2
\$\begingroup\$

I've finally started work on my metal detector project and after a fair start, I'm a bit stuck.

The circuit is based almost entirely on the buccaneer VLF circuit to be found here. I've made some alterations, I've upped the voltage to 9V and made the necessary adjustments to account for that. I've also increased the coil diameter to 20cm (the original was 11cm), but I've kept the inductances more or less the same (my coils measure 2.9mH, whereas the original was somewhere in the region of 2.3mH). With the Tx and Rx circuit built (on the breadboard - yes, I know it's not the best, but I didn't want to commit to a full build just yet) and the Tx oscillator up and running, I tried nulling the Rx coil. I must have spent hours trying to get near where I thought I should be at (millivolt level) without much success.

In the screenshots below, I set the p-p Tx voltage at about 40V by trimming R6. This is considerably less than the expected output of the original circuit (80V on my simulations) and yet the Rx coild is showing just over 2V p-p (1.8Vrms), which is certainly a far cry from the millivolt level I was expecting.

The give away here, is that in the original circuit the pre-amp gain is set to be roughly 10. When I tried that kind of gain, my signal railed badly, so I reduced the gain to 5 but it's still clipping.

So, my first question is this: just exactly what should I expect when I "null" my coils and do my results look normal?

Secondly, I wasn't expecting a perfect induced signal but the Rx coil has these annoying witches hats that appear at twice the operating frequency. If you look really closely at the Tx signal you can just make out the glitch that may be coupled into the Rx coil, but I'm not sure. What can be causing this? One thing to note is that the "hats" move around the signal as I squeeze the coils to try to null them. It's really screwing up my Rx signal.

I know there's a lot to take in here and if you need more info I'll post that up.

UPDATE: Incidentally, for those who'd like to know more about the Tx Oscillator, see this post

enter image description here enter image description here enter image description here enter image description here enter image description here

\$\endgroup\$
  • \$\begingroup\$ Why did you turn one of the transistors upside down, compared to the original schematic? You can see distortion on the first scope waveform which are the likely cause of the distortion products on the second. \$\endgroup\$ – Brian Drummond Nov 4 '16 at 16:38
  • \$\begingroup\$ Hi Brian, which one? It's probably a transcription error I checked the actual circuit pretty thoroughly. \$\endgroup\$ – Buck8pe Nov 4 '16 at 16:41
  • \$\begingroup\$ I tend to document my circuits and notes, etc in TinyCad - just out of habit. \$\endgroup\$ – Buck8pe Nov 4 '16 at 16:42
  • \$\begingroup\$ Yep, I see it now...the push-pull pair right? Let me fix that up. Tks. Good eyes;-) \$\endgroup\$ – Buck8pe Nov 4 '16 at 16:45
  • 1
    \$\begingroup\$ Shouldn't the receive coil be earthed - as it stands it's just a floating loop with a single point connection to an input. BTW, which transistor was upside down and have you fixed the pictures or is it still distorting? Ditto the transmit coil - you won't get any power through that coil as it stands. \$\endgroup\$ – Andy aka Nov 4 '16 at 17:04
5
\$\begingroup\$

The main problem is this: -

enter image description here

In case you can't read my red writing, you are missing a connection from the receive coil to earth.

As for screening, yes, you need it to avoid the search head picking up silver foil confectionary wrappers and other useless bits of stuff you want to ignore. Salty conductive soil is also an annoying problem that is largely solved by a coil screen.

Balancing - if you are still having problems, try a little bit of ferrite core placed adjacent to your coils to get the best "balance minimum". Then, try a little piece of non-magnetic stainless steel to try and take the minimum even lower.

Ferrite is non conductive so you can use it to basically balance the inductive artefacts. SS is good because it's quite resistive and although it will upset the inductive balance (due to eddy current induction) those eddy currents will be quite resistive in nature. Thin bits of tin foil (basically what you see in a cigarette packet) might also be useful for helping to balance the coils.

So, messing around with ferrite and SS (and or tin foil), will give you an idea where you could get by mechanically bending the coils a bit this way or that way. But you can always solidly glue down a piece of ferrite in the search head to avoid all this messing around.

The screening/balancing information provided in this answer is not anyone's IP - I learnt it designing metal detectors for food/pharmaceutical product lines.

\$\endgroup\$
  • 1
    \$\begingroup\$ If I get a chance I'll update this question with images of what a balanced receiver signal looks like, since I noticed that there isn't a whole lot of that around. \$\endgroup\$ – Buck8pe Nov 4 '16 at 17:39
0
\$\begingroup\$

TX coil connects to detector and RX does not. The ends of the RX coil are directly connected to the millivoltmeter. Leave the detector in operation and slowly move the RX spool until you get the least voltage, in my case it was 0.04 mv, but I read somewhere that it was ok and 15 mv. It should protect both coils with a metal foil, but so that the foils between the spools do not touch and that both are connected by cable to the mass of the detector. I use old USB cables because they have an external protection I use for that. on the spools leave a small slit (5-10 mm). I am nullifying this and I made my probe and I have detection for 1e-35cm. Excuse me for bad English Please see the picture attached enter image description here

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.