Announcement

Collapse
No announcement yet.

Tarsacci MDT8000 : your thoughts, analysis

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

  • Tarsacci MDT8000 : your thoughts, analysis

    Those who use Tom Dankowski's forum will have seen this thread about a new 'VLF' machine, the Tarsacci MDT-8000:
    http://www.dankowskidetectors.com/di...,152652,page=1

    It's designed by Demeter Gargov, a former Fisher engineer, the CoinStrike being one of his projects.
    It appears to be fundamentally a selectable single-frequency VLF machine (6.4, 9,12,18kHz), with a regular-looking DD search-coil. It has the advantage of giving greater depth than a standard single-freq VLF, it seems. This makes me conclude that it has some method of reducing the effect of the ground signal.
    It's not inconceivable that it uses multi-frequency like the Fisher CZ-series, so when 9kHz is selected, it's actually (9 and 27 kHz) etc.
    But reading the 'operating instructions' suggests that it is actually something different - the audio has a different sound, there's a special ground-balancing procedure. This makes me think it may be using additional sensors to determine ground signal, probably Hall sensors, though other possibilities exist.

    The appearance of the coil shows that if there's anything extra inside the case, it's quite small. There's probably not any secondary transmit or receive coils.
    It's also stated that the machine is 'matched' to the coil, it's not possible the swap coils. If it had a secondary sensor, this might imply the main control box is used to calibrate the way 'VLF coil' and 'secondary sensor' are processed.

    It's also stated that it is more sweep-speed sensitive than regular machines.
    It could be because the two sensors don't match up perfectly. That is to say, the sweet-spot of the main DD is in the exact centre of the coil, but the 'secondary' sensor is perhaps magnetically centred 7 cm to the right. So a small delay process is needed, to match up the readings from the sensors. This would also require an accelerometer/rate-gyroscope in the coil, to determine sweep direction/speed, as the left-right sweep would be processed differently to the right-left sweep.
    This delay-processing would be optimised for a particular sweep-speed, hence swinging too slow or fast may give odd results.

    My guess is it's got Hall sensors, probably several of them, some near the 'toe' of the TX coil, some near the 'heel'. These would be averaged to give a reading of the mid-point of the coils length.

    Anyone got any thoughts, analysis, clever ideas ?

    Do you reckon there's any chance of us creating anything like this ourselves ?
    I read that Mr. Gargov spent many years on the design, so it's probaly not that easy.

  • #2
    The frequency choices appear to have a simple mathematical relationship. (6.4 : 9) is (1 : 1.414); ( 9 : 18 ) is of course (1:2). It's likely he's kept the same inductance, and increasing capacitance 2x , 4x, and 8x to get the stated freqs.
    Example 620 microHenry with:
    120nF => 18.4 kHz
    270nF => 12.3 kHz
    470nF => 9.3 kHz
    1000nF => 6.4kHz
    Latching relays, such as the Axicoms used in the XP Deus would be a practical way of doing this.

    I've searched for Patents with the name Gargov, there are a few related to underground utility location techniques that appear to be related to his regular job.
    Anyone good at Patent searching?

    Comment


    • #3
      If I were making a selectable frequency detector I would not resonate the coil with a cap. Too much trouble, just drive it with a square wave.

      Until the Tarsacci hits the open market, or until the patent app is published, it's all guessing, and the possibilities are endless.

      P.S. -- Dimitar, not Demeter

      Comment


      • #4
        I got his name correct when I searched for patents, obviously wasn't concentrating when I typed that post. Here's ( some of) his patents:
        https://patents.justia.com/inventor/dimitar-gargov

        Comment


        • #5
          Skippy, ask in the Bulgarian forum http://metaldetecting.eu/ if they heard anything about this detector. Dimitar Gargov is a Bulgarian name.

          Comment


          • #6
            According to Mr. Dankowski, the machines name, MDT-8000 refers to 'Multi Domain Technology' , presumably frequency-domain and time-domain. ?

            Comment


            • #7
              I stopped looking for the patent after about a year, it turns out it was eventually published in May 2019.
              There's been a bit of chat about it on the DetectorProspector Forum, so I thought I would track it down. Here's the best I could come up with - there's not a single pdf of it all ( that works ) , I had to select all the relevant sections individually:

              pdf's:
              https://patentscope.wipo.int/search/...TCOLLDOCUMENTS

              html summary:
              http://www.freepatentsonline.com/y2019/0353818.html
              and
              https://patents.justia.com/patent/20190353818

              Seemingly using a square wave transmit, but with some clever signal processing. And quite dependant on the coil characteristics, which may explain the lack of additional coils, interchangable coils, and Tom Dankowski did state the coil and machine were matched.

              Comment


              • #8
                Originally posted by Skippy View Post
                I stopped looking for the patent after about a year, it turns out it was eventually published in May 2019.
                There's been a bit of chat about it on the DetectorProspector Forum, so I thought I would track it down. Here's the best I could come up with - there's not a single pdf of it all ( that works ) , I had to select all the relevant sections individually:

                pdf's:
                https://patentscope.wipo.int/search/...TCOLLDOCUMENTS

                html summary:
                http://www.freepatentsonline.com/y2019/0353818.html
                and
                https://patents.justia.com/patent/20190353818

                Seemingly using a square wave transmit, but with some clever signal processing. And quite dependant on the coil characteristics, which may explain the lack of additional coils, interchangable coils, and Tom Dankowski did state the coil and machine were matched.
                Thanks for posting. Will be a interesting study. At first glance it appears to be BBS or time domain and lots of processing. There is no patent yet, it is pending. This is the application.
                Attached Files

                Comment


                • #9
                  I'm reading through the patent, it's wordy as patents usually are. It is a single frequency square wave drive. The resulting triangle wave current in the coil is distorted into an exponential by the parasitic resistance, and Gargov uses an analog circuit to correct for this. I suspect the correction circuit is hand-tuned to the coil.

                  The target signals end up as exponentials which are usually processed with (discrete) time-domain sampling but they can also be processed with continuous demods, just like a SF sinusoidal detector. In this case, there is an analog "notch" circuit prior to the "ground" demods which subtracts out the resistive component of ground. Then this signal is continuous-wave demodulated to produce a ground channel. The same ground-notched signal is also applied to N target channels, each of which also have an additional target notch filter prior to their demods.

                  The end result is that there is a fair amount of time-domain processing going on prior to the demods, but the demods themselves are run as traditional frequency-domain types. It's a mixed-domain topology, but leans more toward time-domain than not. But it's very much a single-frequency VLF, not PI. Square-wave drive has been done in all multi-frequency designs so far, and the Equinox does single-frequency square-wave drive. I suspect the Rutus Alter 71 does, too.

                  Comment


                  • #10
                    "I suspect the correction circuit is hand-tuned to the coil."

                    I thought perhaps more effectively, the coil could be tuned to the correction circuit, which might make more sense, as it's the coil characteristics that are the main variable here, not the tolerance of R's, C's etc in the 'correction' electronics.
                    So adding a low-value resistor (and inductor ?) in series with the TX coil to increase it to the required value would be viable. These 'coil-tuning' components would be located in the control box.
                    Example: 'correction electronics' is optimised for 3.0 Ohms (coil+cable) resistance and 800 microhenry inductance. Coils are manufactured to nominally be 2.7 Ohms and 770 uH. Sample coil #1 has R = 2.85 Ohms, L = 775 uH. It gets a 150 mR resistor and a 27 uH inductor added as 'coil-tuning' parts.

                    The coil could be fine-tuned, of course. For example if Tarsacci used the method seen on Fisher/Tek 11" x 7" coils, with the seperate 'external' covered tuning compartment. But photos don't seem to show any such 'bulges' on the Tarsacci's coil.

                    Kinda reminds me of matching a turntable cartridge up to your hifi amplifier. Adding load capacitors, and fine-tuning the damping resistance to get a level frequency response.
                    [ For Info: moving-magnet cartridges have a high inductance, typically 0.3 H , and the cable has significant capacitance. This is an undamped LC circuit, so a damping resistor is needed, 47 k is the standard. Then additional C of about 200 pF. The end result is a second-order low-pass filter response, flat in the pass-band, critically damped ]

                    Comment


                    • #11
                      Tarsacci MDT8000

                      https://md-arena.com/gibridnyj-metal...-novinka-2019/

                      Comment

                      Working...
                      X