Hi,

Ive been a long time reader in these forums and this is my first post.

I am hoping someone here can help me out I recently bought some Silver Britannia 1oz coins on ebay and checked their weight and size. Both these coins are the correct size however the weight is slightly different to that shown on the web, my coins weigh 32.52 (bullion version 2003) and 32.79 (proof version 2001) grams respectivly when they "should" weigh 32.45. Are small differences like this common place with silver britannias or could these be fakes?

How common are fake britannias?

Thanks in Advance.

Where did you buy them from?

Ebay, but they were two sellers with good feedback.. 1000+ positive.

take a picture of them and put it up here.

lightly bang them together - what kind of sound do they make? - should be a clear 'ding'.

Yeah they make a "ding".. at least more of a ding than banging two 50p's togeather, problem with this is I dont have a known good coin to compare with.

Ill post some pictures later when I can get my hands on a good camera.

Thanks.

I'm sorry if this sounds at all patronising, but I'd like to eliminate the easy things first.

How did you weigh them? What instrument was it? Balance or digital scales?
Do you have a reference specimen, ie a known good coin or weight?
How many measurements did you make of each.
If more than one what were the other measurements?

This looks like a lot of hard work, but it should take less then 10 minutes. Don't be put off if it looks complicated; it isn't.

Check the precision of your weighing method by the following actions:

Leave the balance to settle. Zero/Tare it.
Weigh coin.
record weight
remove coin.
check to see if measurement returns to zero.
Record this weight without coin.
re-zero/tare even if display reads zero.

repeat until you have at least 20 measurements for each coin.

Do this with your ebay coins and also with a 10p or 5p etc.

Now using excel, find the arithmetic mean (average) of the values and also the standard deviation using the following excel functions:

=average(seriesofvalues)
=stdev(seriesofvalues)

then express the stdev as a function of the mean.
i.e stdev_value/average

this should give you a value (hopefully) in the range of 0.01 - 0.10
Multiply by 100 to convert to a percentage.

This percentage value is known as the CV (co-efficient of variance) and is the precicion of your measurement method.

The average value is the accuracy of your measurement method.

This will determine how good your analysis METHOD is.

If your METHOD is producing errors of say 1% then it means that you cannot draw any conclusions about the coins as your methedology/analysis method is flawed and is producing INTERNAL MEASUREMENT deviations which are just as much as your observed deviation from the expected value.

If you get a tiny intra-sample variation by you can be sure that your measurement method is precise.
If the average is what is should be, then it can be said to be accurate. (Precision and accuracy are not the same thing).

If you are satisfied that your measurement method is accurate, we can analyse the data.

The values you have given in the original post vary from the 'true' weight by 0.22% and 1.05% from the reference. If the precision (CV%) value of the measurement method is significantly lower than the recorded deviation then you can probably be sure that your observed deviations are from real events, not methodology errors.

I note that both of these values are HIGHER than the 'true' value and may be as a result of fat and sweat (ie dirt) on the coins. All accredited reference weights should be handled with tweezers for this reason.

Not to check the hypothesis that the deviation from the true value can be arttributed to dirt.

If you do this with measurements with a new-ish 5p coin and obtain an average value of x. Compare this with the 'true' value of the 5p coin from the royal mint website.
This will tell you what handling a coin will do to it's weight when compared to a known reference (fromthe website). If the difference (%) between the observed weight of the 5p and the true value is similar to your observations with the silver coins, then you have shown that handling the coins might be a source of the increase in mass.

Believe me, this looks like a lot of work but it isn't.
You can verify if you scales are accurate and precise by this method.
You will also have shown a potential increase in mass caused by handling, which can be compared to your experimental sample.

I have made a quick and dirty excel spreadsheet for you. You just fill in the weights in the yellow boxes and the calculations should be done for you. - I can't post it here it won't let me. DOes anyone know how I can upload an Excel file? I guess it will have to wait.

PM me if you want more information, or just post the 20 values for each of your 2 coins and 20 values for the 5p and I'll do the rest.


Alternatively, take them to a dealer and try to sell them to him. If he throws them back at you, you've been had.

Mr.Mangle

Thanks for the reply.

I did this with a small sample of 12 for each coin. My scales are supposed to be "accurate to 0.01 upto 100g" according to the box!

Unfortinuatly I do not have a known good coin since I only have a couple of coins from ebay. Measurements are below.

10p 2003 2001
1 6.4 32.53 32.57
2 6.41 32.51 32.59
3 6.4 32.54 32.58
4 6.42 32.53 32.58
5 6.38 32.53 32.57
6 6.4 32.51 32.59
7 6.41 32.52 32.59
8 6.4 32.54 32.6
9 6.41 32.51 32.61
10 6.4 32.52 32.57
11 6.4 32.53 32.59
12 6.42 32.54 32.6
Mean 6.404166667 32.52583333 32.58666667
Standard Deviation 0.010836247 0.011645002 0.013026779
Standard Deviation / Mean 0.169% 0.036% 0.040%

Edit: that looked a lot nicer in the editor than on the page!

I have a 100g gold bullion "bar" from goldline which these scales get to be 100.02g and a 100g calibration weight that came with the scales which is measured at 100g.

Never even thought about it being grease on the coins, since I have only had them out there capsules to weigh them.

What do you think?

Well, there is good precision with those scales - very small deviations from the mean, which is a good sign.

The 10p coin gives a value of 6.404g which is 1.5% above -expected
Your silver coins are 0.23% and 0.42% below the expected.

I can't account for the discrepancies - the grease and dirt hypothesis might hold water to raise the mass of the 10p but I am not sure about the silver. In our lab we used to test our scales daily using a range of weights from 0.010mg to 100g which were directly linked to accredited sources. The weights were never touched with human hands. They were taken out of the box with tweezers, the measurement made, and repeated then back in the box. Never had sweaty fingers on them and never allowed to accumulate dust. We did this to avoid grease deposits, so I can't tell you how much it would interfere with the readings as we never tried it.

Things to note:

The gold is almost spot on the expected value and is from a reliable source - so you can probably take this at face value.

The silver is below the expected mass, and is from unreliabel source. I don't think this automatically means it is fake. Probably more to do with the lower production quality specifications of the silver than compared to the gold. Silver is a lot cheaper as it's a lot cheaper metal so they don't try so hard to get it accurate.

I wouldn't worry about it. These are small deviations from the expected value.

I have a few 2007 silver Britannia's at home, I bought them from chards in an investment pack of 10 still in the Royal Mint Packaging. I have just tried weighing them individually and found the weights to be within the range of 32.5 to 32.8