Busbar bolt torque and Belleville washers

Clamp force is the target

A busbar joint carries current through the contact area between two surfaces. The resistance of that contact falls as the clamp force rises, because the real contact area grows and the surface films break down. The bolt does not conduct the current in a well designed joint. Its job is to hold the surfaces together, and the force it produces is the number that matters.

Torque is only a proxy for that force. When a nut is tightened, most of the torque goes into friction under the head and in the thread, and only a fraction becomes preload. The split depends on the lubricant, the plating, the thread fit and the surface finish. Two joints tightened to the same torque can differ in clamp force by a wide margin if one is dry and the other is waxed.

The design should therefore start with the clamp force the joint needs for its current and its temperature rise. The bolt size, the number of bolts and the torque figure follow from that. A drawing that gives only a torque value leaves the supplier and the assembler guessing at the force.

From torque to clamp force

The common estimate is torque equals the nut factor times the clamp force times the nominal diameter. The nut factor is dimensionless and covers the friction in the thread and under the head. A dry steel bolt might use a factor near 0.20, a lubricated bolt near 0.16, and a waxed or coated bolt lower still. These are illustrative values, not constants.

For illustration, an M8 bolt at 20 newton metres with a nut factor of 0.20 gives a clamp force of about 12.5 kilonewtons. The same bolt at the same torque with a nut factor of 0.16 gives about 15.6 kilonewtons. The joint sees a quarter more force from a change in lubrication alone. That is why the assembly instruction has to say whether the thread is dry, waxed or lubricated.

The estimate has limits. It assumes a clean thread, a flat seat and a bolt that is not yielding. It does not account for the stiffness of the clamped stack or for the friction between the busbar faces. For a critical joint, a torque and angle procedure or a direct load measurement is more reliable than a torque figure alone.

Relaxation in a copper joint

A bolted joint does not stay at its initial clamp force. The surfaces bed in, the coating creeps and the copper relaxes under the sustained load. The loss is fastest in the first hours and continues over the first thermal cycles. A joint that measured correctly on the assembly bench can be loose after a week in service.

Copper is softer than steel and creeps at moderate temperatures. Tin plating is softer still, and a tin layer under a washer can flow out of the contact area over time. Aluminium creeps more than copper. The result is a gradual loss of preload, a rise in contact resistance and a hotter joint, which in turn accelerates the creep.

What a Belleville washer adds

A Belleville washer is a conical disc spring. It is tightened towards flat, and the load rises steeply with a small deflection. Compared with a flat washer, it stores more travel for a given change in load. When the joint relaxes, the disc extends and gives back some of that travel, so the clamp force falls more slowly.

The effect depends on the disc being used in its working range. A disc that is already flat has no travel left and behaves like a flat washer. A disc that is barely deflected has little load. The washer should be selected so that the installed deflection sits between about 25 and 75 per cent of its cone height, which is the part of the curve where the load is useful and the travel is still available.

Orientation matters. A single disc is usually fitted with the cone facing the bolt head or the nut, and the large end bears on the clamped face. Two discs in series, with the cones facing each other, add travel. Two discs in parallel, nested the same way, add load. The assembly drawing should show the stack, not leave it to the fitter.

Selecting the washer and the bolt

Start with the clamp force the joint needs and the travel the joint is expected to lose. Then choose a disc that gives that load at a usable deflection. The outside diameter should cover the contact area under the bolt head without overhanging the bar. The inside diameter should clear the thread and centre the disc.

The bolt has to carry the preload without yielding. A higher clamp force needs a larger bolt or a higher grade, and the thread engagement in the bar has to suit. The aim is still a joint where the bolt is the sacrificial part: if the joint is overloaded, the bolt should stretch or break before the thread strips out of the copper.

Torque records and re-torque

A torque record should identify the joint, the bolt, the washer stack, the lubricant, the tool and the value. The tool should be calibrated, and the calibration date should be traceable. A click wrench that has drifted gives a false record even when the operator follows the instruction.

For a joint that sees thermal cycling, a sample re-torque after the first cycle is a useful check. Mark the nut and the bar before the cycle, then see whether the nut has moved when the joint is checked. A nut that turns further has lost preload. A nut that does not move may still have lost force if the joint has bedded in, so the check is a guide rather than a measurement.

The assembly drawing should give the tightening sequence for a multi-bolt joint. Tightening one bolt to full torque before its neighbour can tilt the bar and leave the first bolt loose. A cross pattern in two stages is a simple way to seat the joint evenly.

• Record the bolt grade, the washer stack and the lubricant.
• Calibrate the torque tool and record the calibration date.
• Tighten multi-bolt joints in a cross pattern in two stages.
• Mark the joint so a later movement is visible.
• Review the torque when the plating or the washer changes.

Failure modes to design out

The most common failure is a joint that was never tight enough. The bolt felt tight because the friction was high, but the clamp force was low. A dry, rough thread and a burred seat can absorb most of the torque. Cleaning the thread and specifying a lubricant is often more effective than raising the torque.

The second is over-torque. A soft copper bar can strip the thread or deform under the washer before the bolt reaches its intended preload. A hard washer spreads the load and protects the bar. A thread insert gives a stronger thread in a thin or soft section.

The third is relaxation that nobody checks. A joint that was correct at build can lose force in service and give no visible sign until the temperature rises. A Belleville washer, a re-torque step or a condition-based check on the joint temperature are the practical responses.