Announcement

Collapse
No announcement yet.

Coil calculations, finding RX value

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

  • #16
    Originally posted by golfnut View Post
    I hazard a guess that the Rx coil does not have to resonate with a cap 1x, 2x, or Nx the Tx frequency.

    The fact that its 'receiving' should be good enough. Anyone up for taking their cap off to see impact.

    S
    Just done it Golfnut it reduces detection by about half on my machine

    Comment


    • #17
      Originally posted by simonbaker View Post
      Hi kliner:

      Would you clarify what you mean by "TX is supposed to be at resonance"?

      Regards,

      -SB
      Hi Simon
      I mean that we think we are at resonance and due to an acumulation of errors/component tolerances we could well be a fair way away.
      Most people have only got a voltmeter to set stuff up

      Comment


      • #18
        Originally posted by kliner View Post
        Hi Simon
        I mean that we think we are at resonance and due to an acumulation of errors/component tolerances we could well be a fair way away.
        Most people have only got a voltmeter to set stuff up
        I think you are saying we may not be at the resonant frequencies we think we chose -- very true, some tuning is indeed something we can perform.

        I used a capacitor bank to try to tune my TGSL, but actually didn't achieve much improvement. However, now I know that local EMI noise was probably killing the real depth in my workshop, so all the tuning was futile.

        I was very lucky that dfbowers sent me one of his excellent handmade TGSLs. That gave me a perfect way to confirm various measurements and see the effects of the EMI noise in my workshop.

        Regards,

        -SB

        Comment


        • #19
          Originally posted by kliner View Post
          If you change the RX capa value significantly the voltage on the RX side goes right up and the detection depth is reduced.
          When you look closely the TX is supposed to be at resonance, but I'll bet by the time you take into account the tolerance of the components and the hand wound coils not strictly at the right inductance, (not talking about temperature changes) the frequency is probably a fair way away. Does this matter? it surely doesn't help matters. Perhaps we should look more closely at tuning and stabilising our TX circuits. Perhaps this is one of the reasons we often fall short of the professionally built detectors.
          One of the advantages we have with respect to the pros is that we can afford to spend a lot of time tuning and matching our coils to our detectors.
          True to my findings yesturday with the RX/CX Findings were as CX was increased so was the amplitude reading off the RX preamp which in turn was reducing detection level.
          Still running with 15NF CX even though technically should be 20NF due to inductance of my coil, as soon as increased by small 800pf cap was noticable loss of detection depth, and the more loaded onto it the worse it got.
          Now by tweeking Ivonics preset TX mod was able to get rid of that horrible blip on the right hand top side of sinewave which does clean up the horrable chirps at distance detections.
          So im feelings are the need to maybe gently drop the CX even more to see if theres anything else to be had.
          Can I ask is it the lwr the amplitude on the RX preamp the better the operation or am barking up the wrong tree.
          Which brings me to another question is there a test point on the TGSL that we can measure CX for 16.1KHZ while fully connected, because otherwise its working almost blind, as kliner quite correctly said theres other parts in the circuit, even down to type of PCB board been used because print gravity, flux residue or even the sweat of your hands while construction was under progress, its all very well cleaning with esponal but you cant clean under components and any or all could be throwing the frequency off and theres know way of knowing.
          Anyway lets please try and stay on this particular part of the circuit because maybe this is where the variations of good and not so good builds lay. Im willing to do raw tests to confirm if you prefer doing the spice business because want us to route this out for once an all, know stone unturned as they say.

          Warm regards

          Comment


          • #20
            Originally posted by satdaveuk View Post
            True to my findings yesturday with the RX/CX Findings were as CX was increased so was the amplitude reading off the RX preamp which in turn was reducing detection level.
            Still running with 15NF CX even though technically should be 20NF due to inductance of my coil, as soon as increased by small 800pf cap was noticable loss of detection depth, and the more loaded onto it the worse it got.
            Now by tweeking Ivonics preset TX mod was able to get rid of that horrible blip on the right hand top side of sinewave which does clean up the horrable chirps at distance detections.
            So im feelings are the need to maybe gently drop the CX even more to see if theres anything else to be had.
            Can I ask is it the lwr the amplitude on the RX preamp the better the operation or am barking up the wrong tree.
            Which brings me to another question is there a test point on the TGSL that we can measure CX for 16.1KHZ while fully connected, because otherwise its working almost blind, as kliner quite correctly said theres other parts in the circuit, even down to type of PCB board been used because print gravity, flux residue or even the sweat of your hands while construction was under progress, its all very well cleaning with esponal but you cant clean under components and any or all could be throwing the frequency off and theres know way of knowing.
            Anyway lets please try and stay on this particular part of the circuit because maybe this is where the variations of good and not so good builds lay. Im willing to do raw tests to confirm if you prefer doing the spice business because want us to route this out for once an all, know stone unturned as they say.

            Warm regards
            I spent several hours the other day trying to see how the TX frequency affected the detection depth and the RX voltage, (the RX frequency is always the same as the Tx of course) Well I couldn't find a sharp peak but as I changed the TX capa (either adding or subtracting) there seems to be a plateau where things change slowly, then the RX voltages rises, the more it rises the less good is the detection depth.
            I thought the same thing as Simon, what I would need to do would be to get capacitance decade box and do a proper plot to see exactly what is going on.
            So it seems that Satdave is right when he says the detection depth is best when the RX voltage is low. (Are we swamping the system?)
            Of course the TX voltage also has an effect on the RX voltage even on a well nulled bought coil, I thought I would use Ivconics tweakerpot to push up the TX voltage to around 20 v thinking (naively) that I might get a bit more depth, well no 'cause the RX voltage rises and you start to get problems with the GEB as well.

            Comment


            • #21
              Originally posted by satdaveuk View Post
              True to my findings yesturday with the RX/CX Findings were as CX was increased so was the amplitude reading off the RX preamp which in turn was reducing detection level.
              Still running with 15NF CX even though technically should be 20NF due to inductance of my coil, as soon as increased by small 800pf cap was noticable loss of detection depth, and the more loaded onto it the worse it got.
              Now by tweeking Ivonics preset TX mod was able to get rid of that horrible blip on the right hand top side of sinewave which does clean up the horrable chirps at distance detections.
              So im feelings are the need to maybe gently drop the CX even more to see if theres anything else to be had.
              Can I ask is it the lwr the amplitude on the RX preamp the better the operation or am barking up the wrong tree.
              Which brings me to another question is there a test point on the TGSL that we can measure CX for 16.1KHZ while fully connected, because otherwise its working almost blind, as kliner quite correctly said theres other parts in the circuit, even down to type of PCB board been used because print gravity, flux residue or even the sweat of your hands while construction was under progress, its all very well cleaning with esponal but you cant clean under components and any or all could be throwing the frequency off and theres know way of knowing.
              Anyway lets please try and stay on this particular part of the circuit because maybe this is where the variations of good and not so good builds lay. Im willing to do raw tests to confirm if you prefer doing the spice business because want us to route this out for once an all, know stone unturned as they say.

              Warm regards
              Can I ask is it the lwr the amplitude on the RX preamp the better the operation or am barking up the wrong tree.
              No, I don't think that is true in general, but it depends on what you mean. "Amplitude" is probably the null signal amplitude. Certainly nulling your coils well, which implies lowering the amplitude, helps. Changing the RX capacitor will affect the null amplitude also, but I don't think lower is better; you also reduce the target signal gain that way;rather there is some point which works best, more or less, for the TGSL.

              Which brings me to another question is there a test point on the TGSL that we can measure CX for 16.1KHZ while fully connected
              Do you mean measure CX or measure the RX tank resonant frequency?

              Regarding the resonant frequency: not directly that I can think of. LF353 pin 7 is a good place to monitor the RX signal; you could try varying the TX frequency until you see a peak voltage on that pin, but you have to make sure the TX amplitude stays pretty constant. There are other gyrations you can do with a signal generator to try to locate the RX resonant frequency. I think good RX inductance and capacitor measurements and a calculation is the most feasible technique, if you trust it.

              To measure CX I think the only feasible way is to unplug the RX coil and measure across the capacitor (same as unconnected RX cable leads).

              Now by tweeking Ivonics preset TX mod was able to get rid of that horrible blip on the right hand top side of sinewave which does clean up the horrable chirps at distance detections.
              Not clear why that cleaned up the chirps. It might just be because you lowered the TX amplitude and thus have less target "gain" in general.

              It is so hard to draw conclusions from simple tests because a lot of variables are involved. Practically, I think it's worth tuning and seeing what happens with each MD and mainly just get the best depth you can while keeping the DISC and GB functions working properly.

              It probably takes a lot of really controlled tests to check out all the relationships to come up with a design formula that allows us the most freedom to pick our parameters and still get good performance.

              However, I agree this is what it is all about, so I'll certainly keep it on my experiment radar.

              It is extremely important to do any testing in a noise-free environment. Otherwise, we might change a capacitor that changes the TX frequency and get less depth and conclude that the higher frequency is less good, while actually we just stumbled into a noise band, for example.

              Regards,

              -SB

              Comment


              • #22
                Originally posted by kliner View Post
                I spent several hours the other day trying to see how the TX frequency affected the detection depth and the RX voltage, (the RX frequency is always the same as the Tx of course) Well I couldn't find a sharp peak but as I changed the TX capa (either adding or subtracting) there seems to be a plateau where things change slowly, then the RX voltages rises, the more it rises the less good is the detection depth.
                I thought the same thing as Simon, what I would need to do would be to get capacitance decade box and do a proper plot to see exactly what is going on.
                So it seems that Satdave is right when he says the detection depth is best when the RX voltage is low. (Are we swamping the system?)
                Of course the TX voltage also has an effect on the RX voltage even on a well nulled bought coil, I thought I would use Ivconics tweakerpot to push up the TX voltage to around 20 v thinking (naively) that I might get a bit more depth, well no 'cause the RX voltage rises and you start to get problems with the GEB as well.
                Now that is interesting because with experimenting with the TX output using the 14.536khz frequency was that when tweeking Ivonics preset for best results bring the amplitude up and keep a eye on that blip on the top right of the sinewave and you will find as you start to increase the amplitude you gradually loose the blip and smooth it out, then carry on increasing and the blip starts to appear again, now... its that medium of no blip and the nice near perfect sinewave with preset just over halfway at 15v p-p thats option performance where you have max amplitude without the blip, go over that and you loose depth plus your having to alter ground balance.
                When doing the same procedure running at TX 14.38KHZ option performance was, max amplitude 20v p-p with the blip still there, which brought the preset in bang on halfway.
                These tests I talk about will always be with the factory 8" brown round coil ok, save me repeating myself.
                The performance running in and outside my workshop in ground conditions on both frequencys with those settings appeared to be identical apart from the output tone been that much higher which is to be exspected, controls stay the same sense almost maxed out, disc switch on, ground balance 5 past the hour.
                Will post pics later taken from the scope that relates to the tests.
                Im wondering that maybe the TGSL design using lower TX frequencys can handle more amplitude than higher, its just a wonder maybe.
                still wish there was away to read the true RX operating frequency ie maybe 16.1khz there must be away round it, I say maybe but should say well in theory it is but as we all know we couldnt do without theory but it dont mean its always right proved many a time to be wrong.
                When time permits ill go back to the workshop and gently drop the CX and let you know the results.
                By the way a tip for along time now due to experiments on the LX/CX I have the cap for rx off board mounted on the back of the socket, im going to do the same with the TX caps it saves hammering the PCB and its alot easer to chop and change.
                All the best
                Regards

                Regards

                Regards

                Regards

                Comment


                • #23
                  Originally posted by kliner View Post
                  I spent several hours the other day trying to see how the TX frequency affected the detection depth and the RX voltage, (the RX frequency is always the same as the Tx of course) Well I couldn't find a sharp peak but as I changed the TX capa (either adding or subtracting) there seems to be a plateau where things change slowly, then the RX voltages rises, the more it rises the less good is the detection depth.
                  I thought the same thing as Simon, what I would need to do would be to get capacitance decade box and do a proper plot to see exactly what is going on.
                  So it seems that Satdave is right when he says the detection depth is best when the RX voltage is low. (Are we swamping the system?)
                  Of course the TX voltage also has an effect on the RX voltage even on a well nulled bought coil, I thought I would use Ivconics tweakerpot to push up the TX voltage to around 20 v thinking (naively) that I might get a bit more depth, well no 'cause the RX voltage rises and you start to get problems with the GEB as well.
                  We should really try to measure some relationship of TX frequency to RX resonant frequency when we do such tests, because that may be what really matters.

                  You have to check a couple of things to see if you are "swamping" the system. One is the output at LF353 pin 7 -- if it clips, we've swamped. Also check the DC voltages on C15 and C12 over the range of the DISC control and at the GB setting. If it goes below -.5, it could be messing up the JFets of the Synchronous Detector.

                  Otherwise, a larger null signal is not necessarily bad I think. But by raising the TX frequency, we approach the RX resonant frequency, which causes increased gain and more phase shift of our null and target signals. Both of those can cause some problems.

                  Too much gain probably messes up our target detection further down stream. It might be interesting to put a pot for R15 (the 220k feedback resistor on U101a) so we can adjust the gain as we change the TX frequency to keep it constant, and see what effect that has.

                  Too much phase shift will decalibrate the GB and DISC controls, so you really should readjust those as you play with the TX freq. I think the GB channel is the limiting factor for depth -- it always seems to have a smaller target response -- so that can be affected by phase shifts easily. It is probably not even possible to adjust for large phase shifts as you raise the TX freq.

                  So there are a lot of things going on when you adjust the TX frequency. If we are willing to modify the circuit itself, perhaps there are other TX/RX frequency relationships that might be very useful.

                  Regards,

                  -SB

                  Comment

                  Working...
                  X