By Charles Kubach, Mine-Engineer.Com
August 23, 2014
Sampling is defined as taking a small portion of a whole mass that accurately represents the whole mass.
With mining reserves trending to lower grades, the need for accurate, precise and representative sampling of exploration drilling and production blast holes has become more critical. It has not been too long ago where mines sampled every 5th or so holes and assayed those samples. Grades were good, mining cost much less and 10 percent error one way or the other was not so critical. However nowadays, as ore bodies get leaner, and mines more expensive to build and operate, it is more critical to obtain representative, accurate samples for assay, in order to properly plan and execute the mine plan, and keep a adequate volume of product going out with a adequate amount of capital coming in.
Some negative factors affecting sampling of drill holes are as follows:
(a) Loss of sample (cuttings) in the hole, or during collection, resulting in a somewhat unrepresentative sample.
(b) Sample preparation, where samples are improperly crushed, pulverized and split.
(c) Using poor scientific analytical techniques.
As the Founder of Modern Sampling Theory, Pierre Gy once said, "errors induced by faulty sampling equipment, techniques and poor sample preparation and handling can drastically induce errors into the results of a sampling program, rendering even a good sample plan somewhat inaccurate, and not representative of the whole". These errors can range from a few percent to fifteen or twenty percent in some cases. This has been proven many times, and is written about in numerous papers and books, so it is not a theory, it is a fact.
One factor noticed twenty years ago was that traditional core drilling produced less accurate samples than reverse circulation drilling. This was found to be simply that reverse circulation drilling produced many chips of a half inch or so, and the cores produced one large column of rock, which was often cut in half, one half was analyzed and the other half saved for perhaps a off site assay. Unfortunately, the valuable elements were always not evenly distributed on the core so the halves had differing percents of values. With the multitude of small particles from reverse circulation drilling, a more representative sample was obtained, since basic sampling theory has always held that the smaller the particles, the less mass was required for accurate sampling. And as the few hundred pound sample obtained from drilling, could represent 30,000 to 50.,000 tons of ore, this was a very important factor in obtaining more representative and accurate samples.
So now that three major problems associated with drilling ore bodies and collecting representative samples have been pointed out, what can be done to correct them?
First, regarding loss of sample cuttings or inadequate/inaccurate cut of sample sent for analysis, there are several things that can be done to produce representative samples. First, faulty sampling equipment is the cause of much of the error, and this can be the method the sampler collects the sample, or its design, which does not allow the cut sample to flow freely, and possibly leaves quantities of material in the sampler from previous drill holes to cross contaminate future samples. This is easily solved, by design, where the sampler is functional for the type of drill utilized. Of course, it does require a bit of engineering and some "hands on" testing to fine tune the sampler, but the result will be representative samples "in the bag".
Next, the sample preparation errors can likewise be resolved by utilizing a good sample prep plan, adequate training and suitable equipment. For instance instead of riffle splitters that have a 3-4 percent margin of error use a rotary sample splitter, which has a margin of error of less than 1 percent. Ring and puck mills are great for quickly pulverizing samples of up to a kilogram (2.2 pounds) in 3 minutes, and virtually no loss is experienced, because the grinding container is sealed with a o-ring.
And finally, the easiest factor to correct, using scientific analytical techniques. There must be a thousand peer reviewed papers and even more books devoted to this subject and most are fairly applicable to sampling of ore bodies. However, ore bodies, as any geologist will tell you, vary greatly, so some human interpretation is required to "adapt" any such theory or procedure to a particular site. Some input from the geological staff as to the character of the particular ore body would be helpful in determining the randomness of the values in a particular ore body, combined with some practical experience could provide a well suited sample plan with associated techniques to achieve the plan results of representative samples for analysis for each drill hole.
The author found the paper, Statistical Analysis of Blasthole Sampling Data From Paradise Peak Mine, Nevada by W. D. Delhaut, et.al., to be most helpful in writing this article. (SME Publication)
Return To Sample Index