Sella's separator

Sentences without a label are documented facts; the others are marked "probable" or "uncertain".

Between 1854 and 1860 a machine that separates minerals with electricity is built at Traversella. The idea is Quintino Sella's, the machines are built by Froment in Paris, and in 1860 Burci describes them while they are working, with scale drawings. This page explains the problem, the machine, the physics that makes it work and its figures.

The copper problem

The first documented search for copper at Traversella dates from 1741, when the counts of Castellamonte declare that they have discovered a copper mine at the Riondello; the chalcopyrite of Traversella usually has less than 3% copper. The Castiglione mines are abandoned at the end of the 18th century and reopened around 1825; there, according to Novarese as reported by Colomba, the magnetite occurs with chalcopyrite and pyrrhotite in a serpentinous gangue.

In 1854 Count Riccardi di Netro, one of the owners, asks for a study to recover the copper dispersed in the magnetite. In the waste heaps the copper is still there, but it is mixed with the iron ore and picking it out by hand is too difficult probable.

Sources: Berattino 1988, p. 243; Colomba 1915, pp. 3-4; Jervis 1873, p. 124; Burci 1860, pp. 8-10; Costa et al. 2019, p. 12

Three machines, not one

Burci tells of three machines. Towards the end of 1854 Sella tries out a small machine: a wooden disc with a single temporary magnet, that is, an electromagnet. On 19 July 1855 Sella obtains in Piedmont a certificate of privativa, the forerunner of the patent, and commissions from Froment, in Paris, a machine with 48 electromagnets, which is tested for more than a year and exhibited at the Piedmontese industrial exhibition of 1858. The third is the industrial machine, also built by Froment, which in 1860 is working in Count Riccardi's works. Berattino confirms the sequence: commission in 1854, patent in 1855, machine ready and exhibited in Turin in 1858.

Costa and colleagues describe the machine as a cylinder with 48 "fixed magnets" and give the privativa a duration of fifteen years probable. Burci, who writes while the machines are working, speaks instead of temporary magnets already for the first trial: the portal follows Burci. In the 2022 chronology Costa dates the introduction of Sella's "electromagnetic sorter" to 1855-56.

Sources: Burci 1860, pp. 8-11, 13; Berattino 1988, p. 94; Costa et al. 2019, p. 12; Costa, Garabello 2022, slides 25-26

How the industrial machine is made

The industrial machine has a wheel made of three cast-iron rings on the same axle, with three staggered series of electromagnets. Under the wheel runs an endless canvas belt; above the belt a hopper, regulated by a handle and a gate, spreads out the ore. The commutator, invented by Sella and made in Turin, gives current only to the 11 magnets that are above the belt at that moment.

Each machine is powered by 6 large Bunsen cells and driven by a water wheel. The magnet wheel turns at 14-15 revolutions a minute and the belt moves forward at 0.75 to 1.35 metres a minute.

Sources: Burci 1860, pp. 13-16, pl. II

The physics of separation

An electromagnet is a coil of wire wound around an iron core: when current flows in the wire the core becomes magnetised, when the current is interrupted the field almost disappears. Magnetite, Fe₃O₄, the most widespread mineral at Traversella, is strongly attracted by a magnetic field; chalcopyrite, CuFeS₂, very little. Passing under the switched-on magnets the magnetite lifts off the belt and stays attached to the wheel, while the chalcopyrite carries on.

The commutator is the most ingenious part. By switching on only the magnets above the belt, it makes the magnetite get picked up there and dropped a little further on, when its magnet switches off: the iron ends up on one side and the copper stays on the belt and falls off on the other. With permanent magnets the magnetite would stay attached to the wheel. At Castiglione, in 1873, Jervis calls the machine the "Sella electro-sorter".

Sources: Burci 1860, pp. 13-15; Costa et al. 2019, p. 14; Jervis 1873, p. 125

Work, yields and costs

In Count Riccardi's works, Burci writes in 1860, the machines work day and night for almost the whole year and each one processes about 180 kilos of ore an hour. Every 8 hours the workers change and the battery is renewed; how many workers there were and who they were Burci does not say. The method pays only with ore that has at least 2.5% copper and little pyrite.

In 1873 Count Riccardi is the sole owner of Castiglione, with a concession of 1859, and works those mines only for copper. Jervis writes that with the electro-sorter the result is still a loss because of the costs of the apparatus, and that the separated magnetite is thrown away.

Sources: Burci 1860, pp. 25-27; Jervis 1873, pp. 124-125

After Sella

With Sella's sorter and with the washing plant, after the middle of the 19th century, the "giavinatura" disappears, the search for ore in the waste in front of the diggings carried out by the poorest. In the 20th century the principle returns. The engineer Maglione, who runs the mines from 1919 to 1933, works out from the registers of the old electromagnetic works yields of 60% for iron and 25% for copper; in 1923, with more electromagnets, about 80% is reached. In 1937 FIAT gives the Società Nazionale Cogne the old waste heaps, about 300,000 tonnes, with the magnetic enrichment plants put back into operation.

Meanwhile research takes another road. After 1900 Froment tries to separate the copper sulphides from the magnetite with a thick oil and obtains a froth that carries the sulphides to the top: this is flotation. In 1904 the company, in difficulty, sells the patent for 225 pounds to Minerals Separation Ltd, which exploits it all over the world; Torti, without giving his source, writes 25 pounds probable. In the mid-20th century FIAT installs new magnetic separators at Traversella; in the silo built around 1940, after the crushers and mills, the first step is the separation of the magnetite probable.

Sources: Berattino 1988, pp. 197-199, 241; Beratto (thesis), p. 30; Costa et al. 2019, p. 12; Torti 1973, p. 8; Museo della Miniera website, «Il Museo»

What remains to be cleared up

We do not know where the works stood: Burci does not say, and Berattino places at Le Folle the copper foundry that Countess Augusta Gattino, Riccardi's wife, was given permission to build in 1858; that the machines too stood there is only a hypothesis uncertain. Berattino publishes the photograph of a Sella "magnetic sorter", "year 1855", owned by the Politecnico di Torino, with hopper, electromagnet drum, pulley and flywheel; Beratto writes that the sorter "now" belongs to the Politecnico, without saying which of the machines it is or where probable. Vescovo writes that the machine was "later adopted all over the world", but without sources and without any other confirmation uncertain.

Sella was not only the inventor of the machine: according to Chiappino, as an engineer of the Royal Corps of Mines, he worked on the law of 20 November 1859, which declared the subsoil the property of the Nation probable.

Sources: Burci 1860, p. 25; Berattino 1988, pp. 93-94, fig. 8; Beratto (thesis), p. 28; Vescovo 2002, pp. 22-23; Chiappino 1996, p. 45; Chiappino 2010, p. 25

For the class

A calculation to try. The belt moves forward at 0.75 to 1.35 metres a minute and each machine processes about 180 kilos of ore an hour. How many metres of belt pass under the wheel in an hour? How many kilos of ore does each metre of belt carry, on average, in the two extreme cases? Solution: from 45 to 81 metres an hour, so from about 4 to about 2.2 kilos per metre. Second question: the wheel turns at 14-15 revolutions a minute; what would you need to know to calculate its peripheral speed and compare it with that of the belt?

Sources: Burci 1860, pp. 16, 25