Dear Georgia,

I do remember ISO/TC183's meeting in Manila and our workshop at the PASAR smelter. Does PASAR already select interleaving samples (A- and B-samples in ISO vernacular) to estimate the risk associated with buying copper concentrate?

Many years ago I examined the statistics for trace elements in cathode copper and decided that reliable confidence limits can only be obtained by installing a mechanical sampling module between the strip and strap sections. The module was designed to punch a 20 mm disc from each pair of cathodes after stripping. Punch positions were randomly selected from a matrix of 25 by 25 cells. This is an example of a stratified random sampling protocol where each pair of cathodes is a stratum and each 20 mm disc is a random sample. The set of 640 discs should be alternated between containers such that a pair of interleaving samples is routinely obtained.

This statistics improved significantly and the complaints dropped dramatically. After all, every cathode in a shipment sport a single hole so that the buyer is aware the same sampling protocol cannot possibly be applied at affordable cost.

Here's a cautionary tale! When I visisted the copper refinery where the first sampling module was installed again a few years later, a new metallurgist wanted me to look at it. I told him I knew how it worked but he insisted. So I watched it position the punch, saw the disc drop in a bucket and said it looked quite alright. But when I saw the next disc drop in the same bucket, I asked him whatever happened to interleaving samples. He told me to talk to the manager who said, "I'm comfortable the numbers!" I wonder how anybody could possibly be comfortable when confidence limits for trace elements in cathode copper can no longer be reported! ■