metal detecting using rf waves

[deemon]

For my dipl.ing. thesis I made a CW reflectometer for cable tests that preceded anritsu site master for a decade. It employed a bridge that took care of the Tx signal quite effectively. With proper distancing you'll need to reduce Tx by only 20dB or so because it will remain cyclo-stationary, whereas the targets will have responses variation over time as your cart moves.

By the way , I also have a good idea how to make a GPR with single antenna , using a frequency sweep instead of short pulses .... I wanna use an interesting trick for transmit-receive separation . When I finish my PI projects - I will try this , of course .

[deemon]

Craziest way of RF metal detection I've ever seen is “ballistic radar”, device capable to track single 7g cal 303 bullet some 2000m downrange, and measure it's velocity. May sound like high-tech, but it is not, some 30y+ machine. Technical description goes like this: around 150mW GUNN diode source, at 10.5GHz, nice invar resonator for good thermal stability and low phase noise, diplexer and mixer, to produce Doppler signal. Followed by simple but effective filter, few CA3130 amps and precise LC filtering. Computer was based on some obsolete TI 16bit microprocessor, output on LED display and thermal printer (also HPIB). RF head was something like 20cm parabolic antenna, simple portable device most probably modified old live broadcast STL link adjusted for this purpose.

Once I made quite a similar thing , in the middle 90-s . I assembled a little transmitter for the custom-made TV radio relay station ( 10,7 Ghz , 50 mW , GUNN diode oscillator and varactor FM modulator for video signal ) , and I used not a parabolic antenna , but a round horn , connected to the oscillator via a piece of waveguide . An then , when I tested it , I decided to make an experiment ( just for fun ) with Doppler radar . I connected to my waveguide a diode mixer , amplified the signal and fed it to the scope .... and what the funny thing - it worked , and I managed to see a Doppler signal of a little steel balls from a pneumatic pistol . The distance was very little , a few meters ( because of low reflection from those balls and low sensitivity of my diode mixer ) , but I managed to see a clear Doppled signal "burst" on the screen of my scope . Then I tried to connect a ramp generator to my FM modulator ( making a sweep ) and use it like a GPR - but failed , of course . This Ku-band wave looses all its energy in a few centimeters of the ground , it's a pity . So we need to use less frequencies for GPR , of course .

[WM6]

By the way , I also have a good idea how to make a GPR with single antenna , using a frequency sweep instead of short pulses .... I wanna use an interesting trick for transmit-receive separation . When I finish my PI projects - I will try this , of course .

At which frequency band do you plan to try with this project?

[deemon]

At which frequency band do you plan to try with this project?

I want to start with a frequency band across 500 MHz .... of course , it's a kind of compromise here - less frequencies needs too big antennas and has quite a poor depth resolution ( wave is too long ) , but higher frequencies cannot penetrate deep to the ground ( due to high energy loss ) . So , maybe 500 MHz range is a kind of optimal .... but of course , experiment will show , anyway .

[Davor]

@WM6, frequency wouldn't matter that much, in principle it would work at any band, only the size and depth are affected.

One funny thing about Doppler is that it works on a zero IF principle, so in effect it enables incredible gains virtually unaffected by Tx amplitude, so even the noisy sources as gunn diode makes it quite possible. Speed of light makes it time variant, but at small speeds the approaching or departing target give virtually the same frequency response. If you observe it more thoroughly, it is the phase you are after, not the frequency, and phase response is exactly the same as for the standing waves on a line. If you can imagine two passes over the same environment, only the second one twice as faster, you'll have exactly the same phase response for both of them on a whole trajectory, which makes a speed of cart invariant.

I think daemon knows exactly what I'm talking about.

The best part is that you don't need any directing arrays, as they become effective only if given ample free space, typically 6D (effective antenna height), because this principle equally functions in near field.

Most probably there are some other antenna experts here who know very well how messy high gain antennas perform in near field. Perhaps the tidiest would be a short dipole.

[Funfinder]

interesting topic and fascinating that many metal-detector-electronicians also have experience with ham radio etc.
You should have seen my crazy spiderweb antennas I installed into some trees as child.



The problem is not really the frequency, but the signal-noise ratio.

Some persons nowadays can use there normal medium-sensitive detectors just at half recognition-power
because of the electro-smog interferences or the too heavy mineralized ground.
They wanna have extremly clear signals and identification-response, but they lose alot depth.

How can you improve clear signals?
By more sophisticated "metal-or mineralization value distinction"-methods:

phase shift, frequency shift, transmission energy losses, detection of weak eddy-current effects and side-effects and magnetical interference values

Another huge factor is the users skills and experience AND using different detectors in combination.

Metal detection works somekind like wireless energy transfer:
The more directional and powerful the magnetical loop pulses antenna the better the response or signal to valuable.


The soil or "ground" often also is the "electrical ground", it shortcuts all "positive" energy and sends it to "hell".

Or it reflects it - like most higher frequency waves. Even mountains will reflect radio-waves, depending on their mineralization.

Listen radio behind or infront of a brickwall and you will see the difference.



Frequency from 1khz to 1Mhz would work for underground metal detection
but the higher frequencies which are better for small finds must be compensated
with higher power and this could reduce the battery life extremly.

Another possibility to improve the detection would be creating a multi-segment antenna,
higher than the usual 3 point detection method with transmitter, receiver and "man in the middle" aka find.

The easiest form is a special loop antenna with 5 or more rings where each of those
are used in transceiver status and each of them are able to detect the impedance differences
to their neighbor coils.


Radio transmission ground injection works similar.
Imagine a ring of injectors and the receiver in the middle - 25 meter radius.
Of course that way the receiver can find large metal objects inside that area
because of the changing energy values depending in what direction the receiving-antenna (the treasure locator) moves.

Because the electrical (voltage, potential difference) part of the energy mostly gets soaked up, shortcuted or reflected by the ground,
the magnetical part of the EM-pulses is very important for detection because it penetrates the ground much better,
depending how strong "magnetical mineralized" (iron-oxide etc. minerals) the soil is.

Thats why magnetometers are more sensitive than metal detectors and so
its very important to detect very weak magnetical differences, for real great depth.

Another interesting multi-segmented coil design would be 9 powerful ferrite antennas
in a 3 x 3 grid, perhaps in a square with a size of 50 x30 cm for better sweeping efficiency.

Of course each of those powerful high-inductive ferrite antennas also should work
as transmitter-receiver and sends the sensed field-conditions to the computer-unit.

Ferrite antennas set to borderline sensitivity contain a huge power potential for detection,
I see that everytime when I work with my modified Garrett Pro Pointer, and that is just
using a very small standard ferrite-coil - at around 100kHz.

With higher power such a multi segmented coil, contain 9, 15 or even more single ferrite coil cores,
extremly sensitive and detailed detection results could be made possible, even at 1 MHz frequency
for much better directional power - for sending extremly tough and sharp magnetical directional pulses into the soil!

Have fun with this inspirations and genious ideas!

[Davor]

The soil or "ground" often also is the "electrical ground", it shortcuts all "positive" energy and sends it to "hell".

Or it reflects it - like most higher frequency waves. Even mountains will reflect radio-waves, depending on their mineralization.

True, but only in some special cases. Some of these effects are a bit deceptive as there is a big difference between a far field planar wave and a near field we are interested in.

Regarding reflection, you have actually two distinctive effects, of which neither works completely intuitive. Somewhat easier to grasp is a reflection off of a perfect conductor, e.g. metal object with dimensions larger than half wavelength. With enough area it works as a mirror, but as dimensions go down it continues scattering electromagnetic energy even at ridiculously small sizes. E.g. bb gun pellets against a decimetre wave Doppler radar. That's a very helpful feature when considering GPR and, say, gold nuggets.

The other effect is completely counter-intuitive as it deals with dielectric properties of soil. It is true that radio waves reflect against the soil, but only if the incident angle is smaller than the critical angle (Snell's law of refraction...). It means that there is incredibly less reflection from the ground for near perpendicular emissions. Attenuation is a completely different animal, but ground as a "conductive ground plane" is an over-idealised concept.

The most counter-intuitive concept is antenna gain. It happens in far field. Let me paraphrase this. Gain of antenna happens EXCLUSIVELY in far field. A concept of effective antenna height deals with these effects. In effect only antenna array may have gain over a simple dipole. Even antennas with reflectors may be observed as antenna arrays as the reflector acts as a mirror, hence you have an array formed by the exciter dipole and it's image. All antenna arrays have poor phase response and lower amplitude in near field. Far field conditions start forming at distances roughly 6 times the effective antenna height, which increases with gain. Hence antennas with small effective height are the best choice for close proximity.

[Funfinder]

thx for the very interesting answer

The main issue we have is to think in "micro" versus "macro" dimensions.

Powerful radio program long waves interact different with ground plus ionosphere than metal detector longwaves
that penetrate the surface of the ground almost directly from very low distance by vertical transmitting DD magnet-coil-antennas.

The electrical ground not alway reallys is grounded, a sat-dish reflects the microwaves into the LNB even if it mounted on wood or brick roof
but still represents the "ground".

As you, Davor, told already correctly, it highly depends on wavelenght and reflectors size, too, what the electron-mass-wave-complex does.


Stupid example:
If we put a styropor cube on the ground and put a metal object inside the detector can finds it like in air.
But the more moisture and mineralic or even "metal-combinations" the soil contains, the higher it gets saturated and interferes with the injected EM field.

- Some energy will be absorbed and creates microminimal heat (like moving electrons in wires)
- some energy will be reflected
- some flows away sidewards or deeper
- and some creates detectable patterns or changes on a detectable level the field-strenght etc. (metal-object vs ground contrast)


If the ground would consist of styropor we could get the best possible sensitivity!

But we have to deal many times with EM-energy distorting moisture, electrolyts, soluted minerals plus sensitive detection circuits "overriding" electro-smog.

So far the only solution for better depth was enlarging the coils or more precise: more distance between tx rx coils.

Because intermediate induction value of coil a versus coil b is higher if those are more distant - the measureable area !!!

The contrast factor!

So the easiest way to improve depth would be higher sensitive and more stable circuits who can filter out more contrast
OR
more directional antennas for better, "sharper" depth EM-energy streams.

The best thing would be electron magnetical resonance but this is too expensive.
So we deal with "metal find resonance".

We can get sharper pulses by higher voltage, frequency and directional antennas.
By high effective directional magnetical pulses for whatever magnetical targets.


Seen on the consumers level "simple turn on and go"-detectors are prefered.
Many don't like 40cm plus coils, they are simply too heavy for them, even if they are sensitive for coins etc., too.

But how to achieve high contrast if rx tx coils are very close together - for really extrem depth!?

We would need 10x more powerful energy or batteries than now for higher frequencies (for their penetration energy loss compensation)
so they would penetrate on a much sharper and therefore much better distinctable contrast factor level.
Even wave-polarization changings could be detectable on higher frequency.

And just 2 close together DD coils are simply too few for better "resolution".

The multi-frequency setup is a nice idea, but 2 frequencies at the almost same time would be enough (deep plus small find optimization)
but the coil itself has to be good enough for this task, too. A 1m coil at even 500kHz for gold-nuggets won't work.

A coil with more than 2 different windings will be a must, anyway, for deep plus small thing detecton at the same time.

Per instance in the middle a 10cm DD coil driven at 100kHz for the very small and shallow finds and
on the outside ring coils with 60 and 50 cm diameter for the large deep stuff - at 5kHz.

But this is just a very short overview now - and seen from electronic-level or computer level high sensitivity tec
there is alot air upwards for improved signal-noise ratio definitions.


And last but not least:
The treasure hunter is the limit!

Finding a golden coin in 1m depth is great, but the treasure-hunter also has to like to dig that deep!

good luck

[Davor]

I wouldn't dismiss 100kHz easily. It is well within the useful range of frequencies where eddy currents produce recognisable signatures for most metals, and that particular part of LF band is very quiet, so it can provide the otherwise elusive advantage in metal detecting.

[ant17]

metal detecting with rf waves

I wouldn't dismiss 100kHz easily. It is well within the useful range of frequencies where eddy currents produce recognisable signatures for most metals, and that particular part of LF band is very quiet, so it can provide the otherwise elusive advantage in metal detecting.

hi guys today i made a crude experiment to see if i could a get some result i first made a colpits oscillator with antenna running at 20mhz and then a reciving coil anttena set up so with the 2 anttenna facing each other about 2feet apart i checked the reciver coil waveform and found i got my sinusoidal waveform which seems to change frequency and amplitude when i run my hand or metal target through between the anttenas and cant tell from the waveform which is metal which is not is there any way i could descrimnate or am i wasting my time hope to here your thoughts

[Davor]

So far the easy and non-exotic approach to discriminating metals is by eddy current decay and also permeability in ferrous materials. Both effects tend to vanish with high frequencies, and GPR-s are useful to indicate presence, not composition.

Most coloured metals have eddy current time constant from few to few hundred microseconds. MHz frequencies are too far.

[ant17]

metal detecting with rf waves

So far the easy and non-exotic approach to discriminating metals is by eddy current decay and also permeability in ferrous materials. Both effects tend to vanish with high frequencies, and GPR-s are useful to indicate presence, not composition.

Most coloured metals have eddy current time constant from few to few hundred microseconds. MHz frequencies are too far.

any body else have an opinion