Announcement

Collapse
No announcement yet.

Coil Design for metal detector walk through gate

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

  • Coil Design for metal detector walk through gate

    I have a very serious problem regarding my walk through gate.
    I have opted it as final year project at university.
    I am using BFO technology for metal detection and making a walk through gate with 6 zones.
    For that purpose i made 6 detector circuits of BFO. The detector circuits are working quite fine independently and detect copper piece (PCB) of 4 square inches at about 10 inches but...the problem is that when i mount the coils on the gate (wooden frame) and make them work together they do not work. If at all i tune them with a lot care they do not cover more than 2 or 3 inches range and also gave small beep after some seconds. First i thought that the problem was with the power supply since it was common for the six zones. I changed the power supply and made six different power supplies for the six detectors. They still give no response.
    Now what i conclude is that when i mount the coils together on the gate their magnetic fields interact in a way to cause this problem.
    Can anyone please guide me how to mount the coils so that they may work together and not interfere. The width of my gate is 2.5 feet. Coil specs are diameter:9 inches, SWG: 21 , turns: 16. First i used dia: 9 inches, SWG 26, turns 35 but that was not letting me tune properly.
    Can any body tell me how this SWG , Diamwter and turns affect the range of detection.
    Do i need to change the coil design so that they may not interact or i need to shield my coils some way. Please please help some one.
    Following is how i have attached the coils. kindly some1 guide me if its the right way!
    Attached Files

  • #2
    Originally posted by Rizwan View Post
    I have a very serious problem regarding my walk through gate.
    I have opted it as final year project at university.
    I am using BFO technology for metal detection and making a walk through gate with 6 zones.
    For that purpose i made 6 detector circuits of BFO. The detector circuits are working quite fine independently and detect copper piece (PCB) of 4 square inches at about 10 inches but...the problem is that when i mount the coils on the gate (wooden frame) and make them work together they do not work. If at all i tune them with a lot care they do not cover more than 2 or 3 inches range and also gave small beep after some seconds. First i thought that the problem was with the power supply since it was common for the six zones. I changed the power supply and made six different power supplies for the six detectors. They still give no response.
    Now what i conclude is that when i mount the coils together on the gate their magnetic fields interact in a way to cause this problem.
    Can anyone please guide me how to mount the coils so that they may work together and not interfere. The width of my gate is 2.5 feet. Coil specs are diameter:9 inches, SWG: 21 , turns: 16. First i used dia: 9 inches, SWG 26, turns 35 but that was not letting me tune properly.
    Can any body tell me how this SWG , Diamwter and turns affect the range of detection.
    Do i need to change the coil design so that they may not interact or i need to shield my coils some way. Please please help some one.
    Following is how i have attached the coils. kindly some1 guide me if its the right way!
    The problem with your implementation (and particularly with BFOs) will be the oscillators locking together. If it's not too late, I would strongly suggest that you abandon the BFOs and use a PI circuit instead. This is the technology used by commercial walk-through detectors.

    Also, as a matter of etiquette, please only post your question in one forum at a time. I have removed the 6 other duplicate posts that you made.

    Comment


    • #3
      I agree, BFO is very prone to frequency locking. Therefore, it is probably the worst possible choice for a multi-zone walk-thru. Sorry!

      Most multi-zones use PI (though there are a couple of IB designs) whereby there is one large TX coil and multiple RX coils. The TX coil is placed on one panel, the RX coils on the opposite panel. Most designs double this, using a cross-fired method.

      If you want to attempt to salvage the BFO approach, you can try to skew each oscillator frequency enough to prevent locking. How much you need to skew depends on the design, and it may not succeed at all. Start with 2 channels and skew by 10%, see what happens. Avoid harmonics, i.e., f2 = 2*f1.

      Comment


      • #4
        pi would be the simplest option,the six zones could have staggered pulses and then the "off" zones could be simply blocked whilst the active one is on,would this not provide interferance free operation from itself ?

        Comment


        • #5
          Most PI W/T's have a single large TX coil in one panel and multiple RX coils in the opposite; this gives better sensitivity than doing TX/RX from the same side. The TX coil is pulsed, then the RX coils are listened to simultaneously through parallel receive channels.

          Comment


          • #6
            It would also be a good idea to keep the electronics for each circuit seperate, with good power-supply isolation. Locking is most probably magnetic in cause, but can be caused by other factors - capacitive coupling, power-supply rail variations, leakage between opamps in a multi-amp IC, even acoustic coupling is possible. Running them at a 15 - 25% frequency difference would be worth trying.

            Comment


            • #7
              Thank you all

              Thanks a lot for your valuable suggestion.
              The problem is that i have spent quite a lot money on this BFO design (bad luck).
              Now i m confused how much would the new design PI cost. Would that cost too much??
              And in case I try to improve my current design, Do I need to shift the frequencies of all the six detectors or only those facing each other?? Would placing the coils on each side not exactly in front of each other help??
              I have used 2n3819 in place of BF244. Is this replacement OK?? or i need to change??
              2n3819 has Forward and reverse junction capacitance different than BF244. for 2n3819 these are at 1kHz whereas for BF244 these are at 1 MHz.
              It would be better for me if I could survive with the current design. I am left with 1.5 month only. Can I implement the new design in this time??

              Comment


              • #8
                Originally posted by Rizwan View Post
                Thanks a lot for your valuable suggestion.
                The problem is that i have spent quite a lot money on this BFO design (bad luck).
                Now i m confused how much would the new design PI cost. Would that cost too much??
                And in case I try to improve my current design, Do I need to shift the frequencies of all the six detectors or only those facing each other?? Would placing the coils on each side not exactly in front of each other help??
                I have used 2n3819 in place of BF244. Is this replacement OK?? or i need to change??
                2n3819 has Forward and reverse junction capacitance different than BF244. for 2n3819 these are at 1kHz whereas for BF244 these are at 1 MHz.
                It would be better for me if I could survive with the current design. I am left with 1.5 month only. Can I implement the new design in this time??
                I guess the real question is "Are you flogging a dead horse?".

                Even if you switch to a PI design at this stage, the BFO project will not be wasted. Your final project report can include this valuable learning experience, enabling you to justify the switch to pulse induction.
                A quick fix would be to get a Surf PI kit from Silverdog ->
                http://www.silverdog.co.uk/shop/inde...&product_id=76
                As long as multiple zones are not a strict requirement for the project, you could put an elongated TX coil on one side and a similar RX coil on the other. Once you have that working (with hopefully some time left) it would be "relatively" simple to add a number of other receive circuits (one for each zone).
                As there's only 1.5 months remaining you'll either have to struggle on with a design, that most likely will give poor results at best, or bite the bullet and go for a PI.

                Whatever you decide, we're here to help.

                Comment


                • #9
                  OK, since you may not have the time to completely switch gears let's see if we can get those BFO's to work.

                  Originally posted by Rizwan View Post
                  And in case I try to improve my current design, Do I need to shift the frequencies of all the six detectors or only those facing each other?? Would placing the coils on each side not exactly in front of each other help??
                  I assume you have 3 coils on each panel, is that right? Looks like the coils are spaced maybe 5-6" apart, so compared to the 30" gate width I would expect the coils mounted together to be more prone to frequency locking.

                  Do this: turn off one panel. Adjust the 3 coils on the other panel to run at +10%, 0, and -10%. See if this gives decent sensitivity. If they still seem to lock increase the spread to 15% or 20%. Next, use only the middle coils on both panels, disable all the others. Run them at the same frequency, see if this gives decent sensitivity. If not, skew one of them until they work.

                  BTW, what frequency are you currently using?

                  I have used 2n3819 in place of BF244. Is this replacement OK?? or i need to change??
                  2n3819 has Forward and reverse junction capacitance different than BF244. for 2n3819 these are at 1kHz whereas for BF244 these are at 1 MHz.
                  Depends on what are you using it for. I would expect in most metal detector circuits it is plenty close enough.

                  - Carl

                  Comment


                  • #10
                    It might be worth re-shaping your coils to a square or even a mildly rectangular shape (wide and lower-height) to increase the seperation between them. This will only change (decrease) the inductance by 10% at most, hence your operating frequency will shift little. Just move them further apart in general. And route the cabling to each one seperately.

                    Comment


                    • #11
                      Thank you for the suggestions.
                      Here is my circuit diagram.
                      Attached Files

                      Comment


                      • #12
                        Originally posted by Carl-NC View Post

                        BTW, what frequency are you currently using?

                        - Carl
                        The frequency my colpitt's oscillator is somewhat around 100 kHz.

                        Comment


                        • #13
                          Help needed regarding PI

                          Hello all
                          I did shift the frequency of the oscillators 15-20 % and the zones are now working independently.
                          But the problem that still remains is that i am unable to get a range more than 8 inches.
                          For quite large objects it goes up to 12 inches or so.
                          Bringing metal quite close to the coil makes it go undetected.
                          I have 1 month almost and i am ready to do it pulse induction even with 1 or 2 zones only.
                          Please guide me regarding some good PI design that can be completed within a month.
                          I will really appreciate the guidance.

                          Comment


                          • #14
                            A quick fix would be to get a Surf PI kit from Silverdog ->
                            http://www.silverdog.co.uk/shop/inde...&product_id=76
                            [/QUOTE]

                            I am here in Pakistan. Do you have any idea how long it will take to reach me here???

                            Comment


                            • #15
                              Originally posted by Rizwan View Post
                              Thank you for the suggestions.
                              Here is my circuit diagram.
                              Do you have any local bypass caps on the supply input(s) to the opamp(s)?
                              You don't say what op amp you use. Is it a single supply type that can operate down to zero volts?
                              The way you have the two op amp circuits I guess could be made to work but the trim adjustments affect gain as well as signal level - plus you have large caps in this circuit that could affect the rate of threshold response. Same for the opto isolator - they probably don't all act identically, you may need to tweak threshold for each and their off decays may differ as the optos aren't matched.
                              Was this a proven circuit or how did it come about?
                              Does your system work with just one TX-RX?
                              What if you make up some kind of fet switch sequencer that cycles between the 3 TX/RX pairs and use one receiver. Perhaps add a signal blanker during the switchover period...?

                              Comment

                              Working...
                              X