The Troughed Belt: Two Tipped Rollers That Doubled What a Conveyor Carries

Lay a strip of belting, or any stiff sheet, flat on the bench and pour dry sand onto it in a line. The sand spreads sideways until its surface reaches its angle of repose, and then it stops. Measure the width of the heap and its height at the centre. Notice two things. The heap has spread well outside where you poured it — so on a moving belt this is the spillage line, and it is why a flat belt needs skirts. And the cross-section is a flat triangle: almost all of the belt width is under material only a few millimetres deep. Now lift the two edges of the strip so it folds into a shallow trough, and pour again. The material stays where you put it and stacks. That is the entire invention, and you have just done it by hand. The angle the loose surface stands at is worth measuring while you are here. Sand is not stone and dry is not damp; record what you actually have, because step 4 needs it.

One centre roller horizontal, two side rollers tipped up at the same angle, all three in the same vertical plane and their working faces meeting end to end. Cut a frame from angle iron. Set the centre roll first and get it level with a spirit level — everything else is measured from it. Then set the two side rolls: mark the trough angle with a protractor, and check that the gap at each junction is the same on both sides. An idler whose side rolls are at different angles steers the belt, permanently, and no amount of adjustment further down the run will fix it. The rollers themselves must turn freely and must be **concentric**. A roller with a flat spot or an out-of-true shell hammers the belt once per revolution; over a shift that is hundreds of thousands of hammer blows into the same carcass. Make the stand square to the belt line. The commonest tracking fault on a real conveyor is not a worn belt, it is an idler stand that has been knocked out of square by a loader.

Robins' sheet 2 carries two figures. One is the troughing idler. The other is a pulley whose diameter is slightly larger in the middle than at the ends — a **crown**. A belt climbs toward the tight side. On a crowned pulley the middle is the tight side, because the belt has further to go over the bulge, so the belt is continuously steered back to the centre. It is self-correcting rather than adjusted. Try it. Turn a short wooden roller with a taper each way — a millimetre of crown over a couple of hundred millimetres is plenty — and run a strip of cloth over it. Start the strip deliberately off to one side and turn the roller: the strip walks back to the middle. Turn a second roller dead cylindrical and repeat: the strip walks off the end and keeps going. Crowning is for the *head and tail pulleys*, not for the carrying idlers. Crowned carrying rollers under a troughed belt fight the trough and wear the belt's underside. One caution the old texts all give: crowning steers a belt, it does not fix a misalignment. If a belt runs off on a crowned pulley, the frame is out of square and the crown has run out of authority.

Two loading faults destroy belts, and both are about where the material arrives rather than how much of it there is. **Off-centre loading** puts the whole weight on one side of the trough. The belt cambers, runs off, and rubs a structure until it is cut. Centre the chute on the belt and check it loaded, not empty. **Loading against the belt** is worse. If the material lands with a sideways or backward velocity component, every piece has to be accelerated up to belt speed by scrubbing against the rubber. Turn the chute so the material arrives already moving in the belt's direction and at something near its speed, and that scrubbing disappears. Run the belt with the load and watch the material at the loading point. If you can see the stone skidding, the chute is wrong, and the top cover is being sanded off at that one spot. Guard the nip points before you run anything. The gap where the belt meets a pulley or an idler will take a hand in and will not let go, and a conveyor cannot be stopped fast enough to matter. Pull-cord stops along the length, fixed guards at every pulley, and nobody reaches over a running belt.

**US 571,604, 'Conveyer Belt Apparatus', Thomas Robins, Jr., granted 17 November 1896.** Robins was working on the problem of moving coal and ore in bulk, where flat belts already existed and were disappointing. The patent's sheet 2 shows both halves of the answer: the three-roll troughing idler and the crowned pulley. Everything on a modern overland conveyor is an elaboration of those two figures. The consequence was out of proportion to the mechanism. Bulk material had been moved by bucket elevator, by wagon and by hand; a troughed belt moved it continuously, at low power per tonne, over distances no other method could reach. The mines, the ports and eventually the quarries reorganised themselves around it. **Honest limits.** A belt will not climb steeply — about eighteen to twenty degrees for loose aggregate before the load slides back, which is the material's own limit and not the belt's. It cannot turn a corner; it needs a transfer, and every transfer is a wear point, a dust point and a spillage point. It is a bad choice for anything very hot, very sharp or very sticky. And it is the single most dangerous machine on most sites, because it is quiet, it is long, it has nip points every few metres and it looks harmless.