Rigid Couplings, Knuckle Joints and Lever Design: How Three "Simple" Components Brought a Crushing Plant to...

How Three "Simple" Components Brought a Crushing Plant to Its Knees — And the Design Masterclass That Fixed Everything

Reading time: ~40 minutes | Complete engineering study with formulas, tables & worked examples

A rigid coupling splits at 2:47 AM. A knuckle joint tears itself apart under load. A lever arm buckles at the fulcrum. Three failures. One root cause. Nobody actually understood the design process.

What This Post Will Give You

By the time you finish reading, you will understand the complete design process for three of the most fundamental machine elements in mechanical engineering:

• Rigid Couplings — the silent workhorses connecting rotating shafts
• Knuckle Joints — the elegant pin connections transmitting axial loads
• Levers — the ancient force multipliers still governing modern machinery

You will walk away with every formula, every design proportion, every stress analysis method, and the engineering judgment to know when the textbook answer isn't enough.

This isn't theory for theory's sake. This is how machines stay together under load — or don't.

The Philosophy Nobody Teaches You

Why "Good Proportions" Matter More Than You Think

Before the practitioner let the practitioner touch a single calculation, he sat her down in the control room and said something that changed how she approached every design problem from that point forward.

"In practice, design of simple machine elements is often done by good proportions. The principle is this: if it looks right and it looks wrong, it probably is wrong."

the practitioner frowned. "That sounds... unscientific."

"It sounds unscientific because you haven't designed enough things yet," the practitioner replied. "Good proportions aren't guesswork. They're the condensed wisdom of thousands of engineers who've watched things break. When something looks too thin, too stubby, too spindly — your eye is detecting a stress concentration or a load path problem before your calculator can."

But here's the critical caveat that catches people: this principle does not always apply. A design may require special materials or have specific limitations — space restrictions, weight constraints, unusual loading — that cause the design to deviate from accepted proportions. In those cases, you must calculate. You must verify. You must prove the design works.

The Rule: Start with good proportions. Then do the stress analysis to confirm them. If the analysis says the proportions are wrong, trust the numbers. If the proportions look wrong and the numbers say they're fine, check the numbers again.

Casting vs. Fabrication: The First Decision

Every machine element covered in this chapter can be manufactured by one of two primary methods:

Casting (typically cast iron or cast steel):

• Preferred when many identical units are to be manufactured
• Produces complex shapes economically at volume
• Internal stress patterns differ from fabricated parts
• Surface finish and dimensional tolerance are generally looser

Fabrication/Welding (typically from steel plate and bar):

• Preferred when only one or two units are to be manufactured
• More flexible for modifications and repairs
• Welded joints introduce residual stresses that must be considered
• Generally faster for prototypes and one-offs

The design procedure produces values that vary somewhat with the method of manufacture. This is not a minor footnote — it fundamentally affects how you proportion your elements, what stresses you check, and what safety factors you apply.

Rigid Couplings — The Connection You Never Think About Until It Fails

What the practitioner Found in the Wreckage

The coupling that failed was a standard flanged rigid coupling — the kind used to connect two collinear shafts where no angular or parallel misalignment is expected. It had four distinct zones of damage:

1. The hub (boss) had cracked along its bore
2. The web had sheared where it met the flange
3. The flange bolts had stretched beyond yield
4. The spigot (the locating register) showed fretting damage

the practitioner pointed to each zone in sequence. "Four failure modes. The original designer checked for one — the bolt shear. That's like building a bridge and only checking the paint."

Anatomy of a Rigid Coupling

A rigid coupling consists of four primary components:

┌─────────────────────────────────────────────┐ │ │ │ ┌───────────────────┐ │ │ │ F L A N G E │ │ │ │ │ │ │ ══════╪═══╦═══════╦═══╪══════ │ │ │ ║ HUB ║ │ │ │ │ ║(BOSS) ║ │ │ │ │ ║ ║ │ │ │ ══════╪═══╩═══════╩═══╪══════ │ │ │ S P I G O T │ │ │ │ W E B │ │ │ └───────────────────┘ │ │ │ │ ══════ = Shaft │ └─────────────────────────────────────────────┘

The Design Process for a Rigid Coupling

Here's what the practitioner walked the practitioner through — and what the original designer skipped.

Step 1: Determine the Torque

The coupling must transmit the full torque from the driving shaft to the driven shaft. If the power P and rotational speed n are known:

$$T = \frac{P}{2\pi n}$$

Where:

• T = torque (N·m)
• P = power (W)
• n = rotational speed (rev/s)

Critical Note: Always apply a service factor to account for shock loads, starting torques, and dynamic overloads. Typical service factors range from 1.25 for smooth, steady loads to 3.0+ for heavy shock applications.

Step 2: Design the Hub (Boss)

The hub transmits torque from the shaft to the web. The connection is typically either:

• A key and keyway (most common)
• An interference (shrink) fit
• A combination of both

For a keyed connection, the hub must be checked for:

Bearing stress on key:

$$f_{bearing} = \frac{F}{A_{bearing}} = \frac{T}{r \times \frac{h}{2} \times L}$$

Where:

• r = shaft radius
• h = key height
• L = key length (effective contact length)

Shear stress on key:

$$f_{shear} = \frac{F}{A_{shear}} = \frac{T}{r \times w \times L}$$

Where:

• w = key width

Hub hoop stress (for interference fits):

The hub must not burst under the interference pressure. This is analyzed using thick-cylinder theory (Lamé equations).

Step 3: Design the Web

The web transfers torque from the hub outward to the bolt circle. It experiences:

Shear stress in the web:

$$f_s = \frac{T}{2\pi r_w^2 \times t_w}$$

Where:

• r_w = mean radius of web
• t_w = web thickness

Step 4: Design the Flange and Bolts

The bolts on the flange must resist the tangential force created by the torque at the bolt circle radius:

Tangential force per bolt:

$$F_{bolt} = \frac{T}{R_{bc} \times n_{bolts}}$$

Where:

• R_bc = bolt circle radius
• n_bolts = number of bolts

Each bolt must be checked for shear (if the joint is a bearing-type connection) or tension (if the joint is a friction-grip connection):

Bolt shear stress:

$$f_s = \frac{F_{bolt}}{A_{bolt}}$$

Where A_bolt is the bolt shank area (for bearing type) or tensile stress area (for friction grip).

Step 5: Check the Spigot

The spigot provides location only — it should not transmit torque. However, bearing stress on the spigot face must be checked to prevent fretting:

$$f_{bearing} = \frac{F_{spigot}}{A_{spigot}}$$

The Mistake That Killed the Coupling

"The original designer," the practitioner explained, "used a service factor of 1.0 on a jaw crusher drive. A jaw crusher. The thing that literally breaks rocks."

the practitioner winced.

"A jaw crusher needs a minimum service factor of 2.5 — often 3.0 — because of the cyclic impact loading. The designer calculated the bolts for steady-state torque and called it done. The web was never checked. The hub interference fit was marginal. And the spigot? Pure afterthought."

Your takeaway: The coupling design is only as good as its weakest element. You must check every component — hub, web, flange, bolts, and spigot — under the actual service conditions with appropriate factors.

Knuckle Joints — The Pin Connection That Demands Respect

The Second Failure

While the practitioner and the practitioner examined the coupling wreckage, a maintenance technician brought over the remains of the tensioning mechanism's knuckle joint. The pin had bent. The eye had cracked. The fork showed bearing damage.

"This one's actually a beautiful design problem," the practitioner said, holding up the deformed pin. "A knuckle joint looks simple. Pin through a fork and an eye. A child could sketch it. But the stress analysis? That's where engineers earn their keep."

What Is a Knuckle Joint?

A knuckle joint connects two rods that are subject to axial tensile (or compressive) forces. It consists of three main components:

     ┌──────────┐
═════╡   EYE    ╞═══ Rod
     │    ┌─┐   │
     │    │P│   │       P = Pin
     │    │I│   │
─────╡    │N│   ╞───── Fork (2 prongs)
─────╡    │ │   ╞─────
     │    └─┘   │
     └──────────┘

Key Features:

• The rod transmits the axial load to the eye
• The eye wraps around the pin with width a
• The fork straddles the eye with two prongs, each of width b
• The pin passes through both fork and eye, carrying the load in shear and bending
• In the knuckle joint illustrated, the rods are integral with the eye and fork (forged)
• However, the knuckle joint is often separate to the rods, and then needs to be welded or screwed into the eye and fork

Important Notes:

• Knuckle joints may be cast or fabricated — if they are relatively small, they may also be forged
• In the knuckle joint illustrated, there is no separate bearing and rotational or oscillating motion occurs between the pin and eye or pin and fork (or both)
• If this is the case, the pin is usually a tight fit in the eye or held to the eye via a grub screw and bearings are provided in the fork
• The bearings may be plain bearings or rolling element bearings
• If there is considerable movement, it may be necessary to use bearings to minimise friction and wear

Good Proportions for Knuckle Joints

This is where decades of engineering experience get distilled into a few simple ratios. For steel knuckle joints without bearings, good proportions are:

Why do these proportions work? Because they balance the stress distribution across all failure modes — tensile, shear, bearing, and bending — so that no single component is dramatically weaker than the others. When proportions are "good," no one stress dominates catastrophically.

the practitioner drew a quick sketch on the back of a maintenance form. "Notice that 2__a__ is less than b. That means the stresses in the fork will always be less critical than in the eye. The eye is your weakest link. That's where you focus your analysis."

Complete Stress Analysis of a Knuckle Joint

This is the full engineering methodology. Every stress. Every component. No shortcuts.

Rod Design

Condition: Both ends are normally pinned, so there are no bending or shear loads on the rod. The rod needs to be designed for tensile stress only.

Tensile stress in the rod:

$$f_t = \frac{F}{A_r} = \frac{F}{\frac{\pi d^2}{4}}$$

Where:

• F = axial force (N)
• d = rod diameter (mm)
• A_r = cross-sectional area of rod (mm²)

Rearranging for required rod diameter:

$$d = \sqrt{\frac{4F}{\pi f_t}}$$

But wait — what if the rod sees compression too?

If there is a compressive load as well as a tensile load, the rod must be checked for buckling. This is where the tensioning mechanism failed — the rod was only checked for tension.

Buckling check — Ensure F_c > F:

For short to intermediate columns (Le/k < Le/k_lim):

$$F_c = \sigma_c \times A \times \left[1 - \frac{\sigma_c}{4\pi^2 E} \times \left(\frac{L_e}{k}\right)^2\right]$$

For long (slender) columns (Le/k > Le/k_lim):

$$F_c = \frac{\pi^2 E \times A}{\left(\frac{L_e}{k}\right)^2}$$

Where the transition slenderness ratio is:

$$\frac{L_e}{k}_{lim} = \sqrt{\frac{2\pi^2 E}{\sigma_c}}$$

For a round rod: the radius of gyration k = d/4

Critical Note: Since F_c is the critical buckling force (the force at which the rod will buckle), a safety factor must be applied. The design force must be less than F_c divided by the safety factor.

These are standard column formulae from engineering mechanics. The the practitioner formula (short columns) and Euler formula (long columns) work together to cover the full range of slenderness ratios.

Eye Design

The eye is typically the most critically stressed component in a knuckle joint of standard proportions. Three stresses must be checked:

Tensile stress in the eye:

$$f_t = \frac{F}{(D - d) \times a}$$

Where:

• D = outside diameter of eye
• d = pin diameter (also approximately equal to rod diameter)
• a = width of eye

This formula calculates the stress across the net section — the narrowest cross-section of the eye where material has been removed for the pin hole.

Shear stress in the eye:

$$f_s = \frac{F}{2 \times e \times a}$$

Where:

$$e = \sqrt{\left(\frac{D}{2}\right)^2 - \left(\frac{d}{2}\right)^2}$$

This is the shear area on each side of the pin hole. The factor of 2 accounts for the double shear path through the eye material.

Bearing stress in the eye:

$$f_{brg} = \frac{F}{d \times a}$$

This is the contact pressure between the pin and the bore of the eye. It's critical for wear and for preventing local crushing.

Fork Design

The fork has two prongs, each of width b. For standard proportions where 2a < b, the fork stresses will be less critical than the eye stresses. However, for completeness and for non-standard designs, all fork stresses should be calculated.

Tensile stress in the fork (per prong):

$$f_t = \frac{F}{2(D - d) \times b}$$

Note the factor of 2 — the load is shared between two fork prongs.

Shear stress in the fork (per prong):

$$f_s = \frac{F}{2 \times e \times b}$$

Where e is calculated the same way as for the eye.

Bearing stress in the fork (per prong):

$$f_{brg} = \frac{F}{2 \times d \times b}$$

Again, the factor of 2 accounts for load sharing between the two prongs.

Important Note: The stresses in the fork are the same form as the stresses in the eye, but with 2b in place of a. If the knuckle joint is of standard proportions with the same strength material for the eye and fork, then since 2a < b, the stresses in the eye will always be more critical. In practice, you may only need to analyse the eye. However, if different materials are used, or if proportions are non-standard, you must check both.

Pin Design

The pin is simultaneously loaded in shear and bending. This makes it the most complex analysis in the knuckle joint.

Average shear stress in the pin:

The pin is in double shear — it is sheared on two planes (one on each side of the eye):

$$f_{s,avg} = \frac{F}{2 \times A_{pin}} = \frac{F}{2 \times \frac{\pi d_p^2}{4}}$$

Critical Note: The actual shear stress is higher than the average shear stress because the pin is simultaneously in bending. The bending creates a non-uniform stress distribution across the pin cross-section. The actual peak shear stress in a circular section under combined loading is approximately 4/3 times the average (from the parabolic shear stress distribution).

Bending stress in the pin:

This is where it gets interesting — and where engineers argue.

There are two formulae for the bending moment in the pin, depending on the assumed load distribution:

Formula 1 — Uniformly distributed loading assumption:

$$M = \frac{F(a + b)}{4}$$

This assumes the bearing load is uniformly distributed across the width of the eye (a) and each fork prong (b). The bending moment is maximum at the centre of the pin.

Formula 2 — Concentrated loading assumption:

$$M = \frac{F(a + b)}{4}$$

Both formulae give the same symbolic form, but the interpretation differs. Formula (1) assumes equivalent concentrated loading at the mid-points of the fork and eye. Formula (2) is the most conservative and generally gives a pin size that is too large.

The bending stress is then:

$$f_b = \frac{M}{Z} = \frac{M}{\frac{\pi d_p^3}{32}}$$

Which formula to use? It is recommended that the conservative formula be used for design purposes. If the resulting pin size seems too large, refine using the more detailed distributed load analysis. Both bending moment formulae assume zero side clearance. If there is side clearance between the fork and eye, the bending moment will be greater and must be recalculated using the actual clearance.

Bending moment formulae should be derived from first principles so there is a clear understanding of how they were obtained. This is not a formula to memorize blindly — it's a formula to understand deeply.

Combined stress check:

Once both shear and bending stresses are known, they should be combined using an appropriate failure theory. For ductile materials, the maximum shear stress theory (Tresca) or von Mises criterion are appropriate:

Maximum shear stress (Tresca):

$$f_{s,max} = \sqrt{\left(\frac{f_b}{2}\right)^2 + f_s^2}$$

Von Mises equivalent stress:

$$\sigma_{eq} = \sqrt{f_b^2 + 3f_s^2}$$

Bearing Stress on the Pin

Bearing stress = same material considerations as for eye and fork. Base the design on the weaker material if different materials are used for pin, eye, and fork.

Summary Table: Complete Knuckle Joint Stress Analysis

What Actually Failed on the Tensioning Mechanism

the practitioner ran the numbers under the practitioner's supervision. The results were damning:

"The pin was sized based on shear alone — no bending check at all," she reported. "The bending stress was 2.3 times the shear stress. And the eye tensile stress was at 94% of yield with zero safety margin."

the practitioner nodded slowly. "And the proportions?"

the practitioner checked her sketch against the wreckage measurements. "The eye width a was only 0.4__d__ instead of the recommended 0.75__d__. Someone tried to save material."

"There it is," the practitioner said. "They saved maybe a few units of currency in material. The downtime tonight will be hundreds of times that."

Levers — Ancient Mechanics, Modern Failures

The Third Casualty

The discharge gate lever was the last failure the practitioner examined, and in some ways, the most instructive. A lever is mechanically simple — a beam that rotates about a fulcrum to multiply or redirect force. Archimedes understood levers. Every first-year engineering student studies them.

And yet, this one had buckled.

"How," the practitioner asked, genuinely confused, "does a lever buckle? It's a lever. It bends. It doesn't buckle."

the practitioner smiled grimly. "It buckles when the designer forgets that a bending member also has a compressive flange. And when that compressive flange is too thin relative to its unsupported length, lateral-torsional buckling becomes the governing failure mode."

Lever Design Fundamentals

A lever transmits force and/or motion about a pivot point (fulcrum). Like knuckle joints and couplings, levers can be cast or fabricated. Smaller levers (for example, rocker arms) may also be forged.

The critical design parameters for a lever are:

1. Bending stress is usually the most critical stress

The lever is essentially a beam. The section must resist the bending moment created by the applied forces. Therefore, the lever usually has a cross-section with greater height than width — this maximises the section modulus for a given amount of material.

2. Bearing pressure at the fulcrum is usually the critical design factor

The fulcrum is where the lever pivots. The contact pressure between the lever bore and the pivot pin must be kept within allowable limits to prevent wear and seizure.

In some cases, rolling element bearings (journal bearings) are used. In other cases, plain bearings are used. If wear is not critical, the lever can be designed without a separate bearing. In this case, it is good design practice to fit the boss with grease nipples or oil holes so lubricant can be applied.

3. The section at the boss requires special attention

If the lever has an integral boss (the thickened section around the fulcrum bore), the bending stress may be maximum just outside the boss — where the section transitions from the thick boss to the thinner lever arm. This is a classic stress concentration location.

Lever Cross-Section Selection

The choice of cross-section depends on the application:

Design Insight: A critical design factor is usually the bearing pressure at the fulcrum, particularly if the lever has an integral boss. The bending stress may be maximum just outside the boss where the cross-section reduces. Check both locations.

Lever Stress Analysis

Bending moment at any section:

For a simple lever with force F at distance L from the fulcrum:

$$M = F \times L$$

Bending stress:

$$f_b = \frac{M}{Z}$$

Where Z is the section modulus of the lever cross-section at the point of interest.

For a rectangular section of height h and width w:

$$Z = \frac{wh^2}{6}$$

For an I-section or U-section, the section modulus must be calculated from the full section properties (using the parallel axis theorem if necessary).

Bearing stress at the fulcrum:

$$f_{brg} = \frac{R}{d_{pin} \times L_{bearing}}$$

Where:

• R = reaction force at the fulcrum
• d_pin = fulcrum pin diameter
• L_bearing = bearing length (boss width)

Shear stress in the lever:

For most lever designs, shear stress is secondary to bending stress. However, for short, heavily-loaded levers, shear should be checked:

$$f_s = \frac{V}{A_{web}}$$

Where V is the shear force at the section of interest.

The Lever Failure Explained

The discharge gate lever had been redesigned during a previous plant upgrade. The new design used a thinner section to "save weight" — but the original lever was thick for a reason. The boss bearing was undersized, causing excessive wear that introduced play. The play allowed the lever to oscillate laterally, and the thin compressive flange couldn't resist the lateral-torsional buckling mode.

"Three lessons," the practitioner said, holding up three fingers:

Lesson 1: The lever section must have adequate width as well as height. Height resists bending. Width resists lateral buckling.

Lesson 2: The fulcrum bearing must be properly sized and lubricated. Wear at the fulcrum introduces play, and play introduces dynamic loading that the original design never accounted for.

Lesson 3: If the lever has an integral boss, check the stress at the transition — not just at the point of maximum bending moment. Stress concentrations at geometric transitions are where fatigue cracks initiate.

Pulling It All Together — The Design Process in Practice

A Framework You Can Use Tomorrow

the practitioner and the practitioner spent the rest of that long night developing replacement designs for all three failed elements. Here is the systematic process they followed — and the process you should follow for any machine element design.

Step-by-Step Machine Element Design Process

Step 1: Define the Loading

Step 2: Choose the Manufacturing Method

Step 3: Establish Proportions

Start with accepted good proportions for the element type:

Step 4: Perform Complete Stress Analysis

Check every stress mode for every component:

Step 5: Apply Safety Factors

Step 6: Verify Proportions Against Results

After the stress analysis, check: do the calculated dimensions still look right? Are proportions reasonable? If something looks wrong — a pin that's thicker than the eye, a web that's paper-thin compared to the flange — go back and check your assumptions.

The Worked Examples

Example 1: Knuckle Joint Design

Problem: Design a knuckle joint to transmit an axial tensile force of 50 kN. The material is mild steel with yield stress σ_y = 250 MPa. Use a safety factor of 3.0.

Solution:

Allowable stresses:

Rod diameter:

$$d = \sqrt{\frac{4F}{\pi f_t}} = \sqrt{\frac{4 \times 50{,}000}{\pi \times 83.3}} = \sqrt{764.0} = 27.6 \text{ mm}$$

Round up to d = 28 mm (or use the nearest standard bar size, say 30 mm).

Using d = 30 mm, establish good proportions:

Check eye tensile stress:

$$f_t = \frac{F}{(D - d) \times a} = \frac{50{,}000}{(60 - 30) \times 23} = \frac{50{,}000}{690} = 72.5 \text{ MPa}$$

✅ 72.5 < 83.3 MPa — SAFE

Check eye shear stress:

$$e = \sqrt{\left(\frac{60}{2}\right)^2 - \left(\frac{30}{2}\right)^2} = \sqrt{900 - 225} = \sqrt{675} = 26.0 \text{ mm}$$

$$f_s = \frac{50{,}000}{2 \times 26.0 \times 23} = \frac{50{,}000}{1{,}196} = 41.8 \text{ MPa}$$

⚠️ 41.8 ≈ 41.7 MPa — MARGINAL — Consider increasing eye width to 25 mm

Check eye bearing stress:

$$f_{brg} = \frac{50{,}000}{30 \times 23} = \frac{50{,}000}{690} = 72.5 \text{ MPa}$$

✅ 72.5 < 125 MPa — SAFE

Check pin bending stress:

$$M = \frac{F(a + b)}{4} = \frac{50{,}000 \times (23 + 38)}{4} = \frac{50{,}000 \times 61}{4} = 762{,}500 \text{ N·mm}$$

$$f_b = \frac{32M}{\pi d_p^3} = \frac{32 \times 762{,}500}{\pi \times 30^3} = \frac{24{,}400{,}000}{84{,}823} = 287.6 \text{ MPa}$$

❌ 287.6 >> 83.3 MPa — FAILS!

This is exactly the result the practitioner predicted. The pin bending stress dominates the design. Good proportions give a starting point, but the pin must be enlarged.

Redesign the pin:

$$d_p = \left(\frac{32M}{\pi f_b}\right)^{1/3} = \left(\frac{32 \times 762{,}500}{\pi \times 83.3}\right)^{1/3} = \left(\frac{24{,}400{,}000}{261.7}\right)^{1/3} = (93{,}269)^{1/3} = 45.3 \text{ mm}$$

Use d_p = 46 mm (or nearest standard pin size).

Lesson: The pin diameter from "good proportions" (d_p = d = 30 mm) was woefully inadequate when bending was accounted for. This is precisely why stress analysis cannot be skipped. The pin must be roughly 1.5 times the rod diameter for this load case — not equal to it.

Revised design with d_p = 46 mm:

Now re-check all stresses with the enlarged pin hole in the eye and fork (D must also increase to maintain adequate material around the pin):

Re-check eye tensile stress:

$$f_t = \frac{50{,}000}{(92 - 46) \times 35} = \frac{50{,}000}{1{,}610} = 31.1 \text{ MPa}$$ ✅

Re-check pin bending:

$$M = \frac{50{,}000 \times (35 + 58)}{4} = \frac{50{,}000 \times 93}{4} = 1{,}162{,}500 \text{ N·mm}$$

$$f_b = \frac{32 \times 1{,}162{,}500}{\pi \times 46^3} = \frac{37{,}200{,}000}{305{,}844} = 121.6 \text{ MPa}$$

⚠️ Still exceeds 83.3 MPa. The bending moment increased because the proportions scaled up.

This illustrates a fundamental truth about knuckle joint design: scaling up the pin doesn't automatically fix the problem because the eye and fork widths also scale, increasing the bending moment. The designer must iterate — often increasing the pin diameter while holding eye/fork widths closer to minimum.

Final optimised design (after iteration):