PI metal detector with energy recuperation

[Midas]

2Midas

Yes , I have the same thought now ... but I really did it for some purpose . This design did have a three main goals - to test this "current-reverse and recuperation" concept , to test some specific schematic solutions and to test a ferrite as a search coil core . What does we know about ferrite ? It has 2 main drawbacks - magnetic instability and nonlinearity And these drawbacks makes almost impossible to use it in high-class metal detector . On the other hand , ferrite search coil looks quite attractive in some specific conditions , such as hollows , caves , fractures , etc . And what is interesting , it seems that this circuit is the one and only that can use a ferrite quite correctly and push it to the limit .

What I mean by this - it's instability and nonlinearity cannot make a harm here . You see , when the current is changing in the coil ( charge and discharge intervals , A-B and D-E on timing diagram ) - we don't measure anything , and when we measure the incoming signal ( measuring interval C-D ) - the current doesn't change When we receive , the current is constant , we are sitting on the one point of the magnetization curve , so if the permeability does change , it's not a problem at all . Even if the permeability of our ferrite drops to 1/2 of initial value - it can't disturb the demodulation process of the circuit . All we need from our ferrite core is to "suck" magnetic force lines into the coil , and it's enough ...

So I cannot stop myself to try it anyhow and look how can it work . What about the concept - it works OK , so the schematics does , but about ferrite I am hesitating now . As I think - this performance of course isn't the best because of mediocre quality of the ferrite that I used ( and the core isn't monolithic , it's not good too ) , but what can I achieve when I use the best ferrite - it can show the future experiments , of course .

So the best idea now is to assemble the same device with the same schematics , but optimized to use a conventional search coil ( simple round coil with air core ) - it needs some minor changes in the power chain circuit , and direct compare it's performance with the existing one . And if I find the good long ferrite core with high permeability and good Litz wire , I will be able to replace my search coil in this device and compare them again in order to make a final conclusion .

OK cool! I look forward to seeing your results.

Midas

[Anker]

The main parametres PI of the device - its sensitivity and ability to ignore a ground mineralization. Profitability - not the most important. We will admit, you have made the supereconomic device and the released energy of the battery have started up for increase in capacity of an impulse. Also what? At you sensitivity will essentially increase? Alas, it is not enough - for example, for increase in sensitivity of the device in 2 times you should increase or capacity of the generator in 64 times, or receiver strengthening in 64 times. It turns out, the sense in economic PI is not present...
In respect of adaptability to manufacture and reliability such device also loses - to make the simple scheme with the usual battery, than very difficult with the tiny battery better.
But as it is known, some people do not search for easy ways

[deemon]

The main parametres PI of the device - its sensitivity and ability to ignore a ground mineralization. Profitability - not the most important. We will admit, you have made the supereconomic device and the released energy of the battery have started up for increase in capacity of an impulse. Also what? At you sensitivity will essentially increase? Alas, it is not enough - for example, for increase in sensitivity of the device in 2 times you should increase or capacity of the generator in 64 times, or receiver strengthening in 64 times. It turns out, the sense in economic PI is not present...
In respect of adaptability to manufacture and reliability such device also loses - to make the simple scheme with the usual battery, than very difficult with the tiny battery better.
But as it is known, some people do not search for easy ways

As I told before , I tested several things at the same time , and one of them was a ferrite search coil . So for better sensitivity maybe better to use simple air coil , and I will test it at the next experiment ( I already prepared all the parts and just winding the coil now ) . But the other features works and works OK even in this prototype . I mean , besides power economy , very low sensitivity to the soil ( without any special "ground subtraction") and improved metal discrimination due to constant current in the coil while the incoming signal being received . And don't forget that reversing the current instead of stopping it already doubles the incoming signal from metals , with the same current in the coil . I don't think that it's a bad solution

And anyhow , if anybody find a new idea , it's of course better to spend the time testing it than to assemble circuit based on well-known old concept , rutted back and forth many times . Maybe this new concept will be used , maybe not .... but I made it quite reliable and solved all the problems here . This prototype was tested in the field and worked OK . By the way , I just have an idea of a "new generation" of this concept , even more effective than present one . The schematic idea seems very beautiful , it will be something special And about "production complexity" it's a funny argument , as I think - just open any computer or plasma TV , you'll find much more parts inside

[maikl]

A similar military mine detector was used in the army of Serbia (probably by Russian licenses), replaced by more modern detectors...

[deemon]

A similar military mine detector was used in the army of Serbia (probably by Russian licenses), replaced by more modern detectors...

By the way , I'd seen a quite similar military device here in Russia , but I remember that it didn't use a PI technique , it was a kind of inductive balance .

[Anker]

Complexity of the scheme should be justified new possibilities. For example, in devices JP Minelab rather difficult scheme - but these devices work perfectly well!
That, we wait for results of the test with the air coil.

[maikl]

By the way , I'd seen a quite similar military device here in Russia , but I remember that it didn't use a PI technique , it was a kind of inductive balance .

You are right deemon, is not PI... it's the induction balance detector (model from the sixties) with two RX coils... http://www.saper.etel.ru/texnica/imp.html

[Tepco]

Schematics of this contraption|:

[Tepco]

As you can see , the first idea was not to stop the coil current , but to reverse it , using the LC resonant circuit . The first step is to charge the coil ( phase A-B on the time diagram ) , then we have a flyback ( phase B-C ) and current reverse ( just like TV deflection works ) , and after this - the coil voltage drops to the ground and the coil current begins to recuperate . But the metal response begins just after the flyback , and at this moment we have a rapidly decreasing current in the coil , so it will be a great problem to select our signal .

So I deciced to use a trick - just after the flyback I open the mosfet switch , and let the current circulate in the shorter chain ( coil , mosfet , diode ) . Now I'm wait for about 100 uS - phase C-D on the right diagram ( and receive the metal response ) , and after it - shut down the mosfet switch , allowing the current to flow back to the power supply . In another words - we use the current , and then return it

In order to get the signal from the coil I use the current transformer TR2 ( on the power chain circuit ) . Its output current goes both to tilt compensation circuit and to processing board . Compensation must be used due to losses in the coil chain , so without the compensation we'll have a current decay that will be mixing with the received signal .

This post get more or less unnoticed on this forum, very interesting proposal. I just watching schematics and trying to figure out how it works. As far as I can understand (correct me if i'm wrong) this machine is substantially different from standard PI in at least two aspects:


First, and most important one, almost “textbook” definition of PI is that it takes sample when all initial pulse energy is either dissipated or recovered by some mechanism, so no stored energy exists in system, except one stored in target. This one is completely different, taking sample during period when energy is still present in coil circuit (and recovered later). At first glance, looks like eternal source of trouble, but interesting idea, worth investigating. Achieved using complex substraction circuit. Second, this design takes coil current as a signal, not voltage developed across coil, very interesting solution. (I tried something along these lines before, but unsuccessfully, more or less).


Interesting question is, did you try this concept with standard, air-core coil, not ferrite loaded, inductance is probably somewhere close? And some, potentially useful comments. Large single ferrite core is available for impeder coils for electrical welding, relatively cheap, say this, from www. kaschke.de, something like this: http://www.kaschke.de/en/products/fe...impeder-cores/ , so no more gluing etc. (This is locally available from comet electronics in my region).Another aspect may be relatively long flyback, 8uS as noted, all effort must be made to minimize this time, 8uS is long, this will certainly “undo” effect of main pulse to very small, low TC targets. What detection range you can achieve with prototype, especially for small targets?


We should call this “ Arzamas 16” detector, completely out of the box.

[deemon]

Schematics of this contraption|:

Ha-ha , I remember this schematics ..... many years ago , in the middle 80-s , I repaired it . I recall now its push-pull generator , balance circuit , transformer coupled stages on ancient transistors .... human memory is a very interesting thing

[deemon]

Interesting question is, did you try this concept with standard, air-core coil, not ferrite loaded, inductance is probably somewhere close? Another aspect may be relatively long flyback, 8uS as noted, all effort must be made to minimize this time, 8uS is long, this will certainly “undo” effect of main pulse to very small, low TC targets. What detection range you can achieve with prototype, especially for small targets?

Of course , the principle of operation doesn't depend on ferrite in the coil - it can work with an air coil as well . But for best operation we need here a coil with high Q-factor ( less DC resistance and eddy current losses ) , so the best idea is to wind the coil of the good and thick multi-stranded Litz wire ( as I did on ferrite coil ) . And this is why I still didn't try it with an air coil - my store of proper Litz wire is finished and I cannot get it more . Now I decided to use a simple copper wire , to test it anyhow ... but of course , it's not an optimal solution .

What about a flyback duration - in the first prototype it was about 5 uS , and then I increased it , in order to reduce the sensitivity to a thin metal garbage . I did it because I wanted to get the universal device for a typical "medium" objects ..... but if we need a device specially for a very little objects - we can accommodate it , reducing the flyback duration , and a measuring interval also . And vice versa , for very big and deep objects the optimum is to set a big coil with bigger inductance , longer flyback , etc .

[Prospector_Al]

Hi all !

2 years ago I decided to make my own PI metal detector . For the first time I made some experiments with well known PI technology , but one day a very beautiful idea came to me - I found a way how to store and recuperate the energy that we usually waste in the resistor connected in parallel to the sensor coil . I tried this idea and found that it really works . So we can either decrease power consumption of the device or radically increase the pulse energy with the same current from the battery . And after some research I made the prototype , it shows quite a little energy consumption and some other interesting features . Now I decided to publish the circuit of this device - maybe you find it interesting ... But how can I attach the pictures ?

Best regards
Dmitry

Hi Dmitry,

It's a good idea--it works. It's been patented by Candy:
http://appft.uspto.gov/netacgi/nph-P...andy+AND+Bruce)

It has a drawback, though. The power recovery process slows the return of the coil current to zero. The consequence is that targets with fast Tcs are not energized as well as with a system that doesn't have the power saving feature.

All the best,

Allan

[Altra]

Hi Dmitry,

It's a good idea--it works. It's been patented by Candy:
http://appft.uspto.gov/netacgi/nph-P...andy+AND+Bruce)

All the best,

Allan

And Dave Johnson did it a few years before BC

[Tepco]

This is not what Arzamas16 does with energy recovery:

[Tepco]

And Dave Johnson did it a few years before BC

In "one of it's kind" 1280 Impulse: