SPI-EG4 !

[DOOLEY]

OR ,

poss , what i'm working on at the moment , imagine a see - saw ballance ,

voltage DIFFERENCES between TYPES of metals minus no target signal , can indicate type of metal regardless of size and density or distance ,

the voltage differences between two samples in the right place remains fairly
constant dependant on metal type ,

which can then be used to trigger a simple led setup to indicate metal type ,

1 , 2 or 3 tones not enough for discrimination , tried 10 different metal types so far
alu,steel,iron,copper,nickle,brass,stainless steel,gold,lead , all give quite distinctive voltage differences between 2 samples regardles of size or distance , well , i say regardles of distance , they have to register on both samples ,

but the idea don't work to well when 2 different metals together or if ground a little "iffy"

more experimenting to do yet ,

[Tec]

The 5c coin benchmark.

Hi Dooley can you post the circuit ?

What would be good is if the most recent posts appeared at the top, since my circuit has completely changed since the top post, with it's mux sampling.

I am interested to know typical detection depth of a 5c Australian coin for commercial PIs. ?

Anyway, today I took it to the beach for a test.Found a euro !.

It also detects 5c coins, grass, and sand very well,

But they can be tuned out, compensated for, tweaked, shielded. no problem.

Waiting for a new PCB to arrive.

[DOOLEY]

hello tec ,

don't want to post any circuit yet , it's a jumble of mess , some on paper , some in my head , some on pc , and nothing fully working or worth presenting to the mases yet.

your 5c i think is similar to a uk 5p silver coin , i dont have a commercialy made PI , just a home build based on the blind squirrel , but detection for similar coin is 6-7 inches using a 300mm dia mono coil , poss a bit big to be good for small objects.

you mention shielding , i wanted to shield , but i ran out of space inside my coil made with polly plumb pipe , but funny thing is , when using 20us sample width , hardly any change when nearing the ground , but when using 10us sample width its quite bad ,have to reset threshold all the time , wonder why that is ??
anyway , point is , with no shielding i seem to have "got away with it" when sample is wide.

[Tec]

5c piece 5p piece

Thanks,

I looked up 5p info:
it is 23% bigger than an aus 5c piece.
The composition is the same (copper 75% nickel 25%)


I know why the 10uSec sample width picks up the ground and the 20usec doesn't..
The wider sample averages out the signal (see picture).

Using 10uSec sample and increase the sample delay by 10uSec should also have the same affect...maybe





From wikipedia:

The 5p coin is minted from an alloy of 75% copper and 25% nickel. The 1968 version of the coin weighed 5.65 grams (0.18 troy oz) and had a diameter of 23.59 millimetres (0.929 in). On 27 June 1990 a smaller version weighing 3.25 grams (0.104 troy oz) and with a diameter of 18.00 millimetres (0.709 in) was introduced

Aus 5c:
Mass 2.83 g Diameter 19.41 mm Thickness 1.30 mm Edge Reeded Composition 75% Copper,
25% Ni

[WM6]

Hi Tec,

I am sure that we all follow with big interest your great PI uC work from start. Don't stop!

[DOOLEY]

HELLO TEC ,

wiki inacurate , our new smaller , post 1990 5p is 1.5mm thick , 17mm dia not 18mm , and weight is correct , compared to aus 5c i rekon there is not allot in it , but yes yours is slightly smaller.

sample width explination looks reasonable , but still some doubts.

had a bit of brain storming this evening , looking at scope pics of 10 metal responces , simple way of ID regarding anything with nickle in it + gold +stainless , slope goes down on first sample , up on second , thats easy , just use an op-amp , open gain as a comparitor on both samples , but copper has got me beat , looks the same as alu and steel and iron , and let's face it , no one would be happy passing over copper as allot of old coins were made of it , i certainly wouldn't be happy ,

so there we have it , looking for a simple answer , thats easy to build , and copper has me beat pic below.
ps:the middle pic is copper coin labled 2p , our old 2p coins were copper , new ones are steel.

[DOOLEY]

so lets all play "SPOT THE DIFFERENCE" can you guess what it is yet
NO ! neither can i
never thought copper would be the tricky metal .

[Monolith]

The mystery of the "Hand signal"

When we get into higher sensitivity, we meet with the "Mystery of the hand signal"
The detector detects the hand.
Shielding the coil, this phenomena is attenuated or eliminated. In fact, I use it to test the shielding. If the detector does not detect the hand anymore, the shield is good enough.

But what exactly is being detected through the hand?

I never bothered to look at the causes, maybe it is time to investigate that.

When I touch the probe of the oscilloscope with my hand, I get over 100V, at 60 cycles. The un-shielded coil picks up a wide spectrum of noise, up to 300kHz, as the FFT showed.

The well shielded coil does not pick up none of that.

Does that mean that mean that the shield is material in attenuating the 60Hz environmental noise? How does it do that?

I understand that the shield attenuates RF noise, but 60Hz?

What exactly is the 60Hz (mains) that comes out of my hand?

Current?

A varying magnetic field?

How is it coupled to my body?

How does it radiate 8 inches from my hand to the detector?

How is it coupled or transmitted to the coil? Or to the coil cable, or the un-shielded circuit board.

This is still a mystery to me.

Can we solve this mystery?

Monolith

[Monolith]

so lets all play "SPOT THE DIFFERENCE" can you guess what it is yet
NO ! neither can i
never thought copper would be the tricky metal .

Try alu foil as compared with thick alu.

And then try a crown cork vertical and the same crown cork flat.

This may give us a hint.

Monolith

[WM6]

A varying magnetic field?

How is it coupled to my body?

How does it radiate 8 inches from my hand to the detector?

How is it coupled or transmitted to the coil? Or to the coil cable, or the un-shielded circuit board.

This is still a mystery to me.

Can we solve this mystery?

Monolith

Hi Monolith

This is not mystery. Your body act as an wideband antenna. Our bodies are out of 70% water with conductive mineral (Ferro inclusive) soluble inside. Considering mass of about 70kg we get an impressive antenna part.

[DOOLEY]

hello monolith ,
i have also been using my hand to simulate ground and minimize the effect

yes , should have tried foil , tomorow is another day ,will try ,should be the same ??
i hope , as foil in it's many forms is the pain that gets us all.

whats a "crown cork" ??

any ides's on easy solution on the copper problem with a PI ??

all the best monolith .

[DOOLEY]

hello wm6 ,

same thing happens to some degree in the field with my hand , and no where near ac , electric wires as far as i am aware.

best guess , hand and ground is absorbing / reflecting some signal / magnetic field , and so lowering /raising the target signal resulting in change of target voltage .

only a guess

[mikebg]

According Monolith, "Shielding the coil, this phenomena is attenuated or eliminated. In fact, I use it to test the shielding. If the detector does not detect the hand anymore, the shield is good enough."

According Charles Garrett, the shielding eliminates the "Green grass signal".

I think, the shielding can not suppress conductive signals because they are generated by eddy currents in substances shown below.

[DOOLEY]

hello mikebg ,

so , a correctly shielded coil would minimize "ground signal" to the point of being small enough not to be a problem ??
and so eliminate the need for further circuit to deal with the problem ??

[Monolith]

hello monolith ,
i have also been using my hand to simulate ground and minimize the effect

yes , should have tried foil , tomorow is another day ,will try ,should be the same ??
i hope , as foil in it's many forms is the pain that gets us all.

whats a "crown cork" ??

any ides's on easy solution on the copper problem with a PI ??

all the best monolith .

Hi Dooley,

a crown cork is a beer bottle top. Mikebg posted a nice picture of it somewhere.

To differentiate metals we need to start with targets of the same size and shape. Then the metals can be differentiated by their conductivity and TC. Ideally we use spheres for that.

The conductivity affects the the skin effect.

The skin effect is closely tied to the frequency of eddy current excitement.

To fully understand how the eddy currents are excited and how they propagate through the target and how their decay evolves, we move to targets of different shape.

Monolith