After reading Jason Poe's 2010 article on strike, I am now an expert...(I just ordered a copy of his book, https://www.amazon.com/gp/product/1492356107?tag=petetdavi&creativeASIN=1492356107 ) Knowing there are lots of variables, does anyone have any idea of ~ how many coins can be struck with a die before it has to be retired?
There are too many variables to give an accurate estimate. I'm sure that most Morgan collectors will recall that the first Philly dies used in 1878 shattered after a little more than 300 strikes. Chris
That would depend on what's being struck. Take shield nickels after 9 to 10 k strikes the dies were shot.
Yup. That's what makes some of the 8tf coins so rare and valuable. There must've been a lot of "oh crap not again" being said in Philly during those early days
I've seen records of Morgan dollar die usage, in the form of mint logs, from a couple years that show an average of about 130,000 per die. In the log for 1878-S, there are some that are under 10,000 per die that were from May (SF began Morgan production in mid-April). I assume these are the rare B1 reverses. There are also some that are over 250,000. For 1881-S and 1898-S, there were some outliers that were over 600,000 coins from a die. The 300 number is seen as inaccurate for the first dies (VAM 9). This was the number of coins that was in the first delivery from the coiner to Linderman, the mint director, on March 12. The total mintage estimate of 750,000 for the 78 8TFs is seen as quite reasonable, however, and this was 41 die pairs. I did a quick, seat-of-the-pants deduction as to just how rare some of these die pairs are during the ANA summer seminar, and we decided that there are clearly a few varieties for which one die failed after no more than a couple hundred strikes.
"Die life" will vary based on what period of history you are talking about. Major factors include die manufacturing techniques, striking pressure, denomination, and metal being struck. A smaller coin, in general, is going to have a longer die life. As paddy mentioned, nickel is a very hard metal and will decrease the life of a die. Some dies, as mentioned earlier, failed after a few hundred strikes. This was more common in the early days of the mint. In the early 1800's, dies would often last 50k-100k (again, depending on the series and metal struck, there may be great variety in this number). In the late 1800's, it was common for dies to last 100k - 200k. In the modern mint, the average die lasts much longer. Cent dies last 1m - 1.5m, nickels, quarters, and dimes are around 500k - 700k, and half dollars are 200k - 300k.
According to http://www.pcgs.com/News/The-1982-No-P-Mintmark-Roosevelt-Dime, the expected lifetime of a dime die in 1982 was 75000 strikes, resulting in the 1982 no P dime being scarcer than the 1916-D. I'd hate to see those Morgans that resulted from the 100000th strike of the die. I suspect there are some undiscovered VAMs accounting for the huge die lifetime.
I'll take the heat for this BUT: The majority of VAMS are like sand at the beach. I've been told by two TPGS graders doing dollar attributions that they can take virtually ANY date and (they or anyone else) can find a "new" VAM with no effort at all! IMO, VAMS are just one nice way to ID dies and die states. It's too bad the TPGS's don't have the time or a simple way to mail in photos of all the different dies they could add. One of the graders told me if a customer wants a VAM for a date like 1881-S, he is told to look for 1-2 minutes. If he cannot find it, issue a refund and move on since many VAMS carry no additional value. Nevertheless, they can be interesting are collectible. BTW, a month ago, I wrote that several of my Top-100 VAMS graded AU - MS-62 by ANACS sold for $38 or less and one sold for $44. Each of them "booked for over $100 yet they sold for double melt. Lessons learned: 1. The majority of VAMS are virtually worthless unless they grade MS-63 or higher. 2. Me and other collectors would rather "cherrypick" a good one than buy it. 3. VAM collecting may have lost much of its luster - just as DCAM Kennedy halves.
Ironically, there may be VAMs that are duplicates of others that were listed for traits that appeared due to die wear. These duplicates would then need to be consolidated into a single listing. This is especially true for those dies that were used to strike 600,000+ coins.
There is some pretty good image analysis software that could be applied to the PCGS True View/Secure image database that could ferret out new VAMs. But there might not be any profit to PCGS in doing so. Not sure if they are using image analysis software yet for variety attribution on submissions, but it's just matter of time until it happens. Automated grading is a ways off, but automated variety identification can be done now. Cal
Image analysis software will only be able to attribute varieties if the images for doing so are of sufficiently high quality. When this happens, it won't be with VAMs, but with early type.
It would seem like early type would be relatively easy to make a program to attribute. The elements were hand punched, so varieties are easy to see and attribute with the naked eye. I agree that high quality images are required, but if you have a standardized setup which fully lights the coin, I would think this would be practical. VAMs, which often require an electron microscope to examine the coin on a molecular level, would be significantly more difficult.
A die clash can often cause irreversible damage to the dies. Many major die breaks are the result of die clashing. So if the coin press has a reliable feeding mechanism and consistent planchets are fed into it, the dies are likely to last longer.
Being that mint marks were hand punched into working dies up until 1989 - location and orientation of this device relative to the rest of the coin design fairly simple for adapting image analysis for mint marked coins. I've been using CAD and triangulation of mint mark location relative to a fixed landmark design device as an attribution verification tool for years.
I think PCGS lost a zero in that estimate. Back in the 1960's the average die life for dimes was 500,000. Nickels which have the lowest die life were 300,000.
The difficulty lies in getting images sufficiently similar in quality and lighting. The optical technology is (relatively) easy; I think it would take 24MP or better to create an image large enough to operate with. The software, though, would be a bit....tedious. Something adapted from fingerprint recognition would do. That, and a whole_bunch of processing power.
