Why the belt gets blamed

Belt tracking is a maintenance problem in most people’s minds: adjust the rollers, tension the belt, and if it persists, replace the belt. Those steps work often enough to reinforce the belief, and they work because they compensate for a geometry error rather than correcting it.

The mechanical rule is simple. A belt moves toward the end of a roller that it contacts first. If two rollers are not parallel, one end of the belt reaches its roller slightly before the other, and the belt steers. Every tracking adjustment is a way of introducing a deliberate, opposite misalignment to cancel an unintended one.

That is why adjusted conveyors drift again. The compensation is set for one load, one temperature and one belt tension. Change any of them — or let the frame deflect a little more under a heavier load — and the cancellation no longer balances.

Three geometric conditions, all set during build

An aluminium conveyor frame has to satisfy three conditions that are independent of each other, and all three are established before the belt is ever fitted.

The first is parallelism between the two side frames. If the frames converge or diverge along the length, the rollers mounted across them cannot be parallel, and no amount of adjustment makes the belt run true over the whole length — only at the point where it was adjusted.

The second is level across the width. A frame that is twisted, with one side higher than the other, puts the belt on a slight slope, and gravity does the rest. Twist is the condition most often missed because it is invisible in a plan view and because a frame can be perfectly level at both ends and still be twisted in the middle.

The third is the span between legs. Under load the frame sags, and sag changes the relative position of the rollers along the length. A conveyor that is true when empty and drifts when loaded is telling you the span is too long or the section too light for the load.

Accumulated tolerance: the long-conveyor problem

A short conveyor is easy to build true. A long one is built from sections, and every joint contributes a small error. Individually those errors are inside tolerance; added together along twenty metres they are not.

The mechanism is straightforward. Each frame joint has a positional tolerance, and each is assembled to the previous one rather than to a common datum. A tenth of a millimetre at each of fifty joints is five millimetres of drift at the far end, and five millimetres is more than enough to make a belt wander visibly.

The fix is a datum. Build the conveyor to a measured line rather than joint to joint, so that each section is corrected against a reference instead of inheriting the error of the one before it. On a long run that single change is usually worth more than any upgrade to the section.

Measuring it without a laboratory

Three measurements settle whether a conveyor frame is true, and none of them needs specialist equipment beyond a line and a level.

Parallelism is checked with a taut line or a laser along each side frame, measuring the offset at intervals. The frame’s width should be constant along its whole length; where it is not, the frames are not parallel. Level across the width is checked with a spirit level or a straightedge and level at several points along the length, not just at the ends, which is where twist shows up.

Deflection is measured by recording the frame’s position empty and then loaded at the mid-span. If the mid-span drops enough to change the roller geometry, the span or the section needs attention. The threshold depends on what the conveyor handles and what the process tolerates, so the useful step is to state the limit on the drawing rather than to look up a number.

What to send with a conveyor frame enquiry

A conveyor frame enquiry usually arrives with a length, a width and a belt type. The items that decide whether it will track are the ones about load, support and how the frame will be built.