 5 times faster than the TC of your smallest desired target. The next large improvement comes from separating the TX from the RX circuit in the form of a DD coil or induction balanced (IB) concentric coil. An improvement in coil speed can be in the order of about 2us using a DD or IB concentric coil.

      All of the techniques discused here I have tried down to 7.5us. Those who want to venture below this need to try techniques that might not be obvious at higher delays. If you live in gold territory and seek small gold, you need to try every technique to make a faster response starting with the coil and then moving into the circuit design.

      bbsailor

      Comment


      • #33
        Does anybody tried the corrugated board? Cut in thin stripes, this can be wound to the coil bunch as a spacer. As the (dry) paper has a dielectric constant of 3.5 and there is a lot of air, so this value can be decreased at least to the half or more.
        Teflon has 2.1.
        See dielectric constants:
        http://en.wikipedia.org/wiki/Dielectric_constant

        Aziz
        Attached Files

        Comment


        • #34
          Originally posted by bbsailor View Post
          I see by one of your previous posts that you are using Scotch 24 on a mono coil.
          Yes, that's correct.

          Originally posted by bbsailor View Post
          What is the minimum delay that you are able to sample?
          I don't know yet. I have built the coil, but have only just received the parts that I ordered to build my PI. Now I have to find time between work, family and detecting to start experimenting with my design.

          Originally posted by bbsailor View Post
          I suspect that if you used Scotch 24 on a mono coil and it were to touch around the cross section diameter of the wire bundle, that it would not create any problems at about 10 to 12us. However, my suggestion to leave a gap is to accomodate those who are trying to squeeze the fastest speed out of a coil below 10us. Granted, the shield runs 90 degrees to the coil and should have minimal interaction. But, being so close to the coil and having any amount of eddy currents generated on the shield or within the coil wire could be sensed by high gain circuits at very low delays.
          Once I have completed my PI, I may wind another similar coil, this time leaving a gap around the inside circumference of the shielding to compare the results. I agree that there are still likely some eddy currents that may be eliminated by doing this, because there may be eddy current that run slightly off of the 90 degree mark, yet still circle the circumference of the wire bundle. (Does my wording make sense?)

          Originally posted by bbsailor View Post
          The coil is only one component of the potential speed of a PI. Once the coil has the least amount of capacitance and minimal potential eddy current design using all the techniques discussed here, your attention needs to shift to the TX and RX circuits. The first candidate is looking for a low COSS (output capacitance) MOSFET which contributes to the value of the damping resistor (Rd). Less capacitance means a higher damping resistor value and potentially faster coil discharge TC.
          I have order a few different MOSFETs to test out. Two are commonly used in PIs, the IRF740 and IRF840, but I'm leaning towards an IRF510 right now. The Vdss is lower, 100V vs 400V or 500V, and the Id is lower, 5.6A vs 10A or 8A, but the Coss is 65pf vs 195pf or 190pf. The current rating is fine, I'm just wondering how the 100V avalanche voltage will affect the minimum sample delay or overall sensitivity of the unit. I think the Coss is great, though.

          Also, I'm using an active-off scenario for driving the MOSFET, which should help reduce the sample delay.

          Originally posted by bbsailor View Post
          The coil discharge TC is governed by Rd which is in parallel with the input resistor (Rin) while the discharge voltage is above .7V on the first amplifier stage on a mono coil. You want the coil discharge TC to be 5 times faster than the TC of your smallest desired target.
          I'm currently going with a two-stage receiver, using an NE5534 and a 1K input resistor on the front end, which is pretty typical. The value of the damping resistor will be determined once the coil driver and receiver circuitry is functioning.

          Originally posted by bbsailor View Post
          The next large improvement comes from separating the TX from the RX circuit in the form of a DD coil or induction balanced (IB) concentric coil. An improvement in coil speed can be in the order of about 2us using a DD or IB concentric coil.
          That's on my radar too, once the mono configuration is optimized to my liking.

          Originally posted by bbsailor View Post
          All of the techniques discused here I have tried down to 7.5us. Those who want to venture below this need to try techniques that might not be obvious at higher delays. If you live in gold territory and seek small gold, you need to try every technique to make a faster response starting with the coil and then moving into the circuit design.

          I'm a shallow water surf hunter and beach comber myself. I'll instead be dealing with a different issue of trying to keep the unit quite as waves pass above the coil in knee to waist deep water, while still maintaining a low sample rate. That ought to be fun to figure out how to do.

          Comment


          • #35
            Kyle,

            You said: "I have order a few different MOSFETs to test out. Two are commonly used in PIs, the IRF740 and IRF840, but I'm leaning towards an IRF510 right now. The Vdss is lower, 100V vs 400V or 500V, and the Id is lower, 5.6A vs 10A or 8A, but the Coss is 65pf vs 195pf or 190pf. The current rating is fine, I'm just wondering how the 100V avalanche voltage will affect the minimum sample delay or overall sensitivity of the unit. I think the Coss is great, though."

            Tabular data for COSS is usually listed at 25 V. When looking at the actual COSS, look at the graph data for the voltage you plan to use on your PI. For example, the IRF510 is 80pf at 25V and about 125pf at about 12V. There are better MOSFETS with lower COSS in the 30 to 50pf range. Search this forum for posts by Reg who mentioned some low COSS Fairchild MOSFET part numbers.

            If you use this MOSFET with a 100V rating you will need a series resistor with about 4 X the resistance of the coil to keep the voltage from exceeding the MOSFET clamping voltage of 100V.

            When the MOSFET clamps, it will extend the delay while the flyback voltage is above 100V. Adding a series resistor of about 20 ohms (2 Watts) will keep the MOSFET voltage below 100V and will make the total 300uH coil turn on TC of about 300/25 or 12uS. So, a 60us pulse width will cause the current to reach 5 time constants and reach a constant current level. Read about current rise time constants on the web. Look on the Fairchild web site for some N channel MOSFETS with low capacitance and a voltage in the 300 to 500V range.

            bbsailor

            Comment


            • #36
              Spacer

              Originally posted by Aziz View Post
              Does anybody tried the corrugated board? Cut in thin stripes, this can be wound to the coil bunch as a spacer. As the (dry) paper has a dielectric constant of 3.5 and there is a lot of air, so this value can be decreased at least to the half or more.
              Teflon has 2.1.
              See dielectric constants:
              http://en.wikipedia.org/wiki/Dielectric_constant

              Aziz
              I have tried cardboard and it works OK when it is very dry.However, cardboard being essentially hygroscopic, it absorbs humidity from the air and then changes its characteristics.
              I found the material of picture below, that is made of Polypropylene basically a spacer that is mostly air. It is fairly rigid but easy to cut. I apply a coating of 30 microns of graphite composition with an embedded drain wire. Got good results.

              Are you still doing coil analysis? I want to build a new IB-PI coil, could you test my planned design?
              Tinkerer
              Attached Files

              Comment


              • #37
                spacer

                Tinkerer is that the sign material that all the politico signs are using on the side of the roads? Hard to tell if it is that or a larger dia. material Do you use a heat gun to make it more plyable? I was thinking if that is the political sign stuff, there's going to be a lot available soon
                Wyndham

                Comment


                • #38
                  PP spacer

                  Originally posted by wyndham View Post
                  Tinkerer is that the sign material that all the politico signs are using on the side of the roads? Hard to tell if it is that or a larger dia. material Do you use a heat gun to make it more plyable? I was thinking if that is the political sign stuff, there's going to be a lot available soon
                  Wyndham
                  The stuff I get is from old packing boxes. It has the same purpose as cardboard, but it is waterproof. It comes in different thickness and colors.
                  I cant remember its name, but somewhere within my 10 GB of metaldetector info on the computer I have a manufacturer of it, just need to find it, if only I could remember what it is being called.

                  I know it works just fine as a spacer and base for the shielding. Proof is, I just run a PI to sample at 3.5us delay, very sensitive. (low power)
                  Tinkerer

                  Comment


                  • #39
                    Hi Tinkerer,

                    your spacer material is quite good and I haven't found equal material here yet. I also want to build a new concentric IB coil and have also planned some coil analysis for the coming weeks. The standard concentric IB coil with the dominating TX-coil being outside and the others inside seams to be good one.
                    I will publish next week some coils with different number of windings and diameter.
                    Regards,
                    Aziz

                    Comment


                    • #40
                      IB coil and spacer

                      Originally posted by Aziz View Post
                      Hi Tinkerer,

                      your spacer material is quite good and I haven't found equal material here yet. I also want to build a new concentric IB coil and have also planned some coil analysis for the coming weeks. The standard concentric IB coil with the dominating TX-coil being outside and the others inside seams to be good one.
                      I will publish next week some coils with different number of windings and diameter.
                      Regards,
                      Aziz
                      I will make an effort and spend the time needed to find the manufacturer of the PP "cardboard" so that we get the name of it the rest is Google.

                      About the IB coil.
                      I have some numbers that I dont understand.
                      When I measure the inductance of the TX coil alone, I get about 350uH
                      The Bucking coil has about 130uH
                      The RX coil has about 450uH

                      When I connect the TX and the Bucking coil in series aiding, I get about 650uH why?
                      Then I connect TX and Bu in series opposing and I get 350uH???
                      I wonder what is wrong?
                      The problem is that this coil assembly works marvelously so whatever seems to be wrong is also very right.
                      "The Enigma Coil"

                      Tinkerer

                      Comment


                      • #41
                        Originally posted by Tinkerer View Post
                        I will make an effort and spend the time needed to find the manufacturer of the PP "cardboard" so that we get the name of it the rest is Google.

                        About the IB coil.
                        I have some numbers that I dont understand.
                        When I measure the inductance of the TX coil alone, I get about 350uH
                        The Bucking coil has about 130uH
                        The RX coil has about 450uH

                        When I connect the TX and the Bucking coil in series aiding, I get about 650uH why?
                        Then I connect TX and Bu in series opposing and I get 350uH???
                        I wonder what is wrong?
                        The problem is that this coil assembly works marvelously so whatever seems to be wrong is also very right.
                        "The Enigma Coil"

                        Tinkerer
                        All coil inductance calculations seen on the literature when switching different coils in series or parallel are to be assumed not inductively coupled.
                        When coils sharing the same or part of the flux area, they are inductively coupled. When the magnetic field is increased (bucking coil in series), then the inductivity will be grown. When the magnetic field is decreased (bucking coil in opposite series), the inductivity will be decreased.

                        For inductively coupled coils, it is not trivial to get the exact total coil inductivity. There are some very expensive software on the market for this. Fortunatelly, my software is capable to calculate any arbitrary coupled coils.
                        Aziz

                        Comment


                        • #42
                          IB coil specs

                          TX coil - outer dia. 275mm-inner dia. 220mm-27 Turns
                          RX coil -outer dia. 200mm-inner dia. 50mm-62 Turns
                          Buck coil-outer dia. 170mm-inner dia.110mm-27 Turns

                          The coils are flat wound.

                          TX and RX are on the same plane.
                          A polypropylene spacer of 8mm is placed on top of the TX and RX coils.
                          The PP spacer is coated with 50 microns of Graphite. A drain wire conducts the noise and eddy currents from the Graphite to Analog ground.
                          The Bucking coil is placed on top of the Graphite. A precise NULL is found by moving the Bucking coil slightly off center. The moving of the Bucking coil is extremely sensitive, 1mm in any direction has a big effect.
                          It is also possible to lift one side of the Bu coil for nulling. Again it is a question of mm.
                          Tinkerer
                          Last edited by Tinkerer; 05-10-2008, 02:16 PM. Reason: correction

                          Comment


                          • #43
                            Giving this lost thread a bump as it contains some interesting topics.

                            Comment

                            Working...
                            X