To determine if a gold coin is genuine

[woodchips]

Something a little different.

One of the saving/investment opportunities at the moment is gold coins. These come in many varieties but assume it is the standard 1 troy ounce 4 carat coin.

In case didn't know but tungsten has almost an identical specific gravity to gold, 19.32 to 19.25. So identifying whether you have a gold plated tungsten bar, or gold, is near impossible to tell.

But, their resistivities are very different. I wondered about calculating the side to side, across the diameter, resistance of a fake and a real 1oz coin. I couldn't work out how to do this, my maths struggles. Any suggestions?

Thing is, the fake could be a tungsten disc with gold plate. This is easy to spot, you can't bend it in your hands. But what is the tungsten was a fine powder such as being sintered and then mixed with the gold? It will be as malleable but the resistance would still be far too high.

Measuring an old silver, 95% I think, UK florin came up with just a few 10's of micro-ohms side to side.

[wraper]

You could use inductive sensing. Tungsten is paramagnetic but gold is diamagnetic.

[nowlan]

http://www.amazon.com/gp/product/B00DKW4IYO

Your local dealer might let you use their XRF machine if in doubt.
There are cheaper alternatives like ping test or acid test.

[EEVblog]

You can get phone apps that test for gold and silver coins based on the ping sound when struck.

[Kalvin]

Send it to me. If I send it back to you, it wasn't genuine. You'll know then.

[AF6LJ]

Send it to me. If I send it back to you, it wasn't genuine. You'll know then.

[calexanian]

Just find a gold miner like me! You can tell by dropping it on a hard surface and the sound it creates. Gold makes a thud/splat sound with little or no ring. Tungsten is very hard/brittle and will have a sharper sound with a ring.

[SeanB]

Plenty of fake coins around, and they are often very good fakes, almost perfect replicas. The major differences are that they often want to cut down on the plating thickness, so the stamping will have imperfections.  Another is that the colour is slightly wrong.

The most final test is to melt it down and filter it, which will show up tungsten as it will still be a powder, or to dissolve it in Aqua Regia and filter out the residue, which will be the tungsten. Then you can precipitate out the gold and smelt it back into a nugget. Probably right after giving the faker the tungsten back at very high velocity, lovingly wrapped in a copper Full Metal Jacket.

[mikeselectricstuff]

If resistivity is different, then the eddy-current damping effect on an inductor may be one way to measure it.
 

[Marco]

Your local dealer might let you use their XRF machine if in doubt.

Unless it puts enough X-rays into it to melt it how is it supposed to know what the core is?

If I were to design a machine to test 99+% gold bars/coins I'd use simple templates to test dimensions and then put them on a speaker/microphone pair to test audio range transfer function (getting a pure gold coin to ping is a bit of a pain).

[T3sl4co1l]

Your local dealer might let you use their XRF machine if in doubt.

Unless it puts enough X-rays into it to melt it how is it supposed to know what the core is?

Doesn't have to go all the way through, only the first layer and back.  XRF is any kind of bombardment (usually e-beam in an electron microscope, or a bremsstrahlung x-ray source otherwise) that excites inner electronic transitions, thus emitting characteristic spectra -- in all directions.  So there's plenty of scattering, and depth is only a matter of source penetration and power level.

If you need to go really deep, there's the possible advantage that heavy elements (like gold and tungsten) have quite high energy inner electrons (of course, only being accessible with an even higher energy of source radiation!). But I don't know offhand how that compares with the increased shielding effect, which one is stronger..

Tim

[langwadt]

Your local dealer might let you use their XRF machine if in doubt.

Unless it puts enough X-rays into it to melt it how is it supposed to know what the core is?

Doesn't have to go all the way through, only the first layer and back.  XRF is any kind of bombardment (usually e-beam in an electron microscope, or a bremsstrahlung x-ray source otherwise) that excites inner electronic transitions, thus emitting characteristic spectra -- in all directions.  So there's plenty of scattering, and depth is only a matter of source penetration and power level.

If you need to go really deep, there's the possible advantage that heavy elements (like gold and tungsten) have quite high energy inner electrons (of course, only being accessible with an even higher energy of source radiation!). But I don't know offhand how that compares with the increased shielding effect, which one is stronger..

Tim

I believe those handheld XRF tester specify something like 10s of um of penetration for heavy metals like gold and lead

[nowlan]

http://www.sigmametalytics.com/

Not sure how this one works. I saw it on youtube last year. The price is far too high.
How is it different from ultrasonic?

[akis]

Something a little different.

One of the saving/investment opportunities at the moment is gold coins. These come in many varieties but assume it is the standard 1 troy ounce 4 carat coin.

You mean it is an opportunity to sell not to buy anymore. That was like 1-2 years ago.

[EEVblog]

Unless it puts enough X-rays into it to melt it how is it supposed to know what the core is?

You don't past a certain depth. It relies on the fakes being really cheap fakes and hence very thin plating.

[woodchips]

Thanks, some interesting comments.

Questions;

Does the 'thunk' drop test also work when the tungsten is a fine powder? I would have thought that because each particle is isolated by the gold then it will be much harder to detect than a solid tungsten core?

Magnetism, never thought of that, would need to have a think.

Ultrasonics, used for crack detection but again, if a powder then too fine to see?

X rays, yes, just don't have one!

To tighten the spec down a little, something you could pull out your pocket in a coin shop and test the proffered coins. Doubt they would let you, but would know which shops to avoid.

Has anyone come up with a formula to calculate the side to side resistance of a disc? Resistance still seems the most basic, simplest, test to make.

[mikeselectricstuff]

http://www.sigmametalytics.com/

Not sure how this one works. I saw it on youtube last year. The price is far too high.
How is it different from ultrasonic?

My guess would be inductive - ultrasonic would probably need an interface liquid, and I'm not sure you could determine much with ultrasonic anyway.

[CJay]

You can get phone apps that test for gold and silver coins based on the ping sound when struck.

Now that's interesting, not owning any gold to test them with I wonder how accurate they are or if they're snake oil

[Kleinstein]

The sonic / Ultrasonic test can be very sensitive, if you have good values to compare with. So if the app has data for the real coins, they could detect really small differences. It could even go the way to tell what coin it should be just from the sound.
The nice thing is that it is really low effort: put the coin on a pice cotton, hit it with a small kind of hammer and record the sound for maybe a few points to hit it.

The frequencies are really sensitive to the elastic constants / sound velocities and to a lesser degree (because it is influenced from how to hold the coin) internal damping of the material. The sensitivity for the modulus is mainly limited to small natural variations in size - so something in the percent range should be possible. Gold and tungsten are different by a factor of about 2.5 for the speed of sound - that is way different.

There might be a chance to imitate the rather high damping of gold with a composite of tungsten and lead - but this would make the coin rather sensitive to temperature and the density would be slightly lower.

A combined test of something like density, sound and maybe conductivity is likely a good idea.

[dannyf]

You can melt it and measure the temperature.

Or to take it to a qualified jeweler.

[wraper]

Measuring the volume, weight and magnetism would give almost 100% certain result. You cannot get all of those even remotely right at the same time in the fake. You can make some RC oscillator with a coil in the middle of which coin will be inserted. If the tungsten is inserted, inductance should rise, therefore frequency will decrease. The opposite will happen if the gold in inserted.

[Kleinstein]

Tungsten is diamagnetic, but this is still a weak effect. So at normal frequencies adding a coin to an inductor will still reduce the inductance due to eddy currents - though not as much as the higher conductance gold one. Still the exact placement might make the bigger difference.

To measure the weak magnetic effect it would need more like a sensitive balance and a gradient magnet.

[wraper]

Tungsten is diamagnetic, but this is still a weak effect.

Tungsten is paramagnetic

[nowlan]

Here is a list of methods, in case anyone wanted to know more.
http://www.silvercoins.com/fake-silver-coins-14-ways-to-spot-counterfeits/

[T3sl4co1l]

Powder fakes aren't any worse to identify, because as long as tungsten is the dominant component, it will also be dominant mechanically and acoustically.  The gold acts as solder to join the tungsten particles together.

Tungsten is also rather hard to coin, assuming we're talking about real coins here and not blanks or bars.

If it's just a little bit added, to shave say 10% here or there, and maybe just in the center, keeping a heavy cladding of still legitimate gold on the outside, then that would be harder to identify and you'd be best off doing an ultrasonic or inductive test.

Tim