Computer Says...
May 25, 2011 4:37 PM Subscribe
How accurate are instant assay machines for precious metals?
I took a palladium bar stamped 99.5% pure (with a known Russian producer's mark on it) into a precious metals dealer. He put it in a fancy box that was hooked up to a computer, and the computer screen spit out percentages of every metal that was in it. The computer said roughly 95% palladium, and a few percent other metals. (the whole procedure took about 2 minutes). I don't know what the machine is called but it was very cool. I questioned him on its accuracy... I understand why someone would counterfeit a bar, but it would seem odd to slightly understate the metal in it; the profit margin from such jiggery-pokery would seem to be small. So do these machines have a margin of error? Or might the refiner have somehow produced a bar that was less pure than stated?
I took a palladium bar stamped 99.5% pure (with a known Russian producer's mark on it) into a precious metals dealer. He put it in a fancy box that was hooked up to a computer, and the computer screen spit out percentages of every metal that was in it. The computer said roughly 95% palladium, and a few percent other metals. (the whole procedure took about 2 minutes). I don't know what the machine is called but it was very cool. I questioned him on its accuracy... I understand why someone would counterfeit a bar, but it would seem odd to slightly understate the metal in it; the profit margin from such jiggery-pokery would seem to be small. So do these machines have a margin of error? Or might the refiner have somehow produced a bar that was less pure than stated?
That fast and non-destructive can pretty much only be an XRF. As GEM notes, calibrating an XRF can be tricky and they can drift with time. They most definitly have bias and uncertainties. On my machine, a lower-end model, which is probably similar to the one your assayer had, we figure about +/-0.1 to 0.03% for normal operation by someone who knows what they're doing. A specialist lab with better instrumentation (high power instruments) can do rather better than that.
I wouldn't take it as gospel. Did he give you a certificate of assay or something? That should show bias and uncertaintainty (a +/- number) on the paper. If not, there's really no way of knowing how good his number was.
posted by bonehead at 9:51 AM on May 26, 2011
I wouldn't take it as gospel. Did he give you a certificate of assay or something? That should show bias and uncertaintainty (a +/- number) on the paper. If not, there's really no way of knowing how good his number was.
posted by bonehead at 9:51 AM on May 26, 2011
This thread is closed to new comments.
posted by Green Eyed Monster at 6:49 PM on May 25, 2011