Rolling-Element Bearings: Selection, Fits and Installation: The Four Bearing Types You Actually Need to Know

Quick Revision

• Design life equation: L = (60 × N × h) / 10⁶ — converts hours to millions of revolutions
• Ball bearing life: L₁₀ = (C/P)³
• Roller bearing life: L₁₀ = (C/P)^(10/3)
• Average bearing life ≈ 5 × L₁₀ life
• Three life methods: Basic L₁₀ (load only) → Adjusted Lna (+ reliability, material, lubrication) → New Life Lnaa (+ fatigue limit, contamination)
• Adjusted life formula: Lna = a₁ × a₂₃ × L₁₀
• New life formula: Lnaa = a₁ × a_the bearing supplier × L₁₀
• Viscosity ratio: κ = ν/ν₁ — higher κ means better lubrication and longer life (cap at κ = 4)
• Contamination factor ηc: 1 = very clean, 0.8 = sealed/greased for life, 0.5 = shielded, 0.1–0.5 = contaminated, 0 = heavily contaminated
• If Fa/Fr ≤ e: simpler formula applies (usually P = Fr or P = Fr + Y₁·Fa)
• If Fa/Fr > e: more complex formula with combined load factors applies
• Static load checks: Deep groove ball: Po < Co; Roller bearings: Po < Co/1.5
• Minimum radial load: Ball bearings Fr > 0.01·C; Roller bearings Fr > 0.02·C
• Always check: shaft speed does not exceed bearing speed rating
• 37 standard bearing types — selection based on load direction, speed, accuracy, noise, friction, self-alignment
• Key factors affecting life beyond load: steel type, lubrication quality, and cleanliness

The Four Bearing Types You Actually Need to Know

Out of the 47+ types of rolling element bearings catalogued by major manufacturers, four types cover the vast majority of industrial applications. Here's what separates them:

Key Insight: Ball bearings use a life exponent of 3 in calculations. Roller bearings use 10/3 (≈ 3.33). This seemingly small difference has a massive impact on predicted bearing life. Get the exponent wrong, and your entire calculation is meaningless.

the practitioner had used a deep groove ball bearing in an application that experienced significant shaft deflection. The bearing couldn't self-align. It was fighting the misalignment with every revolution — and it lost.

The Core Formula You Must Memorize

This is the equation that forms the foundation of every bearing selection:

Basic Rating Life (L₁₀):

L₁₀ = (C / P)^p

Where:

• L₁₀ = Basic rating life in millions of revolutions
• C = Dynamic load rating of the bearing (from manufacturer tables, in kN)
• P = Equivalent dynamic bearing load (calculated, in kN)
• p = Life exponent:
• • p = 3 for ball bearings
    • p = 10/3 for roller bearings

Converting Design Life from Hours to Millions of Revolutions:

L = (60 × N × h) / 10⁶

Where:

• L = Design life in millions of revolutions
• N = Average operating speed in rev/min (RPM)
• h = Design life in hours

Putting Numbers to It: A Quick Reality Check

Let's say you need a bearing to last 10 years on a machine running 8 hours per day, 300 days per year, at 1,500 RPM.

Step 1: Calculate design life in hours:

h = 10 × 300 × 8 = 24,000 hours

Step 2: Convert to millions of revolutions:

L = (60 × 1,500 × 24,000) / 10⁶ = 2,160 million revolutions

Step 3: If you've selected a deep groove ball bearing with C = 52.7 kN and your calculated load P = 8 kN:

L₁₀ = (52.7 / 8)³ = (6.59)³ = 286 million revolutions

Result: The bearing's L₁₀ life is 286 million revolutions, but you need 2,160 million revolutions. This bearing will fail long before your design target. You need a bigger bearing or a different type entirely.

This is exactly the mistake the practitioner's plant made. They picked a bearing that fit the shaft — but never checked whether it could survive the required operating life.

Method 1: The ISO L₁₀ Life (The Conservative Baseline)

This is the simplest and most widely used method. It considers only the loads on the bearing:

L₁₀ = (C / P)^p

Pros: Simple, universally understood, conservative. Cons: Ignores lubrication, material quality, and contamination. Often dramatically underestimates actual bearing life in clean, well-lubricated applications.

Method 2: The ISO Adjusted Rating Life (The Refined Estimate)

The ISO recognized that real-world factors massively influence bearing life. In 1977, they introduced adjustment factors:

Lₙₐ = a₁ × a₂₃ × L₁₀

Where:

• a₁ = Reliability adjustment factor (a₁ = 1 for standard 90% reliability)
• a₂₃ = Combined factor for material and operating conditions (lubrication, viscosity)

How a₂₃ is determined:

1. Calculate the mean bearing diameter: d_m = 0.5 × (bore + outside diameter)
2. Determine the required lubricant viscosity (v₁) for your speed and bearing size
3. Calculate the viscosity ratio: κ = v / v₁ (where v = actual viscosity of your lubricant)
4. Look up a₂₃ from the manufacturer's diagram using κ

Key Takeaway: With proper lubrication (κ ≈ 3), you can double or triple the predicted bearing life compared to the basic L₁₀ calculation.

Method 3: The the bearing supplier New Life Theory (The Game Changer)

This is the most advanced method, and it introduced a concept that changed bearing engineering: the fatigue load limit (Pᵤ).

The revolutionary idea: if the load on a bearing stays below the fatigue load limit, and lubrication and cleanliness are adequate, fatigue will not occur. The bearing could theoretically last indefinitely.

Lₙₐₐ = a₁ × a_the bearing supplier × L₁₀

Where a_the bearing supplier is determined from manufacturer diagrams using:

• The viscosity ratio κ
• The contamination factor η_c
• The ratio η_c × Pᵤ / P

The contamination factor (η_c) is critical:

The Three Methods Compared: Same Bearing, Drastically Different Results

Here's what the practitioner found when she ran all three methods on the exact bearing that failed — a 6309 deep groove ball bearing on a 45 mm shaft, running at 5,000 RPM with an 8 kN radial load (no axial load), lubricated with oil at 20 mm²/s viscosity:

Read those numbers again. The same bearing, in the same application, can have a predicted life ranging from 286 to 2,288 million revolutions depending on which method you use and what conditions you account for.

The lesson: The basic L₁₀ method is intentionally conservative. Under clean, well-lubricated conditions, your bearing will likely last far longer than L₁₀ predicts. But under contaminated conditions, even the advanced methods show a dramatic life reduction.

Hidden Killer #1: The Wrong Steel (And You'd Never Know)

All ISO bearing life equations assume standard bearing steel. Most commercial bearings meet or exceed this standard, so factor a₂ (material factor) is typically 1. But here's the nuance: premium bearings from top manufacturers often use superior steels that significantly outperform the ISO baseline.

Your action item: When comparing bearing brands at similar prices, the difference is often in the steel. Premium steel = longer life. This is baked into the manufacturer's specific life calculations, which is why the bearing supplier's method often predicts longer life than the generic ISO method.

Hidden Killer #2: Lubrication (The #1 Cause of Premature Failure)

Lubrication is the single most impactful factor in bearing life. The lubricant creates a thin film between the rolling elements and raceways. When that film breaks down, metal touches metal, and failure accelerates rapidly.

Critical lubrication factors to monitor:

• Type: Oil or grease? Each has different viscosity characteristics.
• Viscosity at operating temperature: This is NOT the viscosity on the bottle. Viscosity drops dramatically as temperature rises.
• Additives: Some additives improve extreme-pressure performance.
• Replenishment schedule: Grease degrades over time. "Lubricated for life" means the bearing's life, not forever.

The viscosity ratio (κ) formula:

κ = v / v₁

Where:

• v = Actual kinematic viscosity of lubricant at operating temperature (mm²/s)
• v₁ = Required minimum viscosity for the bearing size and speed (from manufacturer diagrams)

Target: κ ≥ 1. If κ < 1, your lubricant film is too thin. If κ > 2, you're in excellent territory and bearing life increases substantially.

Fun fact: The standard grease used in pre-lubricated deep groove ball bearings typically has a viscosity of 100 mm²/s at 40°C. But at 80°C operating temperature, that same grease might only deliver 20–30 mm²/s. Always check viscosity at operating temperature, not ambient.

Hidden Killer #3: Contamination (The Silent Assassin)

Even microscopic particles can devastate a bearing. Every time a particle passes between a rolling element and the raceway, it creates a tiny dent. Those dents become stress concentrators. Stress concentrators become cracks. Cracks become spalling. Spalling becomes failure.

Contamination factors include:

• Dirt, dust, and metallic particles entering the bearing
• Water or liquids contaminating the lubricant or causing corrosion
• Poor filtration of circulating oil systems
• Inadequate sealing of the bearing housing

the practitioner's investigation revealed that the failed bearing had no seals. The conveyor operated in a dusty environment, and fine particles had been grinding away at the raceways for months. With an η_c of approximately 0.2, the bearing's effective life was reduced by over 80% compared to clean conditions.

The Step-by-Step Selection Process (Your Cheat Sheet)

After her deep dive, the practitioner created a standardized selection procedure that her entire engineering team now follows. Here's the process, generalized for any of the four main bearing types:

For Deep Groove Ball Bearings

Step 1: Define the required bearing life in operating hours (h). Rule of thumb for mechanical design: plan for 10 years of operation.

Step 2: Convert to millions of revolutions:

L = (60 × N × h) / 10⁶

Step 3: Determine the average radial load (Fr) and axial load (Fa) on the bearing. In most cases, these are the shaft reaction forces. Apply shock factors if dynamic loads are present.

Step 4: Calculate the ratio Fa/Fr.

Step 5: Using the shaft size, select a candidate bearing from manufacturer tables. Note the static load rating (C₀) and dynamic load rating (C).

Step 6: Calculate the ratio Fa/C₀ and read the value of e from the manufacturer's graph.

Step 7: Calculate the equivalent dynamic bearing load (P):

• If Fa/Fr ≤ e: P = Fr
• If Fa/Fr > e: P = X·Fr + Y·Fa where X = 0.56 and Y is read from the Fa/C₀ graph

Step 8: Calculate L₁₀:

L₁₀ = (C / P)³

Step 9: Compare L₁₀ to your required life L:

• If L₁₀ < L → bearing is too small, select a larger one
• If L₁₀ ≈ L → bearing is acceptable
• If L₁₀ >> L → bearing may be oversized (consider a smaller one to save cost)

Step 10: Calculate the equivalent static load: P₀ = 0.6·Fr + 0.5·Fa (if P₀ < Fr, use P₀ = Fr). Verify that P₀ < C₀.

Step 11: Check minimum radial load: Fr > 0.01 × C (rule of thumb for satisfactory operation).

Step 12: Verify shaft speed doesn't exceed the bearing's speed rating.

For Cylindrical Roller Bearings

Follow the same steps, but note these differences:

• Fa/Fr ratio should not exceed 0.5 (cylindrical rollers have limited axial capacity)
• Life exponent changes to 10/3 instead of 3:

L₁₀ = (C / P)^(10/3)

• For bearings without flanges (type NU): P = Fr (they cannot carry axial load)
• For bearings with flanges: use the specific X and Y factors for the bearing series
• Minimum radial load rule of thumb: Fr > 0.02 × C

For Spherical Roller Bearings

• Life exponent: 10/3
• Static load check: P₀ < C₀/1.5
• Static load equation: P₀ = Fr + Y₀·Fa
• These bearings a desktop spreadsheet application where misalignment is present, but check that the axial load doesn't exceed: Fa < 3 × B × d (for tapered bore versions with adapter sleeves)

Where:

• B = width of the bearing (mm)
• d = internal diameter of the bearing (mm)

The Decision Framework: Which Bearing Type Do You Actually Need?

the practitioner eventually turned her learning into a decision matrix that she shared with her entire department:

Performance Comparison at a Glance

(★ = poor, ★★★★ = excellent)

The Complete Formulas Reference Card

Save this section. Print it. Pin it to your wall. These are the equations you'll use over and over.

Design Life Conversion

L = (60 × N × h) / 10⁶

Basic Rating Life (L₁₀)

Ball bearings:

L₁₀ = (C / P)³

Roller bearings:

L₁₀ = (C / P)^(10/3)

ISO Adjusted Rating Life

Lₙₐ = a₁ × a₂₃ × L₁₀

Reliability adjustment factors (a₁):

Notice: If you need 99% reliability instead of the standard 90%, your design life drops to just 21% of the L₁₀ value. Higher reliability demands come at a steep cost.

the bearing supplier New Life Theory

Lₙₐₐ = a₁ × a_the bearing supplier × L₁₀

Where a_the bearing supplier is read from manufacturer diagrams as a function of:

• Viscosity ratio (κ)
• Contamination-adjusted load ratio: η_c × Pᵤ / P

Equivalent Dynamic Bearing Load (P)

Deep groove ball bearings:

• Fa/Fr ≤ e: P = Fr
• Fa/Fr > e: P = 0.56·Fr + Y·Fa (Y from tables based on Fa/C₀)

Self-aligning ball bearings:

• Fa/Fr ≤ e: P = Fr + Y₁·Fa
• Fa/Fr > e: P = 0.65·Fr + Y₂·Fa

Cylindrical roller bearings (with flanges):

• Fa/Fr ≤ e: P = Fr
• Fa/Fr > e: P = 0.92·Fr + Y·Fa

Spherical roller bearings:

• Fa/Fr ≤ e: P = Fr + Y₁·Fa
• Fa/Fr > e: P = 0.67·Fr + Y₂·Fa

Equivalent Static Bearing Load (P₀)

Ball bearings:

P₀ = 0.6·Fr + 0.5·Fa (if P₀ < Fr, use P₀ = Fr)

Spherical roller bearings:

P₀ = Fr + Y₀·Fa

Cylindrical roller bearings:

P₀ = Fr

Minimum Radial Load (Rules of Thumb)

Engineering takeaway

Six months after that midnight phone call, the practitioner presented her findings to the entire engineering department. She stood in front of 40 engineers and said something that stuck with everyone in the room:

"We didn't have a bearing failure. We had a knowledge failure. The bearing did exactly what the physics predicted it would do. We just never bothered to run the calculation."

Here's what changed at her plant — and what should change at yours:

1. Every bearing selection now includes a documented L₁₀ calculation. No exceptions. Even for "standard" applications. The calculation takes 10 minutes. A catastrophic failure takes months to recover from.

2. Lubrication is monitored, not assumed. They now track lubricant viscosity at operating temperature and maintain a target κ ≥ 1 for every bearing in the system.

3. Contamination is treated as the enemy it is. Sealed bearings are used wherever possible. Housings in dusty environments were retrofitted with improved seals. Oil filtration systems were upgraded.

4. The right method is used for the right situation. The basic L₁₀ method for initial sizing. The adjusted life method for design verification. The the bearing supplier new life method when contamination needs to be modeled.

5. They design for 10 years, minimum. Because the cost of oversizing a bearing by one increment is negligible compared to the cost of a single unplanned shutdown.

Your Next Step

Look at the most critical rotating equipment in your facility or in your current design project. Can you answer these three questions right now?

1. What is the calculated L₁₀ life of the bearings in that equipment?
2. What is the viscosity ratio (κ) of the lubricant at actual operating temperature?
3. What is the contamination level (η_c) of the operating environment?

If you can't answer all three, you're where the practitioner was before that midnight phone call. The good news is that you now have every formula, every method, and every decision framework you need to fix that.

Don't wait for the phone call. Run the numbers today.

Have a bearing selection challenge or a war story of your own? Drop it in the comments below. The engineering community learns fastest from shared failures — because every failure somebody else shares is one you'll never have to repeat.

Bookmark this post. The next time you're staring at a bearing catalogue at 2 AM wondering if that 6309 can handle the load, you'll thank yourself.