Metal 'discriminator?'

Bottom Line Up Front: NO ELECTRONICS EXPERIENCE AT ALL! I'm too old/stubborn to learn it at this point, and I have exhausted all other resources. So... please help me b/c you're my only hope!

Background info:
- Some pre-1965 U.S. coins contain 90% silver.
- U.S. half-dollars minted between 1965-1970 contain 40% silver.

My goal:
- I would like to be able to determine whether a coin roll (literally, a 'stack' of coins wrapped in paper) contains silver (either 90% or 40%) WITHOUT opening the roll.

My 'vision':
- Silver, as you probably already know, is the most electrically (and interestingly, thermally) conductive metal on earth, so I'm hoping that by passing the coin roll (containing MOSTLY copper-clad coins) through some sort of a simple coil, I can detect the difference in [not sure - fill in the blank?] as the roll is moved sloooooowly through the coil and the conductivity (or whatever) changes as the silver coin passes by the coil. I don't need to accurately 'measure' anything - I don't need numbers - I just need to be able to detect the difference in copper-clad and silver-containing coins. Something like a needle jump or an LED when a certain threshold is passed, for example. The VAST majority of coins will be copper-clad: silver-containing coins will be the very rare exception.

** If I could discriminate between 40% and 90% silver coins (higher needle jump or different color LED for example), that would be super-cool too... but if that's too hard, it's not a necessity.

Am I off base? If not - what shouuld I try to detect/measure? Inductance? How do I [easily] do this?

Thanks in advance for any help you can provide this electronically-illiterate soul!

Heat the rolls with a induction coil (eddy currents). Then scan the roll length wise with an infra red detector. The temperatures of the silver and copper coins will be different. Count the "copper" vs "silver" temperature occurrences

Thanks! That's a different approach than I had considered. I was hoping there was something quicker/simpler than heating and then measuring - but it sounds like it will work. Is it difficult to build an induction coil? What sort of timeframe do you suppose it will take to adequately heat a roll (20 half dollars, for example)?

BFO Metal Detector will discriminate metals.
You can search for "BFO Metal Detector circuit" by google and start to test it.

klystron said:

Heat the rolls with a induction coil (eddy currents). Then scan the roll length wise with an infra red detector. The temperatures of the silver and copper coins will be different. Count the "copper" vs "silver" temperature occurrences

Click to expand...

Regarding an inductive heated roll, it might be difficult to evaluate
the temperature difference through the wrapping paper.
Also, as the coins are in intimate contact, temperature differences
will level out easily due to good thermal conduction.

If not put the focus primary on heating, but just sensing, one
might investigate the electrical power variation when moving the roll through
the electrically stimulated coil.

A silver coin (of good electrical conductivity) among ordinary
copper plated coins will result in higher damping of the
electrical oscillations, given lower signal amplitude.

pico said:

BFO Metal Detector will discriminate metals.
You can search for "BFO Metal Detector circuit" by google and start to test it.

Click to expand...

A metal detector or cable tester works in quite the same way.
In presence of metal, the frequency of a LC resonant circuit
is changed from its nominal value and the frequncy shift
can be heared.

At least, an individual silver coin should be distingushable from a copper plated one.
Maybe you should try out first on samples.
It would be hard anyway to do it on a whole stack (roll).

Eddy_C said:

one might investigate the electrical power variation when moving the roll through the electrically stimulated coil.

A silver coin (of good electrical conductivity) among ordinary copper plated coins will result in higher damping of the electrical oscillations, given lower signal amplitude.

Click to expand...

This is more inline with what I was thinking. The trouble is - I wasn't joking when I said I was electronically clueless. :-| I really don't know what I would be measuring/sensing! I know/remember that passing a conducter through a coil with a current going through it will affect something... I just don't know what. If the composition of the conducter (the conductivity) changes as it passes through the coil - I envision being able to sense/read a corresponding change in whatever it is I'm measuring - like a higher meter reading as the [hidden] silver coin goes through the coil.

Maybe if i paint a picture of what i envision...

I picture myself reading a meter (of some sort!) while slowly passing a coin roll through a narrow-ish coil with current going through it. As the predominantly copper coins which make up my conductive rod pass by the 'sensor' (coil), the meter does not fluctuate much. If a silver coin (with its higher conductivity) passes by the narrow coil however, I'd see a definite spike on the meter.

That's what I'm hoping for. I just dont know if it's possible - as I don't even know what I should be measuring/sensing.

Thanks again - I'm really hoping this project is possible!

Eddy_C said:

one might investigate the electrical power variation when moving the roll through the electrically stimulated coil.

A silver coin (of good electrical conductivity) among ordinary copper plated coins will result in higher damping of the electrical oscillations, given lower signal amplitude.

Click to expand...

Sorry for the bump, but - what is an easy way to set something like this up? This sounds like just the ticket, but I'm not sure how to measure 'electrical power variations...signal amplitude." Are we talking about an oscilliscope? If so, I don't own one, nor know how to operate one... but if that's the only way... 8-O I guess I could buy one and learn. I was hoping there was something simpler to set up... like some sort of simple meter?

Thanks again.

Back to basics :
The required characteristic must be detectable and different for copper and silver
The required characteristic must have the spatial location resolution of 3 mm (coin thickness)
The relative magnetic permeability of copper and silver are 0.99999 and 0,99998 respectively
Metal detectors use the relative magnetic permeability to detect ; it will not resolve the small difference in permeability
Metal detectors coils cannot give 3mm spatial location resolution.
Metal detectors will not work

The thermal conductivity of copper and silver are 400 and 429 respectively
Heating a roll of coins in an oven (including an induction oven ) for a short while and then viewing the rolls with a suitable FLIR camera will show the silver and copper coins

The trick is to use a cheaper type of "FLIR" camera ; but perhaps the real of the self FLIR camera will be cheaper than the development cost.

klystron said:

The thermal conductivity of copper and silver are 400 and 429 respectively
Heating a roll of coins in an oven (including an induction oven ) for a short while and then viewing the rolls with a suitable FLIR camera will show the silver and copper coins

The trick is to use a cheaper type of "FLIR" camera ; but perhaps the real of the self FLIR camera will be cheaper than the development cost.

Click to expand...

You've sold me on the 'thermal' vs. 'electronic' (for lack of a better term) approach.

I'll shoot for heating inductively (I'll have to Google how to do that), and then either slowly scan lengthwise with an IR thermometer or pay someone to build me a cheap-ish thermal-imaging device. This site (link) has a really neat approach: put a spot-IR-thermometer on motors and have it 'scan' an area! Since I only need to measure differences along the length of the roll, I could eliminate one axis and just have it scan up a predetermined length.

Anyone with Arduino experience out there want to make a few bucks? :lol: Hit the link and then PM me if interested in building something like this (with only 1 axis vs. 2).

Thanks again.

In coin selector box, it discrinate coin by using physical properties. For example
- coin size
- coin weigh
- coin electromagnetic property

https://en.wikipedia.org/wiki/Silver
https://en.wikipedia.org/wiki/Copper

Do we really compare pure silver with pure copper ?
What is the base material of the copper clad coins ?
Can you tell something about the typical material composition ?

If it is steel or nickel then it would be magnetic and thus a metal discrimination by different magnetization could be done.
If I assume that the silver containing alloys are more or less non-magnetic, then one might think of using the different magnetic properties.

For example, they can be simply investigated using a tiny permanent magnet (e.g. rod magnet with pole size of about thickness of a coin).

When moving one pole of the magnet alongside the stack, there should be noticable effect caused by ordinary copper
plated coins (assumed to be magnetic) and the silver containing coin - the attracting magnetic force should decrease slightly
when the pole is located at the silver containing coin.
(It will not vanish totally because of the influence of the adjacent magnetized copper clad steel coins.)

It's worth to think about some simpler means than inductive heating and IR image scan.

pico said:

In coin selector box, it discrinate coin by using physical properties. For example
- coin size
- coin weigh
- coin electromagnetic property

https://en.wikipedia.org/wiki/Silver
https://en.wikipedia.org/wiki/Copper

Click to expand...

That would work - but I need to do the whole stack/roll withOUT unwrapping it.

---------- Post added at 17:32 ---------- Previous post was at 17:30 ----------

Eddy_C - it's not pure silver or pure copper. Wikipedia has some info on the composition here: https://en.wikipedia.org/wiki/Kennedy_half_dollar

None of them are magnetic.

Could you not just weigh the entire roll of coins? If the weight equals (weight of non-silver coin)*(number of coins) then you know there are no silver coins in the roll.
Or is the partial composition of the silver coins some metal that causes the coin to weigh the same?
(Sorry if that is an obvious question to some, but I'm no expert in US coins).

From the link you gave
https://en.wikipedia.org/wiki/Kennedy_half_dollar

Mass 11.34 grams
(as copper-nickel clad)
40% silver clad: 11.50 grams
90% silver: 12.50 grams g
Diameter 30.6 mm
Thickness 2.15 mm

So, the 90% silver coil has weigh more than the 40% silver coin 1 gram per coin (12.5-11.5) gram. If a roll of coin has 10 coin, it wold has a weigh different of 10 gram per roll. It's a very basic math.

sky_123 said:

Could you not just weigh the entire roll of coins? If the weight equals (weight of non-silver coin)*(number of coins) then you know there are no silver coins in the roll.
Or is the partial composition of the silver coins some metal that causes the coin to weigh the same?
(Sorry if that is an obvious question to some, but I'm no expert in US coins).

Click to expand...

That's a great question - and that was my first inclination as well. After weighing several hundred rolls (and then opening them), I discovered that - with a 0.001g precision scale - it was relatively easy to detect a roll with at least one 90% coin in it, but the difference in weight between a copper-clad and a 40% silver coin was WELL within one standard deviation of the average roll weight. I actually measured some rolls that contained TWO 40% coins that weighed LESS than some all-copper-clad rolls. 8-O Those would have slipped right past if I were relying on weight alone.

---------- Post added at 21:38 ---------- Previous post was at 21:34 ----------

pico said:

From the link you gave
https://en.wikipedia.org/wiki/Kennedy_half_dollar

Mass 11.34 grams
(as copper-nickel clad)
40% silver clad: 11.50 grams
90% silver: 12.50 grams g
Diameter 30.6 mm
Thickness 2.15 mm

So, the 90% silver coil has weigh more than the 40% silver coin 1 gram per coin (12.5-11.5) gram. If a roll of coin has 10 coin, it wold has a weigh different of 10 gram per roll. It's a very basic math.

Click to expand...

Basic indeed. See my previous post. I wish it were that easy (I tried). If I were willing to let the 40% coins go (I'm not), that method would (and indeed - does) work.

And - not that it changes anything - but there are 20 half-dollars in a roll ($10 total face value). Not 10.

Do you mean that there are mixed coin type within one roll?
If so, you can detect only it is a mixed coin roll or it is a pure same type of coin in the roll with a very precision scale as you mentioned above.

Annnndd.... right back where I started. Let me try again:

You can't see what's in the rolls (except for the coins on the ends). There could be anything in there: silver, copper, John Gotti, anything... Almost everything (like 99.9%) however, will be Copper-clad - coins with the same properties. I'm looking for a non-destructive way to 'see' what's inside without opening the roll.

Side note: 6 months ago, I thought the precision scale route was on target. Until I realized that there's NO chance of NOT missing a few 40% halves using only weight as a variable. A roll with a 90% in it definitely stands out (heavy). The 40%s are the real problem.

To understand and mitigate development risks I will do the following.
Make 3 group of three coin rolls with a known silver / copper combination.(use the same paper as the actual coin-rolls)
Make a "nozzle for a heatgun with an output port measuring 2 x 100 mm
Invite a Fluke sales-person for a demostration of his FLIR camera

Heat each roll in coin-roll group A for 5 seconds and take FLIR "photos"
Heat each roll in coin-roll group B for 10 seconds and take FLIR "photos"
Heat each roll in coin-roll group C for 20 seconds and take FLIR "photos"

For very little cost you will have a much better understanding of the "performance" of heat detection