It needs no preparation of the pipe end at all.

A threaded joint needs a thread cut. A flanged joint needs a flange welded or screwed on. A grooved joint needs a groove rolled or cut. A welded joint needs hot work, access all round and a qualified welder. Every one of those is difficult or impossible on a pipe already in service, in a trench, or made of a material you cannot weld.

A bolted sleeve seals against the plain outside surface of the pipe. Cut the pipe square, clean the outside, slide the sleeve over and tighten the bolts; the gaskets are compressed radially onto the pipe OD.

That makes it the default for repairs, tie-ins to existing mains, and any connection where the pipe cannot be taken away and prepared. The trade-off is that it takes up more space than a conventional joint and it does not resist axial pull unless it is a restrained type.

Yes, within a published range, and it is one of the main reasons they are specified.

Because the seal is a compressed gasket bearing on the outside of the pipe rather than a machined fit, a coupling can tolerate a spread of outside diameters. Wide-tolerance couplings accommodate the variation between nominally identical pipes in different materials - cast iron, ductile iron, steel, PVC and asbestos cement of the same nominal bore all have different outside diameters. Stepped or transition couplings join two genuinely different ODs by using different gasket sizes at each end.

This matters enormously on older networks, where a repair frequently means connecting new pipe to something laid decades ago in a material no longer used.

Measure the actual outside diameter of both pipes rather than working from nominal bore, and check both fall inside the coupling's stated range. Nominal size is not enough information to select one.

A plain sleeve coupling does not, and assuming otherwise is the most common way they fail.

The coupling seals against pressure by compressing gaskets radially onto the pipe. That grips well enough to hold the seal but is not intended to stop the pipes being pulled apart along their axis. Wherever the pipeline generates axial thrust - at a bend, tee, taper, valve or closed end - an unrestrained coupling nearby can be pushed off the pipe.

There are two answers. Use a restrained or anchor pattern coupling, which incorporates gripping rings that bite into the pipe surface and carry the tension. Or provide external restraint - a concrete thrust block, or tie bars across the fitting.

Calculate the thrust from the pipe diameter and the test pressure, not the working pressure. And remember a straight run is not automatically thrust-free: a closed end during commissioning creates full thrust where there was none in the design.

It is closed over the defect and bolted down so the sleeve, rather than the pipe wall, contains the pressure at that point.

The procedure is to expose and clean the pipe well beyond the damage, check that the surrounding pipe wall is sound enough to seal against - a repair clamp needs good metal either side of the defect - and confirm the pipe OD against the clamp range. The clamp is then positioned centrally over the defect and the bolts tightened evenly and progressively, working around the clamp rather than fully tightening one bolt at a time.

Some patterns can be fitted to a live main; others require the line to be drained. Check which before starting, and never assume.

Treat the result honestly. A repair clamp on a corroded main is a sound engineering repair for a local defect, but where corrosion is general it is a holding measure and the section should be programmed for replacement.

Match it to the fluid, the temperature and, for potable water, the approvals regime.

Nitrile rubber is common for water, air and hydrocarbon service and is often the default. EPDM suits hot water, steam and many chemicals but is attacked by petroleum products, so it must not be used on fuel or oil lines. Fluoroelastomers are specified for aggressive chemicals and high temperatures.

For potable water the gasket must carry the drinking-water approval that applies in the market the pipeline serves. Approvals differ by country and an approval in one market does not automatically transfer to another, so this needs checking against the local requirement rather than assumed from a manufacturer's general statement.

Get this wrong and the failure is slow and confusing: the gasket swells or hardens over months, the joint weeps, and the coupling is blamed for a material selection error.

A small amount at each joint, which is useful for following a trench but is not a substitute for a bend.

Because the seal is made by a compressed gasket rather than a rigid interface, the pipes can sit at a slight angle to one another and the joint still seals. Manufacturers publish the permitted deflection, typically a few degrees per joint, and it lets a pipeline follow a gentle curve or absorb some ground settlement without fittings.

Two cautions. Deflection reduces the coupling's effective pressure rating in some patterns, so check whether the published pressure applies at full deflection. And deflection consumed to correct poor alignment at installation is no longer available to absorb future ground movement, which is when it is actually needed.

Align the pipes properly and treat the deflection allowance as reserve, not as a way of avoiding a fitting.

Even, progressive, alternating around the coupling, to the manufacturer's stated torque - and never to the point where the bolts are simply as tight as they will go.

The gaskets seal by being compressed a controlled amount. Uneven tightening pinches the gasket on one side and leaves it loose on the other, which is the usual reason a coupling weeps after installation. Overtightening extrudes the gasket out of its seat and can distort a thin-walled or plastic pipe, so the joint gets worse the harder it is pulled.

Work around the coupling in stages, taking each bolt part way and coming back, as with a flange. Use a torque wrench and the published figure. Check that the gap between the coupling halves stays even all round - it is the simplest visual indication that compression is uniform.

Re-check the torque after pressurising, and again after the first thermal cycle on hot service.