Android app uses resonance test to detect fake silver coins

[bordsilver]

Oh man. Why am I stuck with an iPhone?

You could pick up some old second hand Android phone for just a few bucks on ebay. Or perhaps you know someone who has an old phone they are not using? Old Android phones are not worth much. It is entirely feasible to have one around just as a "fake detecting droid"
The app should run on any Android phone with Android version 2.2 or newer.

In fact, the old original "HTC Desire" running Android 2.2 has the best microphone for picking up high resonance frequencies of all Android devices I have tried.
It beats my Nexus phone hands down.

But until now I have only come across one Android device where the microphone is weak enough to actually make it a bit difficult to use the app, and that's my Acer Iconia tablet.
I guess Android devices are not originally designed to be used to pick up faint resonance frequencies in the 20kHz range, so I can't really blame them.

Can you physically run the android app without having an account/sim card?

 

[joebush]

Interesting concept, and it seems like a simple way to add an additional test. I have a question.

The list of coins to test is quite limited. Are there plans to allow user updated results? This would be interesting since empirical data would be massive.

The problem of allowing user added coins is that a user might unknowingly have a coin type in their stack that is fake - which (at least for a time) would become the accepted baseline for that coin type.

Yes, that is the primary reason why I am hesitant to include such a feature.
I have a list of other features which would also help improving the number of supported coins, without the risk of introducing fakes as a baseline in the app.
When/if the features will be added depends on how much time I have to spend on improving the app.

That's only a problem if there is an assumption that a single, or only a few entries would determine validity. If a large number of individual tests that coincided were needed for validation, only then can it be considered substantiated evidence as the the validity of a coin.

 

[Nikon]

Oh man. Why am I stuck with an iPhone?

You could pick up some old second hand Android phone for just a few bucks on ebay. Or perhaps you know someone who has an old phone they are not using? Old Android phones are not worth much. It is entirely feasible to have one around just as a "fake detecting droid"
The app should run on any Android phone with Android version 2.2 or newer.

In fact, the old original "HTC Desire" running Android 2.2 has the best microphone for picking up high resonance frequencies of all Android devices I have tried.
It beats my Nexus phone hands down.

But until now I have only come across one Android device where the microphone is weak enough to actually make it a bit difficult to use the app, and that's my Acer Iconia tablet.
I guess Android devices are not originally designed to be used to pick up faint resonance frequencies in the 20kHz range, so I can't really blame them.

Can you physically run the android app without having an account/sim card?

Yes.
The app does not require a sim card.
Your need an internet connection (wifi should work fine) and a google play account to install the app.
But no connection is required to run the app.

 

[Nikon]

Interesting concept, and it seems like a simple way to add an additional test. I have a question.

The list of coins to test is quite limited. Are there plans to allow user updated results? This would be interesting since empirical data would be massive.

The problem of allowing user added coins is that a user might unknowingly have a coin type in their stack that is fake - which (at least for a time) would become the accepted baseline for that coin type.

Yes, that is the primary reason why I am hesitant to include such a feature.
I have a list of other features which would also help improving the number of supported coins, without the risk of introducing fakes as a baseline in the app.
When/if the features will be added depends on how much time I have to spend on improving the app.

That's only a problem if there is an assumption that a single, or only a few entries would determine validity. If a large number of individual tests that coincided were needed for validation, only then can it be considered substantiated evidence as the the validity of a coin.

Som kind of user-side option to record coins is on my list of potential features for future updates of the app.There are a couple of technical problems to solve in order for this to work reliably. But if I manage to find the time, then something like this might be added in the future.

 

[Midnight Man]

Nikon, thanks for many detailed answers!

It's great to see the community working together in these ways, and your app is very interesting indeed!

A thought (in brief, as I'm just scanning posts tonght being busy doing other things) - I hear your comments on bars. I would expect they would be hard to make resonate, but the practical side of me must ask what is probably a very obvious question - have you tried suspending a bar inside an open weave bag that allows a gentle belting with a hammer? I'm thinking a bag along the lines of this design: http://www.ecobags.com/ECOBAGS-Reusbale-Soap-Bag

Will put some things in the back of the brain for a couple days, and keep an eye here in the meantime

 

[worldbubble]

...

A thought (in brief, as I'm just scanning posts tonght being busy doing other things) - I hear your comments on bars. I would expect they would be hard to make resonate, but the practical side of me must ask what ....

hit hard with something hard and bar will resonate ))





I'd like to see more tests comparing fakes and genuine coins before I'll jump on buying this app.
I've noticed you had a resonance chart ... what about exact frequencies for the test?
Another suggestion is to start the community where frequencies can be discussed as different people might get different results ... this might be helpful in determining margin of error as well.

 

[Nikon]

...

A thought (in brief, as I'm just scanning posts tonght being busy doing other things) - I hear your comments on bars. I would expect they would be hard to make resonate, but the practical side of me must ask what ....

hit hard with something hard and bar will resonate ))





I'd like to see more tests comparing fakes and genuine coins before I'll jump on buying this app.
I've noticed you had a resonance chart ... what about exact frequencies for the test?
Another suggestion is to start the community where frequencies can be discussed as different people might get different results ... this might be helpful in determining margin of error as well.

I suggest you install the free version of the app and try it out. Try the fakes in the videos. Your can also try completely different coins and see what happens. Play around with it to get a feel for how it works.
The measured frequencies are displayed in the app. As well as a number indicating hos many percent they deviate from what is expected for the selected coin.

 

[worldbubble]

...

A thought (in brief, as I'm just scanning posts tonght being busy doing other things) - I hear your comments on bars. I would expect they would be hard to make resonate, but the practical side of me must ask what ....

hit hard with something hard and bar will resonate ))





I'd like to see more tests comparing fakes and genuine coins before I'll jump on buying this app.
I've noticed you had a resonance chart ... what about exact frequencies for the test?
Another suggestion is to start the community where frequencies can be discussed as different people might get different results ... this might be helpful in determining margin of error as well.

I suggest you install the free version of the app and try it out. Try the fakes in the videos. Your can also try completely different coins and see what happens. Play around with it to get a feel for how it works.
The measured frequencies are displayed in the app. As well as a number indicating hos many percent they deviate from what is expected for the selected coin.

thanks ... but before I'll go for android phone and then ... ))

anyway - if application does what it promises to do - thumbs up for it!

 

[VeriGoldSoundLtd]

Oh man. Why am I stuck with an iPhone?

Don't worry. There is the same kind of testing app for iPhone -> CoinTrust (www.cointrustapp.com).

You will find it here: http://georiot.co/45jC

 

[RT]

Awesome Idea, well done.

Let me put my conspiracy theory hat on:

Someone makes a similar app that enables the GPS and it logs the amount of coins/bars you test, then sends the results back to the creator. They now have GPS co-ordinates of your stack.

They can already track your phone (try taking out a battery on a smart phone, wait, cant do that either snymore), Im sure if it came to that, it wouldnt be hard to check sites like this, cross referance site names with email accounts, facebook, twitter and the rest and find out a lot more.

My amazement never stops with how much infomation people allow to be collected about themselves in so many ways and now you cant do anything to stop it from happening. It might be fine today and no reason to be wary, but its a bit late in 5yrs time once all this stuff is innocently gathered but then the nazis or commies or some other group come to power and suddenly your quiet, I do nothing wrong, head in the sand approch gets your stack taken away because of the "greater" good. US already made it illegal once to hold gold, could it happen again?

Its ok but, governments never do that type of thing.... right?

Im going to go put my al foil hat on now before it starts

 

[clubman]

Great app, but I don't have any of the coins in your database, I only have maples.

 

[kennethrsnyder]

Very cool, got it to work first try! Thank you and keep up the GREAT work!!!

 

[Nikon]

I just released an update of the full version of the app.

News in this update:
* It now has a built in feature that supports ALL 1oz .999 silver coins with diameters between 36mm and 42mm.
The way it works is you enter the diameter of the coin, and the app will automatically calculate what a silver coin with those dimensions shall sound like.
So no matter what type of coins you like to stack, as long as they are 1oz .999 silver coins, the app will support them.
If you happen to have a 1oz .999 silver coin with a diameter outside of the supported diameter-range, let me know and I'll try to extend the range of supported diameters.

* Added support for Swedish 5kr and 100kr Silver coins.

* Tweaked the qualification limits slightly.

* Any sound quality problems are now signaled more clearly in the form of a yellow question mark instead of a red "X".

Very cool, got it to work first try! Thank you and keep up the GREAT work!!!

Thanks!

Great app, but I don't have any of the coins in your database, I only have maples.

The full version of the app has support for Maples, and a whole lot of other coins as well.

You can find the full version here:
https://play.google.com/store/apps/details?id=se.brolinembedded.bulliontest

 

[mmissinglink]

Nikon,

It's a neat app. I am interested in its development. I have read most of the comments in this thread, but not all. So if my question has already been answered, please forgive my impatience here.

When impacting a coin (tapping it for example) to produce a resonance sound, does the material the coin is held with or impacted with or the manner in which the coin is held (amount of pressure and surface of coin covered by the thing holding it) matter enough to create a variation significant enough to potentially throw the resonance sound results off?

Thanks in advance for the response.

 

[pdkbffwleo]

Awesome Idea, well done.

Let me put my conspiracy theory hat on:

Someone makes a similar app that enables the GPS and it logs the amount of coins/bars you test, then sends the results back to the creator. They now have GPS co-ordinates of your stack.

 

[Nikon]

Nikon,

It's a neat app. I am interested in its development. I have read most of the comments in this thread, but not all. So if my question has already been answered, please forgive my impatience here.

When impacting a coin (tapping it for example) to produce a resonance sound, does the material the coin is held with or impacted with or the manner in which the coin is held (amount of pressure and surface of coin covered by the thing holding it) matter enough to create a variation significant enough to potentially throw the resonance sound results off?

Thanks in advance for the response.

It is a bit like playing a musical instrument. A key on the piano is always going to produce the same frequencies, but it is also possible to hit a "false note".
Such a "false note" usually still contains the expected frequencies, but it also has a bunch of additional unwanted frequencies.

If the app detects that the sound quality is too bad (typically a "false note" or background noise) then it will either simply tell you: "Too much Noise!", or (sometimes) indicate that the coin does not sound like a genuine coin.
Once you learn how to handle the coins, it is actually very easy to repeatably produce good sounds. I recommend you use the technique shown in the youtube video, it is easy and produces consistently great sound with a minimum of noise.

I also recommend that you look at the spectrum diagram and learn to recognize how a bad and a good sound looks. A "perfect ping sound" for a 1oz silver coin should have three or four distinct frequency peaks close to where the blue lines are, and nothing more. So the frequency diagram gives you instant feedback about how good the quality of the sound was.


Also know that most Android phones have very sensitive microphones, so it usually is of no importance how loud the sound is. A barely audible "ping" produced by gently tapping the edge of the coin with your fingernail is sufficient. Sound quality is far more important than loudness.

That said, a few Android pads do have less sensitive microphones and require a louder sound for the microphone to pick it up.

 

[Nikon]

I just published an update of the full version of the app, with a few requested features:

* Support for American "Morgan" and "Peace" silver dollars.
Seated liberty dollars will probably also work, but this is not yet verified.

* Screen will be cleared momentarily between each new measurement.
This makes it easier to test large amounts of coins, as it becomes more clear when a new measurement is done.

* Action button to manually clear screen between tests.
A small but requested feature.

Let me know if you run into any problems!

The update is available here:
https://play.google.com/store/apps/details?id=se.brolinembedded.bulliontest

 

[fromost]

Hello, I hope this is still the correct place to discuss the app, the thread has been idle for more than 1 year.

I think the app should also provide and automatically check sound profiles for known fakes (tungsten, as well as other alloys).

In the case of gold maples, the profile supposedly contains 5 frequencies, however the checkmark is shown even if only 3 of the 5 appear in the spectrogram, see below (in this case, I had selected the gold maple, but was tapping on a Sunshine mint 1oz silver round):
[img=Pinging Sunshine Mint 1oz silver round matches 1oz Gold Maple profile]http://ohmi.org/~radu/files/Screenshot_2014-10-23-22-49-41.png[/img]

In this case a 4th resonant frequency is listed as detected (18626Hz) , however its amplitude is indistinguishable from the noise floor.

It's a brilliant app, please keep working on it.

If the time comes when transacting in gold/silver coins becomes a daily reality, this app will be critical for building trust, given how many high-quality fakes are coming out of China these already. The fakes will improve probably to the point where secondary tests like spectrometer tests or ping tests will be required for daily use.

 

[Nikon]

Hello, I hope this is still the correct place to discuss the app, the thread has been idle for more than 1 year.

There is no better place. But yes, the thread has been a bit idle for a while.
Sorry for the late reply.

I think the app should also provide and automatically check sound profiles for known fakes (tungsten, as well as other alloys).

It is not possible to check a sound profile for a particular alloy, since it is allways the alloy in combination with the physical dimensions that determine the sound profile.
There would be little point in adding known fake coins to the app. There are too many of them, and they may change their alloy at any time. It would be impossible to keep up.
But since you are interested. Here are some examples of what fakes look like:

The image above is from a 25gram Tungsten coin. You don't really need an app to figure out that something is wrong here. The lowest resonance frequency of the tungsten coin is 17420Hz. It is even inaudible to the human ear!
Proper 1oz (31.1gram) tungsten coins will have a slightly lower resonance frequency. Possibly within the range of human hearing. But it will still be waaay up there.

The image above is of a pretty decent fake Silver Eagle. Visually it looks good, and it has the correct dimensions and weight.
But, well, the frequencies are way off.
I think this one is made of silver plated brass.

In the case of gold maples, the profile supposedly contains 5 frequencies, however the checkmark is shown even if only 3 of the 5 appear in the spectrogram, see below (in this case, I had selected the gold maple, but was tapping on a Sunshine mint 1oz silver round):

Some coins, such as gold maples, have two frequencies located very closely together.
Due to manufacturing variations, the frequencies of authentic coins vary by several percent. To handle this, the app just make sure that the average difference between the expected frequencies and the closest measured frequency is less than 4%. This means that a single frequency somewhere close to where the "pair" is expected is sufficient to pass.

The reason why the sunshine mint silver round passes as a gold maple is because of what I wrote above about alloys and dimensions. Given identical dimensions, gold and silver will have very different sounds. But a Sunshine mint round does NOT have the same dimensions as a gold maple. The difference in dimensions has a big impact on the sound, and in the case of sunchine min rounds and gold maples, the difference in dimensions perfectly compensates for the difference in alloys.

It would actually be pretty easy to take piece of silver or brass, plate it with gold, and make it look and sound like a gold maple. There is only one catch: This "perfect fake" would have to be something like 8-15mm bigger than a authentic gold maple for the sound profiles to match!
That's why I print out the appropriate dimensions of each coin in the app.

I always recommend that the dimensions, weight and resonance frequencies of a coin are tested. These three tests taken together will verify two different physical properties of the alloy: The density, and the speed of sound through the alloy.

In this case a 4th resonant frequency is listed as detected (18626Hz) , however its amplitude is indistinguishable from the noise floor.

Indistinguishable from the noise floor? It is something like twice the nosie floor. More than enough for the app to pick it up.

It's a brilliant app, please keep working on it.

Thanks!

 

[Theantifunie]

Is there any possibility of this app coming to IOS??? I hope so!