Announcement

Collapse
No announcement yet.

Inside the TX-2002 pinpointer

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

  • Inside the TX-2002 pinpointer

    Hi all,

    It was suggested I open a thread about this unit, as it is possible that it could be hacked modified or otherwise improved (and it needs some...).
    I'll post some more images in the next post.

    The TX-2002 is a cheap Chinese pinpointer made by http://www.shtianxun.cn/infoi.aspx?Class_id=36

    I bought it to see if it was any good for pinpointing at the beach, but really I could live without one there.

    My opinion of it is not high, it seems to be a motion detector (needing constant moving to keep indicating a metal target) and seems to be very insensitive. With an Australian $1 coin in air I get at very best 20mm with the coin flat side on, and only 3mm with it edge on. So most of the time I can see the coin before I can detect it

    I've tried a few tweaks to see if I can more sensitivity out of it, but no. I made a trivial increase by adding an extra 22nF cap across the search coil (two wire, ferrite rod type, 2 ohms). Also tried other things, including trying about four different brand LM358's I had in my parts bins. I noted that the unit could have a vibrator motor installed on it, the tracks and component pads are there, and the buzzer on/off switch is actually a change-over that connects to the vibrator circuit. But I see little point unless sensitivity is improved...

    I made a start at creating a schematic, but as it's not a trivial circuit I stopped after realised I may actually replace the whole board in the future anyway.
    It may be a good cheap candidate to put in the circuit described here http://www.geotech1.com/forums/showt...tector-project

    The search coil L value I'll measure in a few days (when I can get to an L meter). The coil has a 22nF cap in parallel with it forming a resonant circuit. The oscillation frequency is around 20kHz when measured on the hot side of the coil, not 90kHz like the user manual says !
    (The CRO image frequency was with the extra 22nF cap fitted BTW). The oscillation frequency drops slightly (say around 100-200Hz) in the presence of metal, but the main circuit seems to track it so unless you move the metal out of the search coil field again you don't get another beep out of the buzzer.

    If anyone knows more about the unit please add comment posts. I thought it was going to be like a Cen-Tech circuit, but it's not.

    Cheers,
    Martin


    As it was when first opened (prior to my first molestations).


  • #2








    I added the green IC socket and the two pin header socket myself for easier bench testing.








    The brown ceramic SMD cap is my addition, but it's so trivial an 'improvement' I'd just ignore it.








    The manual says it's a 90kHz device but the CRO reads 20kHz unmodified, and 17kHz after I added a bit more C across the coil.

    Comment


    • #3
      Hello!
      Good teardown,thanks.
      Based on your info the coil inductance is about 2mF. For example use: http://www.calctool.org/CALC/eng/ele...cs/RLC_circuit

      Best wishes

      Comment


      • #4
        Originally posted by Teemo View Post
        Hello!
        Good teardown,thanks.
        Based on your info the coil inductance is about 2mF. For example use: http://www.calctool.org/CALC/eng/ele...cs/RLC_circuit

        Best wishes
        For a frequency of 20kHz and a capacitor value of 22nF, the inductance of the coil is 2.88mH.

        The equation is:

        where:

        Comment


        • #5
          Of cource mH . I took 44nF and 17kHz for calculation.

          Comment


          • #6
            Seems like a lot of parts for a poorly functioning unit. 1% resistors too. I wonder why it doesn't work better?

            Comment


            • #7
              Originally posted by Silver Dollar View Post
              Seems like a lot of parts for a poorly functioning unit. 1% resistors too. I wonder why it doesn't work better?
              My thoughts too. Seems like a lot of trouble for something that works so badly (IMHO). I wonder if the makers have substituted a cheaper higher L coil (hence 20kHz) and that's buggered it up.

              Now I'm a newbie when it come to detector circuit design and don't recognise it as anything I've seen before, so if anyone can suggest a circuit that this might be based on I can try to confirm or dismiss it.

              Also, if anyone is keen and knows how to create a schematic from a pseudo net list (in Eagle etc), I could desolder the whole board. Documenting the part numbers and values in an XL table and then repost bare board images and the table ?
              I'll only bother if somebody IS keen to do the CAD schematic side, as I don't have time for both.

              Teemo,
              is there much hope that your PIC design can use the coil as it is (hi L value) ?
              The coil is sealed with hot melt glue under the cap and some sort of glue down the sides, and I suspect the probe case will break if I try to extract it.

              Cheers,
              Martin

              Comment


              • #8
                It has a 358 and a 4011, from what I see and a single transistor. Not going to do much with that.

                Comment


                • #9
                  Originally posted by mr.sneezy View Post
                  Also, if anyone is keen and knows how to create a schematic from a pseudo net list (in Eagle etc), I could desolder the whole board. Documenting the part numbers and values in an XL table and then repost bare board images and the table ?
                  I'll only bother if somebody IS keen to do the CAD schematic side, as I don't have time for both.
                  If you're prepared to desolder the whole board and document the component values, etc., then please go ahead. I will create an area in the Project section where we can put the data. Then, as others get time to work on it, further information can be added. It's about time we had a small group project here.

                  Comment


                  • #10
                    Originally posted by Thomas View Post
                    It has a 358 and a 4011, from what I see and a single transistor. Not going to do much with that.
                    I was hoping someone would have challenged me on that. oh well.

                    Comment


                    • #11
                      Well I see at least 3 transistors but ignored your comment.

                      The real problem is here;

                      The MD is missing. Just plug in a metal detector there and you'll be all set!
                      Attached Files

                      Comment


                      • #12
                        And it's surprising that no-one has noticed this -> Click image for larger version

Name:	Spelling mistake.jpg
Views:	1
Size:	43.1 KB
ID:	338917

                        Comment


                        • #13
                          That's how you can tell it's a Chinese knock off...

                          Comment


                          • #14
                            Originally posted by mr.sneezy View Post
                            Teemo,
                            is there much hope that your PIC design can use the coil as it is (hi L value) ?
                            The coil is sealed with hot melt glue under the cap and some sort of glue down the sides, and I suspect the probe case will break if I try to extract it.
                            No my circuit can not use that coil. L is too high.
                            Teemo

                            Comment


                            • #15
                              Thanks Teemo,
                              Is it fundamentally too high for a PI technique, or is there another reason ?
                              The thing being is if it's too high to use your PI mode then I'm going to have to make a different PIC circuit for it, or fix the original.

                              BTW, am I right in thinking that it's a VLF type unit (TX-2002), and the slight frequency offset with metal is the likely detection method ?

                              I'll try to strip the board in the next day or two and document regardless.

                              Comment

                              Working...
                              X