Off-resonance revisited

[Davor]

I browsed through the forum to find only a precious little information about the off-resonance solutions. There is an off-resonance pinpointer solution in Carl's and George's book, but apart from it only a few other circuits ever lived. Perhaps it is not giving justice to this approach.

Off-resonance will hardly ever become a big project, but can make for a good pinpointer. That said, I'm imagining a pinpointer with better emphasis to metals, especially poor conductors.

Commonplace are the loaded oscillator pinpointers. They are simple enough, and easily repeatable, so obviously good candidates for mass production. They operate in a way that coloured metals reduce the amplitude of oscillation, while ferrous materials, including earth, increase the amplitude. This happens because the eddy currents response of coloured metals is shifted over 90° against the oscillator phase, and this view reveals the weakest point of these detectors - they suck at poor conducting metals. Poor conductors have eddy current response near 90° and as such have negligible influence on oscillator amplitude. This problem can be partially cured by employing high frequency oscillator, 100kHz or so, where even poor conductors response is shifted towards 180°.

Off-resonance operates by converting the resonant tank phase slope to amplitude. This way or another. In ITMD book solution the oscillator feedback reactance is forcing the tank off-resonance, hence the presence of a metal with eddy current response near 90° will also cause amplitude change, because oscillation of such oscillator is also phase sensitive.

In a way a desired all-metal response that includes poor conductors, such as small gold, is in quadrature to the exciter, so a detector that favours quadrature detection would work better overall, and would simultaneously be less sensitive to ground which is in-phase.

The most obvious example of quadrature detection in everyday life is FM. It rejects the amplitude variation (I) and concentrates on extracting information from phase (Q) of the incoming signal. Ideally amplitude is completely obliterated (I->0) by means of amplitude limiters or some other mechanism. The simplest form of a FM demodulator is a slope detector that incredibly resembles what we have as an off-resonance metal detector. It may as well be possible to employ some more sophisticated FM demodulation technique to come by a better pinpointer and some.

I'll drop here occasionally with examples and solution candidates. Feel free to participate.

[WM6]

Welcome in the Off-resonance world Davor. In my eyes very promisable simple solution for us amateur MD builder (cause we cannot compete with "Big Brothers" in this field. I done some off-res experiment in past and was excited by ferro-non ferro distinguish capabilities. But depth was under my expectation. I hope with your help we can go step forward in this simple MD technology. In metal detector too, not only pinpointer.

And best regards (to all) from your beautiful Paklenica.

[Davor]

Nice, when looking across the sea, you'll notice two small islands and a small village on the other side of a channel at south-west called Ražanac. I spent my summers there as a child.

I guess you are referring to a Heatkit off-resonance solution? That one had a passive tank driven by a crystal oscillator. I think it is the way to go as off-resonance oscillators simply can't be manipulated that easy ... despite the appeal of easy construction. I think there is a way around that constraint.
You must have both amplitude and phase manipulation at hand (as Heatkit does) if you wish to have discrimination, but all-metal would work off-the-bat with a proper frequency/phase demodulation.

Anyway, whatever projects in off-resonance you may have stored, please post them here so that we may use this topic as a hub.

[Davor]

Here is a Heathkit schematic I mentioned earlier

[Overtheedge]

You all got my attention quick by starting an off-resonance thread.

I've wondered about the potential for a crystal oscillator w/emitter follower feeding the coil. Then feeding the output to crystal-filtered (single crystal or half-lattice) amplified RF voltmeter. I tend to think an inline variable attenuator (I have MC3340 on hand, -80dB to 1MHz) might be mandatory.

The other option is putting the coil in series with the crystal in the oscillator. Not only would you have amplitude effects, but frequency shift as well when feeding the output to a crystal-filtered RF voltmeter.

There is such a thing as a VXO regenerative receiver. see "KR1S QRP"

I already picked up some 200KHz ceramic resonators (Digikey)and have some surplus 43KHz and 300KHz crystals ordered (BG Micro).

Oh, the purpose is to check mine dumps for small gold still locked in the gangue. So no ground balancing is required. Why not a full size MD? I have to pack in everything several kilometers uphill and camp for several days above tree-line. I'm old (62) so weight and cubic is limited. Between food, survival necessities and the rock-busting tools needed, I have to trim grams off anyway I can.

Anyway, my off-resonance ideas might just be pipe-dreams. But you have my attention and I still have a lot to learn.
eric

[Gwil]

Don't think an emitter follower would be a good choice for a coil driver - it has a low impedance hence constant voltage output. Note that the Heathkit design drives the coil via 100kohms or so. OTOH the collector of a small transistor can have hundreds of kohms of impedance, making signal changes due to targets more apparent.

G

[Overtheedge]

The only reason I considered the emitter follower was to isolate the oscillator with a constant load. Premise being an off-resonance MD would require constant voltage to the coil between the oscillator and RF voltmeter. Eliminate as many variables as possible.

I don't recall if I have used a FET as a buffer, but perhaps I'm just stuck on the load isolation issue.

In RF theory, reducing the coil's value would reduce the reactance permitting closer matching for the emitter follower. The converse is I've run into problems trying to play that game.

This just ain't the same as typical RF stuff a ham radio homebrewer is use to.

Like I stated earlier, I've got a lot to learn.
eric

[Davor]

Surely there is appeal of a simple and lightweight solution. I initially pitched this technology for a pinpointer, but there is no reason why it shouldn't be used for a more grown-up solution. There are several problems never tackled before, and there are also some conceptual confusions about it, and that's the reason for my meddling with this.

Yes, I think off resonance can be a good choice for prospectors. More about it follows...

Main troubles with off resonance are oscillator noise (e.g. low sensitivity) and the fact that this is a single channel CW detector. Less so is he sensitivity to EMI which can be cured with proper coil arrangement.

By single channel I mean that you can't select a single quadrant for discrimination, but two quadrants simultaneously, and hence it operates as "all-metal" channel. The Heathkit solution shows it is possible to have some level of discrimination, but what is less obvious is that when discriminating there is a strong ground influence. It is because the "slope detector" used there is strongly influenced with "I" phase (AM) response.

Small gold response happens at almost pure "Q" channel (PM) response. That is a very good news for gold prospectors as eliminating "I" channel will also eliminate oscillator AM noise and ground response, and hence provide much more sensitivity.

Strictly speaking, a device that would have zero AM component would not be off-resonance at all. To avoid any AM->PM injection, a pure PM detector would have to operate dead-on resonance, at least the driving oscillator should. In such case even huge S/N can be achieved, as seen in FM receivers. This approach can be a promising one.

Separating AM and PM responses can be achieved simply or complicated, whatever one can stomach or pay for, but as Overtheedge points out - every gram of weight that can fall off is a good riddance. After all, seeking gold in crevices requires a compact device that indicates gold, not a hodgepodge clanker monster with cables and battery compartments jutting at odd places.

[TH Jim]

Hi folks,

I have also done some tinkering with Off-Resonance a while ago. It is something i found very interesting and enjoyable. I was also thinking using it in a probe, as the detection distance is not great, but it does have discrimination. I used a LM555 with variable frequency to push square waves through a tank, essentially using a tank as a filter and measuring the current that passes through it. I got some reasonable results, will post my simple schematic sometime if anyone is interested?

Would like to have another look at off-res in the future, but this time using a PIC.


Regards, Jim

[Davor]

Yes, please put your schematic here. At least I'd be very interested.

In my playing with the concept, I too realised that a driven tank is a way to go if you have any discrimination aspirations. Point is that PM to AM conversion functions well only in a driven tank. It means that if you observe only the amplitude, and by means of steering the driving frequency upper or lower than resonance, you'll notice that on one side amplitude change becomes greater than the other, and also target tau related. Hence discrimination.

There is another facet of this technology that may lend itself nicely to PIC implementation. Instead of amplitude one can observe phase. It is far easier to establish an accurate frequency reference than amplitude reference, hence a phase/frequency response may lead to a far better detector than one that observes amplitude. Among all frequency/phase detection techniques, by far the simplest, and seldom properly explained, is pulse-counting detection. It can be achieved directly by making a pulse integrating detector that converts pulse density to voltage, and such devices are often used in high-end FM receivers (so it actually works ) but in case of a PIC, it can be implemented by literally counting the pulses per 0.01 second or so and playing with running averages to compensate for movement. Such detector would perform well at lowish frequencies, would be very insensitive to ground, but would work well on small gold. It would consist of only a free running oscillator (with no off-resonance offset) and a PIC that senses the oscillator by means of a Schmitt trigger... and that's it. The cheapest detector of all times.

Now, why on Earth should it work?!?

As the (normal) ground response is aligned with the phase of the oscillator, hence influencing the amplitude of oscillation, but NOT the phase (at least not the normal grounds), hence it has some kind of ground balance built in. It can be further improved by properly detuning the oscillator, or just let it be for simplicity. All metals have some level of phase response (as in PM), and their response vectors can be dismantled to AM (discarded) and PM components, having a slight bias for poor conductors. And there you have it. A perfect all-metal probe with ground balance.

Now, why would anyone consider a detector that uses an oscillator dead-on resonance as off-resonance technology? Well, primarily because detuning the oscillator converts a PM response to AM, and such detector is in fact something called a "slope detector". These are the worst of all PM detectors around and not immune to any of the noises coming from both AM and PM worlds. Historically those were the first ever FM/PM detectors, and everything that happens in off-resonance metal detectors is directly related to these. So - why not - off-resonance is a way of recognising the tradition.

I have some problems converting these ideas to a single and down to Earth simulation, as Spice is not the best of friends for PM and pulses, but I'll put it here as soon as I figure out how to make it presentable.