Machine Design #57: Bolted Joints — Clamp Load, Locking, and the Limits of Tightening
1. Start with the function, not the component name
Before opening a catalog or fixing a dimension, write what the bolted joint must do, where the load comes from, how many cycles it must survive, and what counts as failure. “Make it like the old machine” hides the assumptions that later become fit, noise, wear, or safety problems. A testable requirement states the input condition, operating condition, expected result, acceptance limit, and measurement method. Internal standards are a starting point; the final decision must be checked against the model, material, process, and actual installation.
2. Four layers must be reviewed together
Function and load
Separate nominal, start-up, impact, misalignment, and fault loads. For a moving mechanism include inertia, acceleration, dwell, and cycle count. Trace the force through the whole assembly instead of validating a single attractive CAD section. The design value must include a stated boundary and a reason for the selected safety margin.
Material and manufacturing process
The same geometry behaves differently when it is cut from sheet, machined from bar, formed, cast, welded, or heat treated. The drawing should describe what the workshop can make and measure. A tight tolerance is not a substitute for understanding the process, grain direction, residual stress, surface condition, and inspection method.
Assembly and interfaces
Every interface needs a datum, insertion direction, clearance, fastening or contact condition, and a way to prevent the wrong part or orientation. If left/right, front/rear, or model variants exist, use geometry, markings, or keyed features to make a mistake difficult. Record the tool access and the force that the operator is expected to apply.
Operation and maintenance
Review how a technician reaches the part, how a replacement is located again, whether another module must be removed, and which checks restore the machine to baseline. A design that works only with an experienced fitter is not yet a repeatable design.
3. Core design checks
- Tension, shear, vibration, and fault-case loads: separate the tension, the shear and the vibration the joint sees, and add the fault case — an interlock trip or a crash loads the joint differently from normal running.
- Bolt grade and female-thread material: state the bolt property class together with the material it threads into; a high-class bolt in soft aluminium strips the female thread before it reaches clamp load.
- Thread engagement length and stripping margin: give the engagement length and the margin against thread stripping, so the failure, if it comes, is in the bolt and not in the part.
- Torque-angle or other clamp-load control method: name the tightening method, because torque alone scatters widely with friction; if torque is all you have, state the lubrication condition it assumes.
- Locking method selected for the actual vibration: choose the locking method for the vibration that is really present, and say what proves it works — a witness mark that has moved is the cheapest evidence there is.
- Tool clearance, witness marks, and audit evidence: check that the tool actually fits, and record the tightening evidence in a form an auditor can read months later.
The checklist is useful only when each line has evidence. “Reviewed” is not the same as “calculated”, and “calculated” is not the same as “tested at the boundary”. Keep the evidence ID beside the requirement so a later engineering change can be audited.
3b. Clamp load is the goal; torque is only an indirect measure
A bolt holds a joint through the clamp load it creates between two faces, not through the torque figure written on the drawing. The problem is that clamp load is hard to measure on the shop floor, so people tighten to a torque. The bridge between the two quantities is friction, and friction is the most variable term in the whole calculation.
Most of the torque applied at the wrench is consumed overcoming friction in the threads and under the bolt head; only the remainder becomes clamp load. So the same torque produces noticeably different clamp loads depending on whether the threads are dry or lubricated, new or already tightened several times, coated or bare. Stating a torque without stating the condition is stating half a requirement.
| Control method | Accuracy of the resulting clamp load | Used in practice for |
|---|
| Torque wrench, torque only | Lowest of the group, error dominated by friction | Ordinary joints, high quantities |
| Torque to a seating value, then a specified turn angle | Better, because the angle relates to the real bolt extension | Important joints with a defined procedure |
| Measuring bolt elongation | High | Large bolts, critical joints |
| Hydraulic tensioning | High | Large structures, large-diameter bolts |
The usable conclusion: on an important joint, do not state torque alone. Add the thread condition (dry, lubricated, with or without thread locker) and consider a better control method.
Tighter is not automatically safer
Three limits sit above:
- The limit of the bolt itself. Beyond it the bolt yields and loses its elastic behaviour,
which means losing the very clamp load it is there to hold.
- Embedding of the contact faces. Rough surfaces, paint layers, soft washers and plastic
parts all settle under pressure. A settlement of a few per cent of the clamped length is enough to lose most of the initial clamp load. Plastic parts need an insert or a metal bush — see Machine Design #96 — Inserts in plastic parts.
- The strength of the female thread. A thread in aluminium, cast iron or plastic is weaker
than one in steel, so the same bolt class demands a longer thread engagement. Stripping the female thread is the worst failure of the three because it means repairing an expensive part, not replacing a bolt.
Two different loosening mechanisms need two different countermeasures
Calling everything "bolt loosening" leads people to pick the wrong remedy. There are two mechanisms:
Clamp load is lost although the bolt never turns. The causes are embedding of the contact faces, creep of the clamped material, or differential thermal expansion between bolt and part. Anti-rotation devices do nothing here. The remedies are: increase the elastic clamped length, reduce the number of interfaces, use hard washers to spread the pressure, and re-tighten after the first thermal cycle.
The bolt turns itself out because the faces slip sideways. When the transverse load is enough to make the two faces move relative to each other, even minutely, the bolt unwinds. This is the governing mechanism on machines with vibration. Remedies in order of effectiveness:
- Stop the faces slipping. Increase clamp load, increase the friction of the contact faces,
or carry the transverse load on locating pins and shoulders instead of on the bolt shank — see Machine Design #104 — Fits for pins and bushings.
- Chemical or added-friction locking. Thread locking adhesive, prevailing-torque nuts.
- Form locking. Wedge-type locking washers, split pins, lock wire — used where the
consequences are severe.
Ordinary spring and wave washers have limited effect against this transverse-slip mechanism; do not treat them as the only measure on a joint that sees vibration.
Four things that belong on the drawing or in the assembly instruction
- The tightening torque with its condition (dry, lubricated, with or without thread locker).
- The tightening sequence on multi-bolt joints — crosswise and in several steps, because
tightening one bolt changes the clamp load of its neighbour.
- Whether the joint is re-tightened after run-in or after the first thermal cycle.
- Which parts carry transverse load on pins and which carry it by friction, so that a future
repair does not remove a locating pin and refit it by eye.
4. Tolerance stack and variation
Do not judge dimensions independently. Build the stack from the functional datum to the characteristic that must be protected. Distinguish nominal clearance, worst-case accumulation, and a statistical distribution only when the process is stable enough to justify it. Include flatness and squareness of the datum, coating or heat-treatment thickness, joint slip and deformation during tightening, operating temperature, field assembly error, and wear over time.
If assembly succeeds only because a technician “nudges it a little”, the design has no reliable capability. Capture the stack in a simple table, assign each contributor a source, and identify which dimensions are controlled by the supplier and which are verified at incoming inspection.
5. Failure modes to ask before release
Ask: “If this is wrong, what will the machine show, and what evidence will reveal the cause?”
| Failure mode | Machine symptom | Verification direction |
|---|
| Boundary selected too optimistically | Passes a demo, fails at speed, temperature, or load | Test min/max and the defined fault case |
| Tolerance not tied to function | Difficult assembly, play, bind, or noise | Rebuild the stack from functional datums |
| Manufacturing step omitted | Distortion, burr, vibration, or drift after treatment | Review with the shop and measure after each critical step |
| Maintenance not designed | Long replacement time or wrong restoration | Run a maintenance trial and restoration check |
| Documents out of sync | Correct part name but wrong revision or setting | Baseline BOM, drawing, configuration, and work instruction |
For this article, also challenge the specific risk in the title: clamp load, friction scatter, thread engagement, vibration, and controlled tightening. A failure mode is not closed by a sentence in a report; it is closed by a measured result, an owner, and a clear re-test condition.
6. What to put on the drawing and in the record
Specify only functional requirements that can be inspected. For a special requirement state its scope, datum or measurement location, measurement condition, and acceptance limit. Notes such as “machine accurately” or “assemble carefully” do not tell a supplier what to do. Keep the calculation or rationale, source revision, assumptions and limits, review comments, prototype or FAT results, and the point that must be rechecked after an engineering change.
7. Practical design review for bolted joint
The first review should be a short, evidence-led conversation. Start at the input and follow the load to the output. For bolted joint, draw the load path and name the surface, edge, thread, bearing, contact, or follower that actually carries it. Mark where friction, clearance, temperature, lubrication, or operator adjustment can change the result. If a parameter is unknown, do not hide it in a generous factor; assign an owner and a measurement plan.
The second review should use the real production route. Ask the fabricator how the feature is made, which operation creates the most variation, how the feature is inspected, and what happens after coating, heat treatment, deburring, or cleaning. Compare the process capability with the tolerance stack. If the process cannot hold the drawing, change the process or the design before ordering parts.
The third review should be performed on the assembled machine. Check access, orientation, tool engagement, marks, guards, and the time required to replace the part. Run the machine through start-up, normal duty, stop, restart, and a controlled fault. Record sound, temperature, vibration, motion smoothness, and any visible witness mark. These observations are often the earliest evidence of a design assumption that was too optimistic.
8. Boundary conditions and calculation discipline
State the lower and upper values for every influential parameter. Include the worst combination that can occur together, not only the maximum of each parameter in isolation. Keep units consistent and write the equation or reference used. When a catalog rating is used, record its test condition and correction factors. When a simulation is used, compare at least one result with a hand calculation or a measured prototype.
For bolted joint, a useful calculation sheet has columns for input, nominal, lower bound, upper bound, source, result, and pass/fail. Add a final column for “what would make this assumption invalid?” This keeps the analysis alive when a supplier changes material, when a cycle time increases, or when the machine is installed in a warmer or dirtier environment than the prototype.
9. Supplier and incoming-inspection handoff
The supplier package should contain the drawing, revision, material condition, special-process note, inspection points, and a sample acceptance record. Do not outsource the design intent. If a supplier proposes a different material or process, compare strength, fatigue, friction, corrosion, thermal behavior, lead time, and inspection capability before approving the change.
At incoming inspection, measure the features that protect the function rather than every dimension with equal effort. Record the actual value and instrument ID. A part can be “within drawing tolerance” and still fail because the wrong surface, burr, fit, lubricant, or orientation was accepted. Link the inspection record to the serial or lot number used on the machine.
10. Commissioning and maintenance feedback
Create a baseline at commissioning: cycle time, force or torque, temperature, vibration, noise, position, and visual condition as relevant. Define the alarm limit and the action when the limit is reached. After the first service interval, compare the trend with the baseline and update the maintenance instruction. A replacement should restore the same baseline, not merely make the machine move again.
Clamp load is only useful if it STAYS: the joint diagram and preload loss
Reaching the right clamp load at assembly is not enough; the real question is how long it holds. Two mechanics decide this:
- The joint diagram. In a joint where the members are much stiffer than the bolt, an external separating load mostly relieves the clamp and adds only a small part to the bolt tension. That is why a well-preloaded bolt sees little of the fluctuating load — high preload protects the bolt against fatigue. A soft joint (thick gasket, thin flanges) does the opposite and pushes most of the fluctuation into the bolt.
- Loss of preload over time. Clamp load bleeds away through embedding — surface asperities flattening under load — and through creep/relaxation of gaskets and soft coatings. A short grip length stores little elastic stretch, so it loses a larger fraction.
| Risk | Mechanism | Design response |
|---|
| Bolt fatigues even at low mean load | A soft joint pushes the fluctuation into the bolt | Make the joint stiff, preload high enough, keep gaskets out of the load path |
| Loosening after a few thermal/vibration cycles | Embedding plus gasket/coating creep | Adequate grip length; limit soft gaskets; a re-torque schedule for critical joints |
| Same torque, scattered clamp load | Thread and face friction vary | Verify by angle/turn or stretch (DTI), not by torque alone |
State the target clamp load (not just torque), the grip length, whether a gasket sits in the load path, and the method used to verify the preload.
11. MINATA release checklist
- [ ] Function, load path, duty cycle, and fault boundary are written.
- [ ] Material, process, surface condition, and inspection method are agreed.
- [ ] Functional datums and the tolerance stack are visible.
- [ ] The six topic checks above have evidence and pass/fail criteria.
- [ ] Assembly, tool access, orientation, and mistake-proofing were tried on hardware.
- [ ] Failure modes have an owner, evidence, and re-test condition.
- [ ] Supplier and incoming-inspection records are linked to the revision.
- [ ] Commissioning baseline and maintenance response are defined.
Frequently asked questions
The torque was correct but the joint still loosened. Why?
Because torque is only an indirect measure of clamp load, and most of it is consumed by friction. Dry or lubricated threads, coated or bare, first tightening or fifth, all give different clamp loads while the wrench reads the same number.
Do spring washers prevent loosening?
Their effect against transverse-slip self-loosening, the main mechanism on machines with vibration, is limited. The stronger lever is to stop the faces slipping: raise the clamp load and carry the transverse load on locating pins. Thread locker or prevailing-torque nuts come after that.
Should a bolt carry shear?
Better avoided. A good joint transfers transverse load by friction between the faces, generated by the clamp load; the bolt shank carries shear only by deliberate design and with a fitted bolt. The usual approach is a locating pin for the transverse load, leaving the bolt to do the clamping.
What changes for a thread in aluminium?
Increase the thread engagement compared with steel, and consider a steel thread insert where the joint is opened often. Stripping the female thread in an expensive aluminium part is the most costly failure mode of a bolted joint.
Does a joint need re-tightening after the machine runs?
On assemblies that get hot or that have many interfaces, yes. Embedding and differential thermal expansion remove clamp load during the first running hours without the bolt turning at all. If re-tightening is needed, put it in the assembly instruction and in the maintenance schedule rather than leaving it to memory.
12. Closing note
Good mechanical design is not a collection of perfect-looking dimensions. It is a chain of explicit assumptions that survives manufacturing, assembly, operation, maintenance, and change. For bolted joint, the right question is not only “will it work?” but “what evidence will show that it continues to work when the boundary moves?” That is the standard MINATA uses to turn a drawing into a dependable machine.
Frequently asked questions, continued
Why does a longer grip length resist loosening better?
A longer bolt stores more elastic stretch at the same clamp load, so when a small amount of embedding or creep occurs, the fraction of clamp load lost is smaller. A short joint that is stiff on the bolt side loses a lot of clamp load from only a little settling.
Should a soft gasket sit directly in the clamp load path?
Avoid it. A soft gasket creeps and makes the joint "soft", which pushes external load into the bolt and lets the clamp load decay. If sealing is required, separate the sealing function from the clamp path, or use a compression-limited gasket or a stop ring.
Conclusion
Bolted Joints is not a matter of picking a single number. Good design ties function, load, material, process, tolerance, assembly and maintenance into one contract that can be checked. An internal standard preserves the experience already accumulated; the designer's job is to understand the conditions under which it applies and turn it into a clear decision on the product at hand.
View all MINATA technical articles