How to Subsample a Large Bulk Sample or Soil, Ore, or Rocks

Before assaying, original samples must be cut down, often in several stages. Each cut involves, in effect, the sampling of a sample, the final portion comprising only the ore that actually goes into the crucible or flask and upon which the actual determinations are made.

In cutting down samples to reduce the bulk to the final amount required for assay, it is important to observe the general principles already mentioned regarding particle size, richness of the ore, and uniformity of mineral distribution in relation to the size of the sample and of the cuts made. Thus, it may be necessary to reduce the particle size at one or more stages to maintain the proper relationships.

The required fineness before cutting at any stage depends on (a) size of sample, (b) distribution, size, and number of particles of ore mineral in the sample, (c) range in value between high- and low-grade ore mineral particles, and (d) percentage of the total weight of the sample comprised in the valuable mineral.

Tables have been devised by various investigators as a guide to the relationships that should be maintained; these have been based both on mathematical calculations and on experimental data.

The following table by Woodbridge, based on extensive experiments and calculations by Brunton, is an example.

smallest-permissible-weight-of-sample

This table was devised as a guide for sampling large lots of medium-grade gold ore, but the principle involved is the same as that for cutting down the kinds of samples under consideration. If the ore minerals are very unevenly distributed in the gangue matrix or the ore is high-grade, finer crushing or larger samples will be required. For more uniformly mineralized base ores, considerably smaller weights would be allowable.

The handling and treatment of drill samples have been touched upon briefly already. Sludge from drill holes usually is fine enough to permit splitting or cutting at the drill, and it has been noted that in churn-drill sampling at some mines the sludge is split in the wet state when it is bailed from the hole, sometimes only ½, ¼, or 1/8 of the total sample being retained.

Face samples often are sacked and sent directly to the assay office without being cut down. When prospects or mines remote from an assay office are sampled, it may be advisable or even necessary to reduce the bulk and weight of the samples before shipment.

Samples may be cut down either by hand or mechanically. Mechanical splitting is preferred for accuracy, but mechanical splitters may not be available and hand-cutting may be the only alternative. Large samples usually are first cut down by “quartering,” in which process the ore should first be spread on a clean, smooth floor and pieces larger than allowable should be broken.

The sample is then shoveled into a conical pile, each shovelful being dropped on the tip of the cone. The fines will tend to remain near the tip or slide part way down the sides of the cone, whereas the larger lumps will roll to the bottom. Thus, the fines will tend to become more or less segregated from the lumps, which is not conducive to thorough mixing. In practice, the tip of the cone will tend to move in one direction, so that when the cone is completed the tip will be off center. If this occurs, there will obviously be a concentration of lumps on one side; when the cone is flattened and quartered, some quarters will have too large a proportion of lumps; and as the lumps often contain less or more of the valuable mineral than the fines, the quarters retained in the sample will not represent the average grade.

Some samplers employ an upright iron rod to mark the center of the cone and thus avoid “drawing” the tip to one side. Others employ a wooden or steel cross having arms of equal length, which is laid on the floor, and each shovelful is then dropped over the intersection of the arms.

After all the ore has been swept up and coned, the next step is to flatten the cone into a disk by dragging it out spirally with a shovel or flat blade, beginning at the tip of the cone and taking care to spread

method-of-splitting-samples

the material uniformly in all directions. The thickness of the finished disk should be about one-tenth its diameter, and, if flattened properly, it will be circular. It may be necessary to recone and flatten one or more times (three conings is common practice) to mix the sample thoroughly before quartering.

The final disk is then divided into quadrants by marking with a straight-edge or a right-angle cross, and two opposite quadrants are removed and rejected. The two remaining quadrants constitute the sample. This may be reconed and requartered in the same manner as the original sample if still of sufficient bulk in proportion to the size of the particles, or it may be first crushed or ground to smaller size before quartering or splitting it further.

First cuts of original samples may be made also with such devices as the Brunton quarter shovel (fig. 9, A) or a split shovel. In splitting a sample with the Brunton shovel, the shovel is forced into the sample pile until all compartments are filled. The shovel is then removed and tilted backward to discharge the material in the two outside compartments, which is the reject. The portion in the middle compartment is emptied into the sample box. The operation is continued until the entire sample has been handled. The split shovel is constructed with alternate openings and prongs, the latter forming narrow boxes, the area of which equals that of the openings between them. The shovel is held over a box or other receptacle and an assistant shovels the sample over it. Half the material is supposed to pass through the openings into the box, the half remaining on the shovel being rejected.

A bank or combination riffle sampler like that shown in figure 9, B, can be constructed easily, and by shoveling the sample into the top, a 1/8-sample is cut. Figure 9, C, illustrates the Jones splitter, which is usually used for the last splitting after the sample has been ground to pass a 1/8-inch or finer screen. The final split for the portion that actually goes into the assay is made upon finely pulverized material that has been mixed on a sample cloth. The mixed sample is spread into a thin cake, and the portion for assay consists of dabs taken from points well distributed over the cake.

Automatic samplers of the revolving type, which cut small portions from a stream of ore at regular intervals, are employed extensively for sampling in ore-dressing plants. Samplers of this type give quite accurate results, provided they are kept in good working order and they are watched to see that they do not become clogged and that the sample receptacle does not fill and run over.

Whatever the system of cutting-down employed, the danger of error caused by inclusion of disproportionate amounts of valuable mineral particles, either in the reject or in the sample, increases as the size of the sample becomes smaller. By way of illustration, assume a sample cut in which only one particle of valuable mineral remains. If the particle goes into the reject during cutting-down, the final sample will have a zero value, whereas, if it goes into the final sample for assay, the result will be 2, 4, 8, or more times the correct value. Now if this particle is broken up into many smaller pieces by grinding the sample, the likelihood of obtaining a proportionate amount in the final sample is much greater. It therefore follows that the finer the particles comprising the sample at each stage of splitting, the better is the chance that the final portion used for assay will be truly representative.

In sampling gold ores it is customary to reject any visible gold. Particles of free gold are not pulverized during crushing and grinding but merely have their shape changed. Therefore, coarse particles of gold in the original sample may carry over into the final sample and give an assay value many times the actual value of the ore.