Application Note
Why a hot joint loses preload
A joint that holds torque on the bench can relax in service, because steel and aluminium expand at different rates. What the material choice does to preload at temperature.
150 °C
Aluminium bolt service temperature
>96%
Preload loss, steel bolt at 150 °C
40–55%
Preload loss, aluminium bolt
65%
Weight saving vs steel
The situation
What actually happens in service
A bolted joint is tightened to its specified torque at room temperature, passes its audit, and ships. In service the assembly reaches operating temperature, cools, and repeats that cycle for years. At some point the joint is found to be loose, even though nothing about the installation changed.
The mechanism is thermal expansion. A bolt and the parts it clamps are a stack of materials, and if they expand at different rates the stack grows by different amounts. The bolt stretches or the joint closes up, and the preload that was set at assembly is no longer the preload the joint is running at.
Analysis
The mechanisms, in order
Each mechanism below produces a different symptom, and each points to a different fix.
01
The mismatch is in the material pair
Steel and aluminium have different coefficients of thermal expansion. In a joint that combines them, a temperature rise changes the relative length of the bolt and the clamped stack. Where the housing grows more than the bolt, the joint relaxes and preload falls; the bolt is not loosening, the joint is unloading.
02
The measured effect is large
In the aluminium bolt programme documented in our material, a class 10.9 steel bolt in a light-alloy joint loses more than 96% of its preload at 150 °C. An aluminium bolt in the same condition loses 40–55%, because its expansion is much closer to that of the housing it clamps.
03
Material choice addresses both failure modes
Choosing a fastener metallurgically close to the structure solves two problems with one decision. The thermal mismatch is reduced, so preload survives at temperature, and the galvanic driving force is reduced at the same time, so the housing is not consumed by corrosion. Neither benefit requires an added coating or an isolating washer.
04
Strength is not the only specification
A high-strength aluminium bolt averages 629 MPa tensile and 510 MPa yield, which is comparable to a class 8.8 steel bolt at roughly one third of the density. It will not match a 12.9 steel bolt in absolute strength, and its lower modulus changes how the joint behaves — so the tightening process has to be specified for the material rather than carried over from steel.
What to specify
Design points worth putting on the drawing
- Specify preload at operating temperature, not only at assembly temperature
- Match the fastener material to the structure where thermal cycling is significant
- Control friction with an anodised or lubricated finish so torque converts predictably to preload
- Account for the lower modulus of aluminium when setting the tightening process
- Check the galvanic couple at the same time as the thermal one — one material choice can address both
- Verify on the actual joint rather than on a nominal table, because joint stiffness is part of the result
Related products
Parts that address these mechanisms
Keep reading
Other application notes
Embodied Intelligence
Keeping play out of a finger joint
Play in a dexterous finger usually starts at the pivot, not at the screw thread. A short engineering look at why a shoulder screw behaves differently from a plain screw in a rotating joint.
Consumer Electronics
When the stack height is already fixed
In a thin device the fastener is usually the last thing chosen and the first thing that has to give. How head form and drive choice recover height without changing the screw diameter.
Send us the joint
A drawing, the service conditions and the failure you are seeing are enough to start. The answer is usually a mechanism, not a part number.