VLF Newbie

[der_fisherman]

Looking at the first trace, you appear to have a lot of HF noise on the rx coil. What frequency is it? Expand the trace and measure it. This is not good!!

If I am right, you need to get rid of this, or thin down the trace on the scope!! It's far too thick.....

[FreezonePK]

Hello geocommunity,

Sorry I haven't given an update on the progress of my metal detection adventures till now. School and research are bogging me down (last year of engineering).

The coils work awesome, I put 2 more turns on the bucking coil and ended up with a 2-3mV null (maybe too small). Now my big issue is my SNR, after I amplify the signal I end up amplifying a ton of noise which wrecks havic on my logic comparotor network. It is probably because this is being done on a breadboarded as a test. Is this noise filterable with say a 4 or 5 stage(less if I can) chehby low pass filter? If not, I will make use of a synchronous demodulator scheme, but I need to create the simplest detection scheme so that the MCU can do something with the signal. Any suggestions would be awesome and I will post a schematic once I figure it out.

Cheers,

Justin

[Carl-NC]

If you perfectly balance the coil so the residual RX signal is zero, then a simple digital phase comparator might work. But a tiny target signal riding on top of a large residual signal won't give you much of any phase shift.

A typical VLF design uses demods to produce quasi-DC signals and high-pass filters to extract the dphi/dt. This filters out the large residual and all you see is the target signal.

[FreezonePK]

Carl,

This is great encouragement, tomorrow I will get the preamp implemented and let you know how it turns out. Also I will try adding some more turns on the bucking coil and move it around to get a better null.

Is a synchronous demodulator absolutely necessary? I planned to directly compare the receive signal to the transmitted signal using logic gates in which a phase difference would produce to a pulse on its output. The pulse width would correspond to a phase difference. I am up to changing this though, just need to make this decision sooner rather than later. *Edit: I am wondering if this is even necessary because I was just reading it is possible to do the synchronours demodulation with the ATMega32 that I plan to use.*

Cheers,

Justin

[Carl-NC]

Justin,

Looks to me like you have a good coil. If you look really closely, I think you'll see that even small targets are slightly modulating the 5.8mV RX signal. You're not going to visually see small targets (easily) unless you get the nulling at least 10x better. You have about a 40dB null right now, you'll need at least 60dB. A preamp will help you get the better null, and to better see the targets.

A common trick used by manufacturers is to add an extra turn to the bucking coil, then pull that turn off the coil and move it around between the TX and RX coils to fine-null.

VLF detectors use synchronous demodulation to convert the RX signal to quadrature quasi-DC signals. With this kind of design super-deep nulling isn't really required, and even a slight variation of your 5.8mV residual RX signal is easy to pick out.

- Carl

[FreezonePK]

Hello fellows,

4th Attempt the charm? I am almost thinking so. Check my new coils out in which I used 30 AWG wire. The tape is to hold it still after I took out my screw hooks and is only a temporary thing. Hot glue was used to hold the wires together, making them quite stiff. Here are some pictures which I would embedd, but they are huge and it doesn't give them justice to shrink them. Coils O-scope I still haven't implemented the pre-amp, I would of but I was busy building the coils.

If you can't read the numbers on o-scope 800mV in, 5.8mV out. Also I ended up doing a frequency sweep with the signal generator and found that the transmit circuit resonates at 12.4 kHz, not my orignal 8kHz. I figured this resonance can be adjusted by switching my 0.1uF cap to a slightly larger value(0.44 uF I think I calculated). I pull the detector a couple feet off the lab bench and had it so the floor is about 3feet below the coils and started doing some detection tests. What my problem is, is that it still isn't sensitive to small (less than 4") metal objects being placed in its detection area. I am glad though that the receive signal's waveform is way clearer than previous models and I am able to obtain a better null. You can see though that my receive signal is not 60-100 degrees out of phase with the transmit signal (as specified by White). It could be the cap across the receive coil that is causing it. I am able to detect large metal objects (5"x2" and up) but only when it is within 4" of the coils. A metal sheet works the best and causes the receive signal to flip signs.

Factors that I thought could be ruining my sensivitiy : 1) No shielding on transmit and receive wires that lead to breadboard. 2) The lab bench has a grounding mat that is distorting fields in its vicinity up to a couple feet. 3) The floor is ruining the results. 4) Global warming (jk) 5) Transmit coil isn't properly tuned (Hopefully not the case).

I will work on it again tomorrow, but any insight over tonight would be awesome. Thanks again.

[FreezonePK]

Yesm you are correct. I do directly connect it to an oscilliscope. I will try to pass it through a preamp and see if I can tell difference then. I have a feeling I will not, but we will see.

[Carl-NC]

What do you have the coil hooked up to? If directly to an o-scope, you may want to add a preamp to see the small targets.

[FreezonePK]

Carl-NC,

My issue is that I do not see an appreciable phase shift when I introduce a small metal object (1'' to 5'') in the detection area. When I place a large metal object (a metal sheet 6"x6") I do see a phase shift of 180 degrees (the waveform flips with respect to the horizonal).

[Carl-NC]

Justin,

Your White's design should work just fine. Doesn't make much difference where you place the caps.

What problem are you having with it?

[FreezonePK]

@ dfbowers:
Thank you for the suggestion but I am afraid I am rather commited to this particular detection method. It was kind of stupid of me but I constructed a pcb for my circuit design before testing the coils. It is rather more intricate than the testing circuit using an ATmega32 to do some signal processing and other nifty things.

I understand that getting it to work on the ground is very different than in work bench but if I can get it to work in the lab I should be able to compensate for ground effects in the future.

[dfbowers]

Alright, I will post the testing Tx and Rx circuits. I am a shotty explainer, so I am not suprised you are confused. Also,as long as the circuit is fine I have an additional question. Do you place the tuning cap directly onto the coil assembly, thereby making it closer to the coils, or do you put the cap at the terminals of the bread board (away from the coil assembly)?

Thanks again for your help.
Here is the circuit: (Figured how to embed images..Yay)

Making concentric coils are great for learning. I took that route first just to understand how they worked!! Then I realized that I needed to approach my project with some sort of goal in mind, or I was going nowhere fast.

Might I suggest setting your sights a little higher? Building a VLF detector from scratch would be very difficult. Most coils are made to spec for a particular detector, not the other way around.

It's not too hard to get a detector to work on your workbench but getting it to work in the ground is going to be something completely different!!

[FreezonePK]

Alright, I will post the testing Tx and Rx circuits. I am a shotty explainer, so I am not suprised you are confused. Also,as long as the circuit is fine I have an additional question. Do you place the tuning cap directly onto the coil assembly, thereby making it closer to the coils, or do you put the cap at the terminals of the bread board (away from the coil assembly)?

Thanks again for your help.
Here is the circuit: (Figured how to embed images..Yay)

[Qiaozhi]

I am sorry, I meant I put a 100uF capacitor across the recieve coil. And Instead of 44uF I mean 0.44uF capacitors across the transmit and bucking coil.

I decided to redo the coils and more closely follow White's patent.
How does this sounds:
f0=8 KHz.
Tx:69 turns, 10" diameter, L=3.819mH
Bucking: 17.25 turns, 5" diameter L=0.125mH
Rx=207 turns, 5" diameter

0.1uF capacitors across RX and Bucking coil.

If you could response quickly that would be great on this new configuration. Does it sound alright? Is Tx turns correct?

As Porkluvr said, we really need to see your schematic. There appears to be some confusion here, where you talk about putting a tuning capacitor across the bucking coil. Also, the capacitance across the RX coil seems far too large.

Let's see the schematic. Then we can understand what you think you're trying to achieve.

[FreezonePK]

Lol Sorry I meant the balanced search loop for metal detector patent.