Announcement

Collapse
No announcement yet.

My Homemade VLF Detector

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

  • There is some merit to summing-in a fixed DC voltage to bring the standing signal closer to zero, for example during testing with a 'scope or DVM connected. But it's not strictly needed for real-world operation, so don't try too hard to make it zero, you will never achieve it, and you will see that the residual coil null drifts around, due to thermally-induced mechanical effects. Take the coil out in the direct sunshine, watch the null move significantly! Indoors, the coil will pick up larger items 'nearby' ie. within a metre or two. Moving the coil from 1 metre away from a metal shelf to 1.1 metre away will be detectable. When placed close to real ground, you will get a measureable signal, but it will still be much less than the residual coil null value.
    Also, the full-wave rectifier should have its' output buffered, so the subsequent stage doesn't 'ruin' its precision.

    Comment


    • Originally posted by ivconic View Post
      "...Disconnect coil and check again if there are still those "jumps".

      I disconnected the coil. No jumps. Extremely steady. Hooked up the coil again....the millivolt "jumps" return. Also, while experimenting, I noticed if I just barely place my finger on the cable for the coil, doesn't matter where, in the middle, on the ends, the voltage output will change as much as -40 mV. Even getting close, 1 - 2 cm away, I see the voltage begin to change.

      Does this mean this is begging to be shielded? or does it simply mean HANDS OFF!!

      Comment


      • Originally posted by Tepco View Post
        • This is absolute metal detector, no discrimination is possible this way, you can detect very small, microvolt changes in coil signal, but phase information is lost in rectification process.

        • No ability to provide ground compensation, so no matter how sensitive this circuit can be, will be more or less (probably completely) useless in any soil and achieve its full sensitivity only in air.
        I plan on adding more to this circuit for discrimination. Basically come off the Rx preamp into a high impedance buffer and do comparisons between the Rx and Tx phase changes. Currently, I'm just interested in understanding the detection process. This also helps me to understand op amps, noise, offset, etc. A good learning tool for me.

        For ground compensation, my theory is that the soil type will simply change the peak dc level after the full wave rectifier. So that should be fine since I'm not looking for a "certain" voltage level, just a fast change in voltage as the ground is swept side to side. I think that is the advantage I will have running this into a ADC and using the software to make the decision if metal has been detected.

        Comment


        • Originally posted by Skippy View Post
          There is some merit to summing-in a fixed DC voltage to bring the standing signal closer to zero, for example during testing with a 'scope or DVM connected. But it's not strictly needed for real-world operation, so don't try too hard to make it zero, you will never achieve it...
          yes, agreed. I do need to bring it down fairly close since the dc level is amplified. I can get the final signal to about 100 mV easily without too much trouble. That then gives lots of range in case I hit a treasure chest full of gold coins.

          Comment


          • Thanks Skippy for telling me I need to buffer so not to load down the rectifier stage. That seems to have completely cleared up the "jumping" of the signal on the output. It looks to be extremely solid now. Its late and I'll do more testing tomorrow. I basically changed over from a summing amp to a differential amp. I use 250K pot to zero out the dc.

            If this does the trick for me, then I can feed this output into a gain amp now then buffer for the ADC.

            Comment


            • Originally posted by JamesPicard View Post
              I disconnected the coil. No jumps. Extremely steady. Hooked up the coil again....the millivolt "jumps" return. Also, while experimenting, I noticed if I just barely place my finger on the cable for the coil, doesn't matter where, in the middle, on the ends, the voltage output will change as much as -40 mV. Even getting close, 1 - 2 cm away, I see the voltage begin to change.

              Does this mean this is begging to be shielded? or does it simply mean HANDS OFF!!
              Good! Now we know it is up to connection with coil.
              I think somehow an unwanted loopback formed there.
              You must be sure about coil shield and than connect it properly to rest of electronics.
              Otherwise it will not do.
              Cheers!

              Comment


              • You could make the offset adjustment circuit less sensitive by modifying the diff-amp. The 10K/10K divider on the non-inverting input could be changed to 56K/10K, for example. Or alternatively, put fixed value resistors in series with either end of the 'offset' pot to reduce its range.
                I think it would also be worth dropping down all the input amp resistor values by a factor of 10, and adding a low-pass filter cap across the other 220K resistor too (or 22K if you reduce them all)

                Comment


                • Originally posted by JamesPicard View Post
                  I disconnected the coil. No jumps. Extremely steady. Hooked up the coil again....the millivolt "jumps" return. Also, while experimenting, I noticed if I just barely place my finger on the cable for the coil, doesn't matter where, in the middle, on the ends, the voltage output will change as much as -40 mV. Even getting close, 1 - 2 cm away, I see the voltage begin to change.

                  Does this mean this is begging to be shielded? or does it simply mean HANDS OFF!!
                  There are 3 possibilities here:

                  1. The coil shell does have a Faraday shield, but you do not have it connected to 0V. Presumably one of the cable shields is connected to the Faraday shield in the coil shell, although there is no easy way to find out which. If Carl is correct, and the RX cable shield is the one, then you should make the connections in the order I stated earlier, and also connect both cable shields together.

                  2. The Faraday shield is good enough for the Garrett ACE, but is ineffective with your circuit due to the high DC gain. The cable itself may also become microphonic at high gains.

                  3. The coil shell does not have a Faraday shield. This is highly unlikely.

                  Comment


                  • Assuming it is #1, then I'll try using the Rx Shield of the cable. I'll make certain that the shield wire of both tx and rx go to ground.

                    However, since adding that buffer after the full wave rect. output, the circuit is much more stable in all respects. Even when I put my hand around the coil or close to the cable, it is barely noticeable on a scope....still can see it change, maybe 4-5 mV, but definitely not the 40+mV it was changing earlier.

                    Comment


                    • Well, I added a gain circuit to the output, but sadly the jumping voltage returns. Apparently the gain circuit is multiplying the extremely small changes on the input to the amp. Not sure what to do.

                      Comment


                      • Originally posted by JamesPicard View Post
                        Well, I added a gain circuit to the output, but sadly the jumping voltage returns. Apparently the gain circuit is multiplying the extremely small changes on the input to the amp. Not sure what to do.
                        Sounds like it's oscillating at a very low frequency.
                        Try adding some decoupling caps near the power pins of the ICs, and a 10uF cap from the center wiper of the 250k pot to 0V.
                        If that doesn't produce any improvement, then it might be external interference.

                        Comment


                        • There's a number of problems that could afflict your circuit. Your offset circuit uses the supply rails as it's reference - not good, you need a stable low-noise source, like a reference IC/diode. You still have all this joined up to bench PSU's and a sig-gen? Might be an idea to work on a clean battery power supply, eg. two PP3 batteries and linear regulators, and self-contained oscillator. You might then be able to drag it all outside away from metal and EMI and see how it does/doesn't behave.

                          Comment


                          • The idea of connecting the circuit to batteries, rather than the bench power supply (as mentioned by Skippy), should also be investigated. This can often be the cause of unwanted oscillation due to [unintentional] ground loops.

                            Hopefully you're learning a lot from this exercise.
                            Designing and building your own metal detector from scratch is not anywhere near as simple as some people would have you believe. That's why there are so many "workshop wonders" around, that completely fail when taken outside in the real world.

                            Comment

                            Working...
                            X