Announcement

Collapse
No announcement yet.

ML EQUINOX coil discussion thread

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

  • ML EQUINOX coil discussion thread

    There's been some previous chat about the workings of the Equinox, including the coil, in this thread:
    https://www.geotech1.com/forums/show...ineLab-Machine

    Of particular interest are the X-ray images of the coil, that have been posted on the MD-Hunter and MD-Arena blogs:
    https://md-hunter.com/minelab-equino...lf-of-machine/
    and
    https://translate.google.com/transla...o-polovina-md/

    Points worth noting from these images include: The TX winding is multi-stranded, probably Litz wire. The RX appears conventional, a single strand of wire.
    The electronics is all contained inside the coil, there are no extra parts, like 'security/anti-piracy' devices in the coil free-socket ('plug'). It would be sensible to assume the circuit board has anti-copying/identification device(s) , and a signal pre-amplifier.

    Thanks to the great work of Foma, it seems the circuit is just a pre-amplifier, and a security device. They are distinct functional blocks, so could be seperated. The security parts could be fitted in the coil socket, or in a short extension cable. The pre-amp alone fitted in the coil. Multiple coils could then be fitted onto one 'security dongle'.
    ------------
    The connector and cable:

    These are a common connector, usually known as the " M12 " series, the Equinox using the 8-pin variant. The control box has a chassis-mounted male plug, the coil has a 'free socket' with female contacts. Pins are numbered 1 to 8, but don't seem to be commonly marked on the actual connector. The pinout diagrams are attached below.

    --
    The cable ML use has a single-core screened cable, with a grey outer sheath, and 5 single insulated wires: brown, red, blue, yellow, green , for a total of 7 connections. There are no additional drain wires, foil shield, or braided shield. Photo below of mouse-chewed cable:
    ------

    Pin 1 : TX hot (Shielded wire inner conductor)
    Pin 2 : TX cold (Shielded wire outer)
    Pin 3 : Brown
    Pin 4 : Red
    Pin 5 : Blue
    Pin 6 : NC ?
    Pin 7 : Green
    Pin 8 : Yellow
    Only the shielded cable has thick wires, the other 5 are low-power, and small. Inside the control-box, Pins 1 and 2 have much thicker tracks on the flexi-rigid cable that solders onto the coil plug. Resistance measured between Pin1 and Pin2 is approximately 0.5 Ohms ( TX coil + cable ). Inside the coil, the shielded cable joins directly to the Litz wire of the TX coil. Unlike some earlier ML machines, power is not 'robbed' from the TX drive to operate the pre-amp etc. There is a seperate DC power wire, of 3.3 Volts.

    Foma has given us a photo of the PCB. By comparing it with the X-ray images, it can be seen that it is fitted 'upside down' in the coil housing. The component side of the board is next to the underside of the coil. The wire connection pads are on the component side, too.

    Foma has provided a partial reverse-engineered schematic of the PCB. It is a low-noise design, with a discrete transistor differential pair at the front end, followed by a single op-amp. The opamp output feeds back into the front end, to control the gain, at what looks like about G = 16. This opamp output then goes to a unity-gain inverter opamp circuit. This makes the circuit have a differential output, feeding back to the control box.

    The security device is in a single 3-pin package, with just 0V, +3.3 V power, and a data signal. This device could easily be relocated to the coil socket, perhaps with a PSU decoupling capacitor for company. The PCB can be hand cut, 4mm square is enough space. If this is done, only 6 wires need to go down to the coil: TX +/- ; Ground; +3.3V supply; RX +/- signals.
    Attached Files
    Last edited by Qiaozhi; 03-05-2021, 09:50 PM. Reason: Corrected coil pinout list to match post #10.

  • #2
    There are some obvious problems with the schematic.

    Comment


    • #3
      Yes, Carl, plenty of errors, we can work on them. The two diodes on the front-end are probably between the bases of the npn transistors.
      I think there's a virtual ground in there, probably what's labelled on the PCB as 'REF', though not shown on the schematic.
      Last edited by Qiaozhi; 01-03-2020, 10:45 AM.

      Comment


      • #4
        Something like this might be the security device? Data sheet states that package markings are intentionally non-part related.

        http://ww1.microchip.com/downloads/e...-40002025A.pdf

        Comment


        • #5
          I wondered if it may be this part:
          https://eu.mouser.com/Search/Downloa...7A-1220769.pdf
          But there's a fair number of these devices, from several manufacturers, and it's not important, really, it's the analog stuff we need to reverse-engineer properly.

          I'm wondering about the SOT23-6 device attached to the differential output. It could be a protection device, to limit spikes on a long cable. Something with diode clamps, transzorbs/zeners that limit voltage pulses to the 0V and 3.3V supply rails of the device. I've noticed them quite a lot in commercial equipment.
          These type of things:
          https://www.littelfuse.com/products/...rotection.aspx

          Comment


          • #6
            Hi Skippy, did that broken blue coil wire get repaired by you, did it come to you broken ?

            Comment


            • #7
              Originally posted by Carl-NC View Post
              There are some obvious problems with the schematic.
              A cursory look at the schem points out some obvious curiousities, for example, why is the positive rail connected to the non inverting input.

              Comment


              • #8
                Upon further examination of the schem, it appears as though the rx emf is intended to swing between?????

                Comment


                • #9
                  Hey guys,

                  Upon looking at the schematic and the x-ray photos, aren't amplifier outputs connected to the Tx side of the coil? It says Rx? Judging from the x-ray photos, the thick shielded wire is connected directly to the Rx coil and is fed right into ADC though low pass filter in the control unit? Looks like the amplification stage for the Tx is moved into the coil itself and is fed with low level signal, but it is also unshielded. There is probably no need to shield low level Tx signal because it get smoothed out by the coil and the ground...

                  I don't think any other coil will work, even if one defeats the coil ID. It is an active coil.

                  It is probably too much to ask. Is there a good picture of the coil board and/or which contacts correspond to which pins in the connector?

                  Tremendous work trying to reverse engineer it.

                  Comment


                  • #10
                    Pinout update:

                    Pin 1 : TX hot (Shielded wire inner conductor)
                    Pin 2 : TX cold (Shielded wire outer)
                    Pin 3 : 3.3 V DC power (Brown)
                    Pin 4 : RX + or - (Red)
                    Pin 5 : RX - or + (Blue)
                    Pin 6 : NC ?
                    Pin 7 : 0 V ground (Green)
                    Pin 8 : Data (Yellow)

                    The 3V3 supply is split, probably with just a resistive divider, to give a virtual ground of about 1.6 Volts.
                    The RX outputs are biased at this 1.6 V level, the AC signal added to this. So one increases in voltage while the other decreases, etc.

                    Notice the revised TX hot/cold wiring, I got them the wrong way round in my original post.

                    I made a little breakout cable for experiments. Pic attached.
                    Attached Files

                    Comment


                    • #11
                      Originally posted by Skippy View Post
                      Pinout update:

                      Pin 1 : TX hot (Shielded wire inner conductor)
                      Pin 2 : TX cold (Shielded wire outer)
                      Pin 3 : 3.3 V DC power (Brown)
                      Pin 4 : RX + or - (Red)
                      Pin 5 : RX - or + (Blue)
                      Pin 6 : NC ?
                      Pin 7 : 0 V ground (Green)
                      Pin 8 : Data (Yellow)

                      The 3V3 supply is split, probably with just a resistive divider, to give a virtual ground of about 1.6 Volts.
                      The RX outputs are biased at this 1.6 V level, the AC signal added to this. So one increases in voltage while the other decreases, etc.

                      Notice the revised TX hot/cold wiring, I got them the wrong way round in my original post.

                      I made a little breakout cable for experiments. Pic attached.


                      neat

                      Comment


                      • #12
                        Legend!

                        Comment


                        • #13
                          Some analysis of the TX coil, based on measurement, not by dismantling the winding:

                          Measured L = 119 microHenry ( 0.119 mH )
                          Measured R across free socket pins = 448 mOhm
                          Resistance of screened cable outer = 50 mOhm
                          Estimated resistance of screened cable inner = 50 mOhm
                          So ... resistance of coil winding = 350 mOhm approximately.

                          For the stock 11" coil, the TX winding measures up with a circumference of 640mm, equivalent to a circular coil of R = 102mm.

                          A bit of work with the coil calculator suggests the number of turns is N = 15T , total wire length = 9.6 m.
                          To get the correct resistance, a wire equivalent to 0.8mm diameter is needed. R = 36 mOhms per metre.
                          Common Litz wire is based on 0.04mm, 0.05mm or 0.071mm diameter ECW.
                          0.8mm diam solid is equivalent to (400 x 0.04mm ), or ( 256 x 0.05mm )

                          See here for some possible choices, such as ( 243 x 0.05mm ):
                          https://www.wires.co.uk/cgi-bin/sh00...00710090_2d100

                          For a theoretical 17" round coil, no. of turns = 11T , needing about 11 metres of 0.85mm diameter equiv. wire.
                          And a 5" round would need 25 T , length 7.5 metres, of 0.67 mm diameter equiv.

                          So that gives a ballpark figure of a typical TX winding - 10 metres of 0.8mm equiv.

                          Comment


                          • #14
                            In the schematic posted, if the 2835(2) is supposed to be the split rail supply op-amp, then it looks strange to me. Shouldn't both 1.8K's be connected to pin 7 to form the voltage divider? I think the schematic should look more like this.
                            Attached Files

                            Comment


                            • #15
                              I just hooked up the coil and control box to the oscilloscope and can see it registers items and displays TID when RX amplitude changes within 1-2mV at sensitivity 20. Can barely see the change as RX idles at 0.3-0.5V Pk-Pk and is biased.

                              Nothing too new really, just showing. I was interested if ML have exhausted room for sensitivity in analog domain.

                              Square is TX signal, triangle is RX.

                              Click image for larger version

Name:	IMG_2853.JPG
Views:	1
Size:	152.4 KB
ID:	356205
                              Click image for larger version

Name:	IMG_2854.JPG
Views:	1
Size:	98.2 KB
ID:	356206

                              Comment

                              Working...
                              X