How good can discrimination be?

[andyt]

Hello All

I'm new to this, but from a tech background...so trying to understand the practicalities of the technology.

Ive been playing around with my new mxt pro. I notice that discrimination readings on the VDI scale are hugely reliant on the orientation of a ring or coin.

For example waving a British 10p over the coil parallel will get a good high positive reading, while doing the same perpendicular (on end) greats a -20 or so which is well in the "ignore" range. Ive seen many references to "dig positive".

Is this just a limitation caused by the fact that the phase response of say a coin changes dramatically depending on orientation? Or my detector!?

If phase does just change dramatically depending on orientation, it seems to me that anything above a very low level of discrimination would be worse than not having it...as you would walk by any coins etc on end!

Interesting to hear your thoughts!

Thanks
Andy

[Davor]

Discrimination happens by means of eddy currents in a target, and they depend somewhat on the target geometry, and much more on the target thickness if they are very thin. You can't do much about it, that's simply a limitation of this technology.

You can imagine some target as a short circuited coil. It has some inductance that depends upon the geometry, and some resistance that is due to the material specific resistivity. In ideal case of a smooth surface being perfectly aligned with your detector it will act as a parallel LR "filter" with tau=L/R, and you'll find tau for copper to be around 150us. You may extrapolate this picture for more demanding geometries and the resulting faster taus.

[kt315]

Is this just a limitation caused by the fact that the phase response of say a coin changes dramatically depending on orientation?

not. in reality the coin lies in soil and surrounded by gallo. this is some area, sphera, of conductor soil. this equalizes the response in both cases because the responce goes from the gallo, not from the coin.

[andyt]

Thanks

Slightly confused...Davor says "It has some inductance that depends upon the geometry",

Kt315 you are saying that the "in air" test is not correct, and that if the coin is buried in soil it will equalise and you will get more similar phase shift between the two orientations (edge vs parallel) ?

The phase surely must change on orientation to some degree as I have read posts that a coin at 45 degrees to the surface will register high say left to right and low right to left, this certainly happens on air test. On multi tone I can get a consistent tone different depending on the direction of sweep from left to right or right to left.

Not sure what gallo is?! That one is missing from my education...I will have to google (presume its a generic term for the soil or sand etc).

I designed electronics products...but not metal detectors.

Thanks again for your help
Andy

[mikebg]

Good morning Andy,
If you are professional designer of electronics, you are familiar with terms "impulse response", "step response"
and "frequency response" of a system. I'm not professional; my hobby is to design transmitters, receivers and
antennas for QRP amateur radio using frequency domain. For me is dificult to use timeconstants and responses in
time domain to design a metal detector. In addition, experiments showed that each response is formed by several
timeconstants inversely proportional to squares of odd natural numbers. That forms several cutoff frequencies in ratio
1:9:25:49 etc. Only the shorted turn responses with single timeconstant.

To design metal detectors, I need information for frequency responses of different soils, of salt water and of different
targets different oriented. The usable information is in enough wide frequency spectrum, for example 1kHz ...
100kHz. That means my test equipment for frequency responses of targets and soils operates as wide band metal detector.

XLT is a "narrow band" metal detector. The transmitter excites environment with sine induction 14kHz. This is only a
point in frequency spectrum. I guess the designers of your machine have the above mentioned spectral information, but they decided
that at 14kHz there is significant phase difference between responses of different targets.

However timeconstants of conductive soil depend on diameter of TX coil. Using 12'' coil at 14kHz generates too
strong GND signal, which limits sensitivity. In addition, the conductive soil shifts target phase which distorts
discrimination. The attached simple block diagram illustrates why discrimination of deep targets is distorted - the energy passes twice depth in ground.

[andyt]

Thanks Mike

Often in radio the Radio Amateurs have more experience that the pros at practically what happens. The pros spend too much time in the lab with calcs and simulations! Pros do have the edge on having test equipment worth more than a house though.

Ive been reading up. The MXT Pro apparently uses 14k ish as a compromise in finding things like coins, depth and also being able to see small gold in prospecting mode. It comes with a 12" concentric (spider) coil. Depth doesnt seem to be a problem in my tests, though as you might expect close iron junk throw the large coil off.

I also read that the resistance of the ground can be ignored but the reactance(inductance, phase) is high. But this is also fairly constant and changes slowly. Hence of course the need to sweep and keep the wanted signal from an object changing more quickly than everything else.

The answer to my question seems to be that orientation does play a part in the response, multi-frequency (or using lower freq) would help but not solve it completely as objects aspect to the field changes its response as well as the material composition.

Andy

[Davor]

Correct.

Your best approach is going from ideal to realistic and adding effects as you go. In every effect family you'll find some extremes that spoil your reading, but knowing the ideal will help you understand the annoyance you are dealing with. Ideal is a clean coin laying flat on a surface. Annoyances? Odd geometry, thin sheets or foils, salty soils, hot rocks, and a few more. However, some of the annoyances are a product of misalignment of your rig, and then you start imagining hot rocks and stuff where there aren't any. In reality you'll seldom stumble upon a real hot rock unless you are in Australia.

Happy hunting

[Qiaozhi]

Thanks Mike

Often in radio the Radio Amateurs have more experience that the pros at practically what happens. The pros spend too much time in the lab with calcs and simulations! Pros do have the edge on having test equipment worth more than a house though.

Ive been reading up. The MXT Pro apparently uses 14k ish as a compromise in finding things like coins, depth and also being able to see small gold in prospecting mode. It comes with a 12" concentric (spider) coil. Depth doesnt seem to be a problem in my tests, though as you might expect close iron junk throw the large coil off.

I also read that the resistance of the ground can be ignored but the reactance(inductance, phase) is high. But this is also fairly constant and changes slowly. Hence of course the need to sweep and keep the wanted signal from an object changing more quickly than everything else.

The answer to my question seems to be that orientation does play a part in the response, multi-frequency (or using lower freq) would help but not solve it completely as objects aspect to the field changes its response as well as the material composition.

Andy

The target response depends on the conductivity, thickness, orientation and any other surrounding targets (including the ground). Discrimination is more like intelligent guesswork. For UK sites it is advisable to do as you suggested, and use as little discrimination as possible (usually just iron rejection).

[satdaveuk]

Yes yes and yes
I have quite a few detectors, some hand built some made by manufactures, and all of them are the same, if your working in UK soil back off the discrimination to very low and dig everything that you can muster a two way hit with.
I tend to detect on all metal when ground permits then once a target is detected toggel into disc mode at lowest settings, if you then get a oneway hit, move around target while sweeping lisening for that two way hit, if you get that then dig.
If your on land without hotrock problems then either turn it off or minium settings because 95+ can often run through as well as other numbers and the last thing you want is a nulling because you will miss whats around it.
Ive often circuled a target 6-7 times before ive moved on because with them deep targets your looking for the quirky chirpy sounds that vary from detector to detector that tells you whats down there.
Your display is ok as a reference far as metal ID but thats where it ends, the only exception to the rule is when using detectors like the Whites DFX which has a signagraph which speaks a million words, very much like looking at a scope when seeing different phases for different metals and also seeing the ground make up, but with the rest like TGSL ML Quattro, Fisher 1265x Whites 1V and so on its all about sounds and knowing how your detector ticks which will come in time as you get more experienced with it.

[golfnut]

In line with your question.

I would get info from an Air test and compare to it 6" under your soil at flat and inclined and see what numbers you get.


I think air-head tests are a time wasters dream activity.

A bit like saying my car doesnt go real well in a lake!

S

[andyt]

I guess air test takes 10mins for a set of coins (or poked in the dirt near the surface as my test was) vs digging a whole several times! So may not be representative of actual circumstances but a bit less time intensive

But I did try burying a coin about 4 inchs and the results were the same. Coin diagonal, one tone to the left, another to the right!

I'm not too experienced with running up and down fields, so I guess after a while you guys train your ears and start to read the signs.

For me its a lot of digging I expect

[Tinkerer]

I guess air test takes 10mins for a set of coins (or poked in the dirt near the surface as my test was) vs digging a whole several times! So may not be representative of actual circumstances but a bit less time intensive

But I did try burying a coin about 4 inchs and the results were the same. Coin diagonal, one tone to the left, another to the right!

I'm not too experienced with running up and down fields, so I guess after a while you guys train your ears and start to read the signs.

For me its a lot of digging I expect

Si, in the end it always comes down to the same:


I'm not too experienced with running up and down fields, so I guess after a while you guys train your ears and start to read the signs.

It is the operator's interpretation of the machine's signal interface.

It depends on the ear and how good the audio output of the detector is.

Morale of the story: Design the best audio.

Tinkerer

[satdaveuk]

Good point Tinker better the audio more chance of hearing different tell tales of the object under the surface, but also lets not deceive are selfs once these metal objects are in the real world in real soil planted or lost many years ago some 20-50-100-1000s years is a totaly different ball game to doing air tests or digging a 6" hole in the earth, or doing spice simulations through a computer.
There is know exact formula because of all the varibals involved, not just from country to country but from one meters of land to the next
I will give you a example: If you want to go for serious detecting in say plough field here in England looking for roman or saxon gold you need to be well into the - numbers and upwards, your basically looking for a alloy thats not just buried with iron particals.shreads of farm tools use for thousands of years plus other alloys which include remains of drink cans remains of farmhand shirt buttons which have turned are mineral into a mine field of different minerals which is in abundance at most farmlands here.
So put all that together then try discrimination?
Comment up the page was not experienced so will proberly be digging many holes''
Well I can tell you now ive got 40years under my belt and still dig many holes and end up with crap.
If anybody for one miniute thinks with the science/formats we are using at the moment will give them accurate discrimination in the real world for anything over 6" in typical UK soil should take up fishing.
Any machine todate that ive had experience with is the same, only way is learn the machine, then you deside wether to cross that fine devideing line, do you dig or not, 9 times out of 10 its that close with the deep stuff.
Me personally if im not knackered I dig.