Factotum # 37 in a series on how do we find the gold in this here hill structure?
1, Our assay method is bad. Its recovery is 30% or worse. We have to crush the rock by a hammer and pan. We weigh the gold we get by a scale to the nearest 1/20 of a gram. this requires huge 50 pound sample sizes. As there is only 0.75 grams in our sample and we get 30% of that for a measly 0.25 grams. There is no way of testing this because we can only guess at it. If we had a perfect assay method to test, we would use that.
2. We are concerened about our sampling method being representative but the only thing we can do is cut across the known gold precipitative structures to get as large a sample of them as possible. This requires channelling across the vein roughly sub parallel to the veinlets which cross the main quartz structure at 35 degrees to the strike. (D. Hume -- northwestern Ontario gold deposit) We do a channel every 7 feet and pray it works.
3,. Our mill has an unknown recovery too. It may be 40% and it may be 80%. Can we tell? If we run the ore we have sampled we only get to know there is a discrepancy between our assays of it, and the mill's recovery. Which is worse?
4. We do the same assaying of the mill heads and tails, and cons taking 50 pound samples every 3 tons, by collecting them carefully in small increments as unbiased a way as possible. We can only know that whatever the loss rate of the mill we have to increase the size of the tails sample to 250 pounds, as it may be that the mill takes 80%. In order to get the same accuracy we have to get 5 times the material. With all these things being the same, we now know the loss rate of our mill, as the accuracy of the tails assay has to be the same as the heads, so the relative losses are constant. With C-T/F-T we know the ratio of concentration. We can say we get 90% of the gold by smelting our concentrate, so we know the amount that went to tails. Smelting here is assumed to be our friend. We have to have friends somewhere.
The last variable is the variable of our sampling versus the mill recovery. We now know that and the nugget effect or sampling variability can be calculated.
By pan.
In fact the pan is BETTER than fire assay of smaller samples. Why? Because it eliminates variability of the sampling size. We take 800 times more material, and the constant loss rate of the assay method and its crudity are more than made up for by bulkosity of the sample.
That is why in 1975 they panned the tails of the Canadaka Silver mill flotation cells, and visually inspected the result to determine the efficiency of their flotation operation for silver. I believe the silver flotation method at Cobalt was the first such one employed and figured highly in their mining economics.
EC<:-} |
