Bolt Torque and Preload: Where the Numbers Come From
Short answer
- Torque is only a proxy for bolt preload, because the real objective of tightening is to generate a specific clamp force that holds the joint members together.
- In a standard unlubricated steel bolt, roughly 85 to 90 percent of the applied torque is consumed by friction, and a 20 percent variation in the Nut Factor K produces a 20 percent variation in preload for the same torque.
- Critical applications use angle control, bolt elongation measurement, or hydraulic tensioners to measure or control tension directly instead of relying on torque as a proxy.
The objective of tightening a bolt is not to hit a specific torque value; it is to generate a specific clamp force, or preload, holding the joint members together. Torque is merely the input handle we use to get there. The confusion arises because the relationship between applied torque and achieved preload is highly variable, dominated by friction rather than the mechanics of stretching the bolt.
Where does bolt tightening torque actually go?
When you apply torque to a fastener, that energy has to go somewhere. Mechanics typically break down the torque distribution into three components: the torque required to overcome thread friction, the torque required to overcome bearing friction under the bolt head or nut face, and the torque that actually creates bolt tension. In a standard unlubricated steel bolt, roughly 85 to 90 percent of the applied torque is consumed by friction. Only the remaining 10 to 15 percent goes into stretching the bolt to create clamp force.
This means the friction coefficient dominates the outcome. If you take a standard steel bolt and apply a specified torque of 100 N·m, you will get a specific preload. If you then apply a lubricant or anti-seize to that same bolt and apply the same 100 N·m, the friction coefficient drops significantly. Because less torque is consumed by friction, a much larger portion of that 100 N·m goes into stretching the bolt. The bolt can easily be pushed past its yield point and fail, or clamp the joint so hard that it crushes the mating surfaces. This is why a lubricated bolt must never be tightened to dry torque values.
When sourcing fasteners and hardware, always verify whether the published torque values account for lubrication or assume a dry, as-received finish. Plating also alters the friction coefficient; zinc plating, for instance, behaves differently than black oxide or plain oil-finished steel.
How are torque and bolt preload related?
The engineering shorthand for this relationship is the Nut Factor equation: T = K × D × F. In this formula, T is the applied torque, D is the nominal bolt diameter, F is the achieved preload, and K is the empirical Nut Factor. The Nut Factor captures all the friction variables—thread friction, bearing face friction, and the geometry of the threads—into a single dimensionless coefficient.
For standard unlubricated steel bolts, K is often approximated at 0.20. If the threads and bearing face are lubricated, K might drop to 0.15 or lower. The variability of K is exactly why torque control is inherently imprecise. A 20 percent variation in K due to surface finish, plating, or lubrication results in a 20 percent variation in preload for the exact same applied torque. Since we want a specific clamp force to prevent joint separation, fatigue, or gasket leaks, relying purely on a torque wrench reading leaves a wide margin of uncertainty.
Using calibrated hand tools ensures you are applying the torque you intend, but it does not solve the friction variability problem. The wrench measures torque accurately; it simply does not know how much of that torque is becoming clamp force.
What preload should a bolt be tightened to?
Design engineers specify a target preload as a percentage of the bolt's proof load. Proof load is the maximum tensile force a bolt can withstand without taking a permanent set—it is the yield strength boundary for the fastener. For standard static joints, the target preload is typically 75 percent of proof load. For joints requiring high fatigue resistance or where separation forces are extreme, designers might target 90 percent of proof load.
The problem on the maintenance floor is that if K is lower than expected—say, due to the inadvertent presence of oil or the use of lubricants on the threads—targeting 75 percent of proof load via torque can easily generate a preload exceeding 100 percent of proof load. The bolt yields, loses its elastic clamping force, and must be replaced. Conversely, if corrosion or surface scale increases K, the bolt might only reach 50 percent of proof load, leaving the joint susceptible to separation and fatigue failure under dynamic loads.
This is also why the choice of thread coatings and chemical aids matters. Applying thread lockers and anti-seize changes the Nut Factor drastically. If a procedure calls for a thread-locking compound but specifies a dry torque value, the joint will be over-tensioned. Always cross-reference the chemical manufacturer's data for adjusted torque values or Nut Factors when altering the bolt's surface condition.
Why do bolted joints need multiple tightening passes?
Even if you perfectly calculate the torque and account for friction, a single-pass tightening strategy will not achieve the target preload across a joint. Joints relax after tightening. This occurs due to embedment, where the microscopic high points of the machined or cast surfaces plastically deform and flatten out under the high contact pressure of the clamp. Gasket creep and micro-slip also contribute to relaxation. Typically, a joint can lose 5 to 10 percent of its initial preload simply due to embedment shortly after tightening.
To combat this, tightening sequences must be executed in multiple passes. For flanged joints, a standard cross-bolting or star-pattern sequence is required. The first pass might be to 30 percent of final torque, the second to 70 percent, and the final pass to 100 percent. A fourth pass or a verification pass is often specified for critical joints to confirm that embedment has stabilized and all fasteners maintain equal tension. Skipping passes or tightening a single bolt to full torque in one go distorts the flange and prevents adjacent bolts from achieving their required preload.
What tightening methods work better than torque control?
Because torque is an imprecise proxy for preload, critical applications use methods that measure or control tension directly.
Angle-Controlled Tightening (Torque-to-Yield)
Angle control, often called torque-to-yield (TTY), separates the tightening process into two phases. First, a low snug torque is applied to pull the joint members into contact and remove slack. Second, the fastener is turned a specific number of degrees. Since bolt elongation is directly proportional to the angle of rotation once the joint is solid, controlling the angle controls the preload. This method largely ignores friction. Whether the bolt is dry or lubricated, turning it 90 degrees past the snug point stretches it by the same amount. TTY is common in automotive engine assembly, but it permanently yields the bolt, meaning the bolt must be replaced if the joint is ever disassembled.
Bolt Elongation Measurement
For large structural bolts or highly critical equipment, preload is verified by measuring the physical elongation of the bolt. This is done using micrometers to measure the bolt length before and after tightening, or by using ultrasonic measurement devices that send a sound pulse through the bolt and measure the transit time, which changes as the bolt stretches. This is the most accurate method of determining preload, completely bypassing the friction variable.
Hydraulic Tensioners
For very large fasteners, hydraulic tensioners are used. A tensioner fits over the bolt, grips the protruding threads, and pulls the bolt axially, stretching it hydraulically. Once the bolt is stretched, the nut is run down against the joint face with minimal torque. The hydraulic pressure is released, and the nut holds the stretched bolt in place, creating pure tension without introducing torsional stress or relying on friction to generate the clamp force.
If you are specifying fasteners, chemical aids, or tightening procedures for a critical application and need to cross-reference components or source specific grades and finishes, reach out to our team. We can help you source the correct hardware and chemical products to ensure your joints achieve and maintain their designed clamp force.
Need a quote on a specific part?
Request a QuoteFrequently asked questions
Why doesn't applied torque equal bolt preload?
Torque is only a proxy for preload because friction dominates the relationship. In a standard unlubricated steel bolt, roughly 85 to 90 percent of the applied torque is consumed by thread friction and bearing friction under the bolt head or nut face. Only the remaining 10 to 15 percent goes into stretching the bolt to create clamp force.
What is the nut factor in the torque-tension equation?
The Nut Factor, K, is the empirical coefficient in T = K × D × F, where T is applied torque, D is nominal bolt diameter, and F is achieved preload. It captures thread friction, bearing face friction, and thread geometry in one dimensionless number. For standard unlubricated steel bolts, K is often approximated at 0.20, and 0.15 or lower when lubricated.
What happens if you tighten a lubricated bolt to dry torque values?
Lubricant or anti-seize lowers the friction coefficient, so a much larger portion of the applied torque goes into stretching the bolt. The bolt can be pushed past its yield point and fail, or clamp the joint so hard that it crushes the mating surfaces. For this reason, a lubricated bolt must never be tightened to dry torque values.
How should flanged bolted joints be tightened to account for relaxation?
Joints relax after tightening through embedment, gasket creep, and micro-slip, and a joint can lose 5 to 10 percent of its initial preload to embedment alone. Flanged joints are tightened in multiple passes using a cross-bolting or star pattern, such as 30 percent, 70 percent, then 100 percent of final torque, with a verification pass for critical joints.