I have chosen this subject in view of the fact that, while there are many and ample opportunities of getting reliable assays made, there are but few who are in a position to obtain reliable data on the sampling of minerals and metals; my long experience, where, for five years, I had charge of the Refining and Sampling Mill, showed me how important it was that a correct sample should be taken.
This, to the ordinary layman, seems so simple that no expert knowledge is needed, but in actual practice there are hundreds of pounds to be lost by the lack of knowledge in this department.
- There is the matter of weighing in large or small quantities.
- The class of weighing machine used.
- The accuracy of weights used.
1. We all know that, say, 1000 tons of ore cannot be weighed in one parcel; but the fewer the weighings the less loss to the seller (and he usually wishes it weighed in small lots carefully) for it is a well-known fact that the most sensitive scales take a certain weight to show distinctly the indication of the balance, or, in other words, what is known as “ the pull of the scales.” As for instance, 1000 tons weighed in lots of 1 cwt. each, taking the pull of the scale as being equal to 1 lb., would give a result of 9 tons loss to the seller, and as most of the ore sent to Customs works is worth 5 ozs.
The only absolutely reliable machine to be used is the beam scales, for the reason that the parts which can be tampered with are, in this class of scale, open to the view of the seller during the whole course of weighing, and any loss of sensitiveness is readily detected as each weighing is made. There are objections raised to these scales on the plea of their being slow in weighing, and the impracticability of weighing trucks or barrow loads of ore; but in my experience in weighing bagged ore, they have been proved quite as quick as platform or any other scales. While in the case of weighing in barrows or trucks, it is only necessary to detach the chains and pan from one end of the beam, and attach three chains, two of which have rings for the handles, and one a hook for the wheel, and thus make the barrow itself the pan of the scales.
In the matter of the weights used, it is necessary to have them carefully checked and marked before using, as cast-iron weights (the sort generally used) are liable to lighten by small pieces chipping off.
Measure the Moisture of a Sample
In most cases the moisture sample is taken after the ore is crushed, and it is very obvious that in a hot country like Australia, there must be a considerable loss of moisture during the operation of crushing and quartering, it is, therefore, apparent that to get correct moisture assays the sample must be taken while the ore is being weighed, and the sample taken to be subjected to as little handling as possible. After many experiments I found that the highest moisture assays were obtained by taking one shovelful of ore from each barrow as it had passed over the scales, this sample was thrown into a large covered iron flat tray, capable of holding, at least, one cwt. of ore, kept shaded from the sun adjacent to the scales ; at the completion of the weighing, the whole of the moisture sample is carefully weighed and covered with a perforated cover to prevent both the dust and the loss of ore, and at the same time to allow for the evaporation of the moisture. The sample is then placed in a drying oven (which should be provided with lock and key) for 24 hours, then taken, cooled and reweighed ; the gain in moisture by adopting this method, in comparison to the system generally in vogue, is ten per cent, in favour of the buyer.
General Sample of Ore
There are various methods of taking the sample, the most generally used of which is that of taking either each 4th up to 10th shovel or barrowful and crush or mix a quarter down to the required size for assay, or to take the whole parcel, crush or mix, and then quarter. In order to arrive at the comparative accuracy of these methods, I took a parcel of 300 tons of 5 oz. sulpho-telluride bagged ore from Kalgoorlie, and divided it into ten lots of 30 tons each ; from each of these 30 ton parcels I took every 10th bag, these I crushed, quartered and assayed. I then took the whole 30 tons, including the 10th bags, and sampled as a whole, with the following results : The assays from the 10th bag sample varied from 4 oz. 18 dwt. to 5 ozs. 12 dwt., an average of 5 ozs. 5 dwt.; whilst the greatest variation in the 30 ton parcel, as a whole, was 3 dwt., and an average of 5 oz. 3 dwt., whilst the difference in the general average amounts to more. The individual differences are much too great to allow of large Custom works adopting any but the method of crushing and quartering the whole sample.
In dealing with ores containing free gold, the sample, after having been quartered down to about 3 cwt., should be carefully weighed previous to passing through an 1/8th mesh screen, so that in the event of free gold being found on the screen a computation can be made of the amount of free gold in the parcel. This likewise applies to the finer screens in a greater degree, and every care should be taken that even the very finest pieces of metallic gold should be carefully weighed and estimated.
The method adopted by the Broken Hill Proprietary Company, at Port Pirie (the largest treatment works in the world) is to divide the consignment into 30 ton lots, weigh and crush with a Gates’ crusher, then pass it through rolls, then through a 3/8th screen, which discharges into a bin. Any ore after coming through the rolls being too large to go through the screen is discharged into the boot of an elevator, and returned to the rolls, until every article of ore passes through the screen. The ore from the bin is received in barrows by workmen, and every fourth barrowload is dumped on to a brick sample floor—the other three going direct to the smelters. This sample, of about 7 tons, is twice piled into the ordinary cone, and then quartered ; piled again and quartered until reduced to about three cwt. This portion is weighed and broken down by hand on a large cast-iron plate, and passed through an 1/8th mesh screen and quartered down to about 5 lb. This is carefully weighed and passed through a 120 mesh screen (that is about 14,400 holes and 14,400 pieces of wire to the square inch); then divided into three parts, packed and sealed in the ordinary way. The reason for adopting hand quartering in preference to the very excellent automatic machines, was the great difficulty and loss of time experienced in cleaning the sampler after each parcel, and an expressed wish by representatives of the sellers that they should be able to see the whole operation, and not lose sight of the sample once it was cut from the parcel. In the sampling of millions of tons of their own even grade ore the automatic sampler was always used. The reason for passing the final parcel through such an extremely fine screen was that the agreed assays had to be within 2 dwt. of one another, and the coarser mesh sieve made this impossible on high grade ores.
A very able and exhaustive paper showing how different are the gold values throughout a bar of lead containing gold, and that sampling by either boring, chipping, or sawing is entirely unreliable. To overcome this, BHP adopt a method of melting the whole of the parcel in a cast-iron kettle, heating up to redness, and then liquefying the lead mechanically held in the dross formed by repeated additions of sawdust to the top of the molten metal, then rapidly skimming, the dross with perforated ladles and granulating whilst red hot. By this means the ordinary differences in dross samples are entirely obviated, the variations only being those found in the coarse assaying. The skimmed metal is thoroughly, stirred, and a ladle of metal about 20 lb. dipped out and rapidly poured (to stop any separation that may occur) into a series of 10 bullet moulds made of such a size that each ballet, weighs one gramme assay ton, and the whole of the bullet is taken for assay. By this means any separation that may take place in the bullet itself will not affect the assay. Although this method is apparently costly, working experience, of it showed it to be as cheap as boring or sawing.
The difference exists equally in a bar of gold, and more so if alloyed with baser metals, and an excellent method of taking a dip sample while the metal is molten is to attach a fluxing crucible to a thin rod of iron and dip the sample and granulate in distilled water. The use of iron ladles is objectionable, in that small flakes of iron get mixed, with and attached to, the sample.