Do NGC or PCGS ever use XRF machines for coin authentication?

I’d place a $100 bet on that .

Keep in mind as technology gets more advanced the counterfeiters also have more tools at their disposal.

It like how militaries keep making more powerful weapons and in response armor industries have to keep upping their game to defend from those new weapons.

But I do think one day coins will be graded by AI and computer to get more accurate grading standards.

If a computer can map out and measure all the neuropathways in a brain it can certainly identify the scratches and luster of a coin.

 

Log in or Sign up to hide this ad.

You can throw away your scales, hand-held's and hand lenses. What we are dealing with is experience. That and an image library is what works now.

All it takes to detect just about any modern fake is a matching image from a similar suspicious coin. AI is only as smart as its programmer and ability to learn. So when the counterfeiters start making just one or three ancients from one die, the fun will begin.

 

That’s not always true. One of the dealers I went to in Florida was able to identify several fake American Silver Eagles with a scale.

They weighed 28.8 grams instead of 31.1 and since they were all in BU condition they wouldn’t have lost over 2 grams of silver in wear and tear.


I’m not saying they will identify every counterfeit out there. Obviously that isn’t the case.

But I find having a scale, calipers, and neodymium magnet to be the first line of defense in detecting fake

Most importantly they can all be used as non-invasive tests.

These days people are actually faking ASEs in NGC slabs graded MS70.

Sometimes it’s kinda funny because they have scratches all over and are still faked as MS70.

 

"Okay, yeah, I'll admit the APS was over the top."

I may be guessing low at the cost of a handheld, but you did get my point about the subject of discussion.

I read as much of those links provided, as I could in the time I had, and after reading the cpl's link, I clicked on a continuation link in that one, and if I understand it correctly, it states that it can go completely through some items that are not too thick, and will "report" back along the way what elements it finds. It really depends on the thickness of the item and the intensity of the beam. Remember, most coins are not extremely thick.
If I remember correctly, It was the second of his first two links that I followed through another link and it was in that one that I read the info about 'depth of discovery'. Please read it for yourself and see if you understand it the same way I did.
Be all, and any, of that as it may, I am very comfortable trusting the results of an XRF gun scan to tell me what my coins are made of.

 

If anyone wants a nondestructive technique that can provide true bulk analysis (i.e. average elemental composition of the whole coin), then it’s neutron activation analysis. The coin is placed in a nuclear reactor and bombarded with neutrons. Alternately, certain radioactive elements or particle accelerators bombarding a light-weight element can be used as a neutron source. Some neutrons will “stick” to the nuclei of the metal atoms of the coin, and these will now be radioactive. As they decay, they give off characteristic radiation. Usually it’s gamma radiation that is used for analysis. By gauging the energy and intensity of the radiation, the gamma spectrometer can determine the elemental composition of the object. Because neutrons and gamma rays can penetrate metal fairly well, the composition of the entire coin can be determined.

Although technically nondestructive, the analysis leaves you with a radioactive coin.

See the links for more info.

Cal

https://en.wikipedia.org/wiki/Neutron_activation_analysis

https://www.google.com/url?sa=t&rct.../chap2_6.pdf&usg=AOvVaw1VaA9-s7di44yfC_iJXqLT

 

A good, general value for XRF penetration depth is 100 microns (0.004"). As said in the articles, the depth can vary depending on the material and energy of the beam, but more than 0.010" would be difficult, especially with a handheld unit and no sample prep. This is usually deep enough to go through most plating processes (The copper plating on Zincoln cents is approx 8 microns).

I've frequently used XRF data (as well as XRD and SEM-EDS) as part of my job for years. Understanding how the equipment works and its limitations is important when analyzing the data. XRF is considered a surface analysis, but understanding the penetration depth is a key factor in understanding the data.

 

Now I really want a coin slabbed as "DETAILS - IMPROPER TRANSMUTATION".

 

Radiant toning

 

Dear Gam,

I'll repeat this so it sinks in: You can throw away your scales, hand-held's and hand lenses.

Gam3rBlake, posted: "That’s not always true. One of the dealers I went to in Florida was able to identify several fake American Silver Eagles with a scale.

They weighed 28.8 grams instead of 31.1 and since they were all in BU condition they wouldn’t have lost over 2 grams of silver in wear and tear.


I’m not saying they will identify every counterfeit out there. Obviously that isn’t the case.

But I find having a scale, calipers, and neodymium magnet to be the first line of defense in detecting fake

Most importantly they can all be used as non-invasive tests.

These days people are actually faking ASEs in NGC slabs graded MS70.

Sometimes it’s kinda funny because they have scratches all over and are still faked as MS70."

Unfortunately there is no way to defend my comment without coming off as a jerk; nevertheless, I have found that except for some long-time specialist dealers and collectors who MAY often know more about what they collect than many long time experienced authenticators at a TPGS, most dealers know swat. Slabbed coins guarantee that any good business person can be a successful coin dealer.

As for your example using SE, before your dealer got it on his balance the experienced authenticator would have know it was bad 99% of the time. EVENTUALLY, this may not be true but we are not there yet!

One of the biggest waste of times for me is to take a crappy, across the room fake and place it on the scale so I can write the low weight of the fake on the flip in order to prevent any controversy.

So you keep your scale, magnet, and other tools, as I believe EVERYONE should have them. I do. I just don't find them very useful anymore. There is a much bigger Boogieman coming.

PS The only magnetic fakes I've seen were gifts that were so crude looking that I would not embarrass any of my students by asking if the thing were genuine. Nevertheless, I do understand that the average person in the US has never seen a Peace Dollar and the Internet is full of "Ex-Perts" scaring everyone with crude magnetic reproductions.

 

That sounds ominous...?

 

Check this out!

https://www.sciencedirect.com/science/article/abs/pii/S0584854711001686

X-ray Fluorescence analytical criteria to assess the fineness of ancient silver coins: Application on Ptolemaic coinage
Author links open overlay panelVasilikiKantarelouaFrancisco JoséAgerbcDespoinaEugenidoudFranciscaChaveseAlexandrosAndreoudElenaKontoudNikiKatsikostadMiguel AngelRespaldizabPatriziaSerafinfDimosthenisSokarasaCharalambosZarkadasgKyriakiPolikretihAndreas GermanosKarydasai

Cite
https://doi.org/10.1016/j.sab.2011.08.001Get rights and content
Abstract
The application of X-ray Fluorescence (XRF) analysis in a non-invasive manner on ancient silver coins may not provide reliable bulk compositional data due to possible presence of a surface, silver enriched layer. The present work proposes a set of three complementary analytical methodologies to assess and improve the reliability of XRF data in such cases: a) comparison of XRF data on original and cleaned micro-spots on coin surface, b) Ag K/L ratio test and c) comparison of experimental and theoretically simulated intensities of the Rayleigh characteristic radiation emitted from the anode.

The proposed methodology was applied on 82 silver coins from the collection of Ioannes Demetriou, donated to the Numismatic Museum of Athens in the 1890s. The coins originate from different mints and are attributed to the first five Ptolemaic kings' reign (321–180 B.C.). They were analyzed in-situ by using a milli-probe XRF spectrometer.

The presence of an Ag-enriched layer was excluded for the majority of them. The silver fineness was found to be high, with very low concentrations of copper and lead. The composition data provide important information about possible sources of silver during the Ptolemaic period and indications of a gradual coinage debasement after 270 B.C. due to economic or technical reasons.

 

You would lose. You obviously have never seen that AI can do, up close.

 

https://itrexgroup.com/blog/ai-in-cancer-detection-treatment-applications-benefits-challenges/

AI is already way past what human capacity is in everything it is applied to. Coins are easy.

https://www.newsmax.com/us/facial-recognition-technology-nyc-cameras/2021/06/09/id/1024440/

 

We’ll see .

 

Not yet. But AI algorithms are steadily improving. And more importantly, the hardware they run on is speeding up exponentially. That's definitely not true for human minds.

AI will be better than human minds in a whole lot of applications. Some people find that frightening. I'm not one of them, usually. I've never been able to run faster than a sports car, or even a loaded minivan. I can't sing louder or more on-pitch than a synth playing through an amp and speaker. I can't do math faster than a PC running Excel. None of these facts diminish my human worth or dignity.

 

I will just remind you that AI actually IS already, today, better than human beings at ever task in intelligence that it is applied to, from reading amamograms, to playing go, you targeting minority ethnic groups for genocide in some aisan countries, to ticketing people for jay walking, to guilding space craft, to building airplanes, to targeting military targets in war...

And that is just today.

You can deny it, but it is true today,

And it is not fancy algorithms. It is artifical learning through neuronets based on biological intelligence. These are not programs, in the traditional sense. They are self programming programs that learn from experience and extroplate.

https://www.ritchievink.com/blog/2017/07/10/programming-a-neural-network-from-scratch/

 

I'm sorry, but that's just not true. You only see news releases about the applications where it fares well -- or, occasionally, the ones where it fails spectacularly.

You're also throwing in a bit of circular reasoning, where "ever[y] task in intelligence that it is applied to" only includes tasks where it performs well, because nobody applies it to tasks where it performs poorly. Or, if they do, they don't brag about it.

Neural nets are inspired by the organization of biological nervous systems. They aren't "based on" "biological intelligence". They do "learn", or rather "get trained"; the process is somewhere between training a dog to sniff for drugs, and "training" water to cut gullies as it flows down a sandy hill.

AI systems aren't little imps that are furiously studying to become gods. They might end up turning into that, but they aren't there yet.

 

No - we use it at the museum and my family works with it in medicine. These are not select cases, and they are not even news any longer. The recent mars landing was AI driven. AI is driving drug development and was essential for the COVId-19 vaccine work and for the tracking of the pandemic, and for the weather forcast. It is fair to say that almost no part of the modern world is unaffected by AI, doing things human beings simply can't do and doing it better. Did we discuss driverless cars and the Freight rail system and even Mass Transit?

 

I am sorry but driverless cars are NOWHERE close to even the most average human driven vehicles, much less professional humans at the top of their game. AI is not used in driverless cars edit except as it relates to gathering data used to refine the programs to sense things better. But they are not self programming otherwise every vehicle would "educate" themselves at a different pace and there wide be a wide gap of efficiency between individual vehicles. They use sensors and very fast computers to detect what is going on around the vehicle but are only as good as the sensors, which are nowhere close to human senses and the human brain, yet.