I've read a lot about the ring test and that people say there is a distinguishable difference in the tone of a known 80% quarter or dime and a 50% quarter or dime. The usual comments are that the 80% will ring higher (not completely correct) and and have more sustain (also partially correct). Here's the shake down. I did a sound frequency spectrum analysis of the drops of some pre 67 and then '68 - meaning there was no question of the content. Briefly, for those that are not familiar, a spectrum analysis plots the amplitude over the frequency range. What I found was that the higher silver content coins had resonant peaks that were from 1.5 to 1.8 kHz lower than the 50% coins. If the numbers were removed, the plots for each type of coin would look very similar, because they are the same shape object (a disc). Some peaks would be higher than others, but they'd mostly all be there. However, the difference in numbers between the two types of coin was stark. This makes a lot of sense since the speed of sound in copper is a lot higher than that in silver, meaning, the more copper in the coin, the higher resonance frequencies it would be able to support. Looking at the plots I also noticed, that with the 50% silver/copper, the sharpest peaks (the ones we perceive the most clearly - the ones with highest signal/noise ratio) was on the higher side of the frequency spectrum, and it was at the 13.3 to 14.2 kHz range (mostly around 13.6 kHz) for quarters and even higher for dimes (17-18 kHz), a range that a large population will lose as they get older. That is why the 80% silver one seems to ring louder with more sustain - its sharpest and most distinct peak is at about the 11.5 to 12.4 kHz range (mostly around 12 kHz) and is more in the middle of the sound spectrum that humans hear. Most people throughout their life (again, MOST people) will not experience as great a loss of hearing damage closer to the middle of the human hearing range. So that's why the silver ones sound sweeter - because we can hear it better! The frequencies that resonate in the 50% coins that are more in the range that we can hear the best are the lower ones that do not have as high of a signal to noise ratio, therefore they sound deader. I think that's because they get muddled up in the sound of the coin clashing into whatever you've dropped it on. Anyway, I'm really excited about this. I'm including an example of each type of coin. Picture1 is of a 50% 1967 quarter and picture2 is of a 80% 1967 quarter. Even though each quarter's frequency plot was unique in its own way and even varied about 7% for each type, it was not at all difficult to tell the difference, there was never one where I was wondering, is this one or the other? I did not do a formal scientific analysis of the data, these are just rough observations so far. But the differences between the two types of coins is different enough that an "eyeball" test can even make sense of the results. Picture1: Picture2:
Yep, that would be correct. Copper density (8.94) being lower than silver (10.49), a 50% coin will have a lower mass, and therfore a higher spectrum of harmonics. But as you point out, owing to the human ear (and fletcher-munson contours will show that at the phons level you hear a coin, there's a possibly around 10dB descepancy between 15 and 17KHz), this effect can be masked. However, what's intriguing about your spectrum plots is they seem to show a marked suppression of the fundamental and first harmonic, and exacerbation of the higher order odd harmonics... Curious... Possibly due to cancellation owing to disk shape? Sorry, being a research engineer makes this kind of stuff very interesting for me.... (yeah, I know - geek alert! )
Wow ! that's really cool. I can usually quickly search through a bunch of coins & find the silver ones based on the tone they give out. I've noticed the 45's give off a really nice tone. The 80's as you say are harder to pick but interestingly, & confirming your theories above, my daughter has no trouble detecting the specific tone. According to her I'm just old, decrepit and immature.
Samboyellowsub, that was a great first post, something usefull to read finally I am assuming that the coins were hitting the same surface and were dropped from the same height and the angles were the same.
I like how interested you are! A fellow science person, very cool. The masses are the same actually. That is why it is harder to tell for the average person. The thicknesses are slightly different. Because the densities are so close, it is hard to tell even with a good micrometer. The neat thing about the different modes was, that depending on where the coin hit, the edge or the face or anything inbetween, different modes would ring louder than others according to the plot. All the modes would be there all the time, but they might be at different decibels each time. I thought that maybe that higher harmonic with the higher signal to noise ratio might just have that higher ration because the lower modes were drowned out by some of the initial lower register noises of the coin first hitting the wooden floor?
Initially i payed very close attention to keeping the drop height and microphone distance as consistent as I could make it, but then I realized that those peaks are there no matter what and it didn't really matter if I kept the the drop conditions the same. Differences affected the overall volume or the relative height of each peak, but the frequencies at which the peaks would occur would remain the same within the same coin. I verified this with several coins by trying to drop them in very different ways to achieve a different results, yet after that, i could still tell what the composition was. good questions.
Samboyellowsub, When i was writing that question it did occur to me in the 'back' of the mind that the spectra would be independent of the mechanical inital considerations, and since you also left that out in the description. I completed a physics degree at uni, ( some years now ), so it was 'fun' and nostalgic watching a 'true' scientific experiment being reported on. Initially I thought it was to do with light reflection ( spectroscopy ). I will be mailing a gold coin into the USA tmrw ( bought here sold there ) and might as well mail to you a coin as free gift and a little reward from fellow stacker here in oz. ( you are welcome to do the same exp on this coin ) Here it is, Herr Professor Doktor Karl Friedrich Gauss, the Prince of Mathematics It is 5 DM, 0.625 Ag and 0.225 oz asw. So if your in, than PM your postal address and will ship tomorrow. ( dont worry, i aint a stalker )
Hi Samboyyellowsub, I've sometimes contemplated if I could develop an acoustic resonance excitation device designed to do a spectrum analysis of metals, based on similar principles to your experiment. The idea would be to strap two transducers on to the item for test, and "excite" the metal with a sweep of dual phase/ frequency modulated waveforms. The second transducer would look for spectrum "nulls" and through a series of algorithms would thus determine the average density of the metal. Obviously this method is not going to tell you which individual metals are in a particular alloy, but if you already know that, it should be able to give you a "percentage of purity" indication. It certainly wouldn't replace an XRF, but it's not going to cost anything like one either! Unfortunately it's would take a paid 6 month holiday from work to develop, so I guess for now it remains on the "ideas" list...
Ah-ha! I wasn't aware of that! I'd be interested to know the exact distance from rim corner to opposing rim corner, to see how that correlates (or doesn't) with the theoretical lowest standing wavelength.....
I'll measure it with a micrometer with .02 mm graduations. you wanna work out the math too? I'll have to dig up the pertinent physical relationships for speed of sound vs. elastic constants of silver copper alloys and we can see if the lowest fundamental makes sense. Can't measure until tuesday though. I'll PM you or something.
it'd be very useful to have a device that automatically can recognise specific alloys, so you can just plink down a coin and the gizmo says '999 silver' - or not - and the stamp on the coin can be verified. i love the sound of silver, it has this delightful clink and diminished sustain compared to copper and nickel.
Thucy, I finally did the drop test on the German coin that you sent me. My wife loves the simple and clean artwork of German coins. The two pictures show a different kind of drop. Some harmonics look stronger than others in the two plots. I noticed there is a lot of noise, but I'm afraid that's the best that can be done with a computer microphone with the fan running. I was thinking that something could be developed to test coins of an assumed composition of metals. Take the composition of what a coin is supposed to be and calculate what the expected harmonics would be based on the dimensions of the coin and then test a sample. I've been thinking about this for a while now. It would prove extremely difficult or impossible to test a coin and guess at its composition based on a plot - that sort of analysis would give you a wide range of compositions at best. Cheers, Sam
Find the resonant peaks of copper, nickel & zinc and you have identified 95% of basemetals found in coins. There is a project in for you. Maybe it could be done, where you put the coin in and from the peaks you deduce what metals its made out of.
I think that Silvertongue's idea is great. Use an audio frequency sweep generator to ring the coin with a pickup attached and see where the peaks are. Use piezo pickups as the exciter and the sensor and map out the resonances in the pickups first to take them out of the equation.
