Bearing Fault Frequency Calculator: How to Use It

A bearing fault frequency calculator estimates the characteristic vibration frequencies produced by defects in rolling element bearings, such as inner race, outer race, ball, or cage faults. You input the bearing geometry (pitch diameter, ball diameter, number of balls, contact angle) and shaft speed, and the calculator outputs the four primary fault frequencies in Hertz. These values help you identify the source of vibration in rotating machinery before disassembly.

Formula

The four standard fault frequency formulas for a rolling element bearing are:

• BPFO = (N_s/60) × (n_b/2) × (1 − (B_d/P_d) × cos θ)
• BPFI = (N_s/60) × (n_b/2) × (1 + (B_d/P_d) × cos θ)
• BSF = (N_s/60) × (P_d/(2 × B_d)) × (1 − ((B_d/P_d) × cos θ)²)
• FTF = (N_s/60) × 0.5 × (1 − (B_d/P_d) × cos θ)

Where:

• N_s = shaft speed in revolutions per minute (RPM)
• n_b = number of rolling elements (balls or rollers)
• B_d = ball or roller diameter (same unit as P_d)
• P_d = pitch diameter (mean diameter of the raceway contact circle)
• θ = contact angle in degrees (typically 0° for deep groove ball bearings)

All outputs are in Hertz (cycles per second). The shaft speed in RPM is converted to revolutions per second by dividing by 60.

Worked Example

Using the default values from the calculator: shaft speed = 1800 RPM, ball diameter = 8 mm, pitch diameter = 40 mm, number of balls = 8, contact angle = 0°.

1. Convert RPM to Hz: 1800 / 60 = 30 Hz.
2. Calculate the ratio (B_d/P_d) × cos θ = (8/40) × cos(0°) = 0.2 × 1 = 0.2.
3. BPFO = 30 × (8/2) × (1 − 0.2) = 30 × 4 × 0.8 = 96.0 Hz.
4. BPFI = 30 × (8/2) × (1 + 0.2) = 30 × 4 × 1.2 = 144.0 Hz.
5. BSF = 30 × (40/(2×8)) × (1 − 0.2²) = 30 × 2.5 × 0.96 = 72.0 Hz.
6. FTF = 30 × 0.5 × (1 − 0.2) = 30 × 0.5 × 0.8 = 12.0 Hz.

These match the calculator outputs exactly.

What Each Frequency Means

The four fault frequencies correspond to different physical defects:

• BPFO (Ball Pass Frequency Outer): appears when a defect is on the outer race. It is the rate at which balls pass a fixed point on the outer race. This is the most common fault frequency because the outer race is stationary and often takes the most load.
• BPFI (Ball Pass Frequency Inner): appears when a defect is on the inner race. It is slightly higher than BPFO because the inner race rotates and adds to the relative motion.
• BSF (Ball Spin Frequency): appears when a ball or roller has a flat spot or spall. It is the rotation rate of the rolling element itself around its own axis.
• FTF (Fundamental Train Frequency): appears when the cage (ball separator) is damaged or when there is excessive clearance. It is the rotational speed of the cage assembly.

In a vibration spectrum, these frequencies often appear with harmonics (multiples: 2×, 3×, etc.). The presence of harmonics strengthens the diagnosis. Also, real bearings may show slight variations from the calculated values due to sliding, load distribution, and manufacturing tolerances.

How to Use the Calculator Correctly

To get accurate fault frequencies, you need correct bearing geometry. Here are the steps:

1. Find the bearing number (e.g., 6205) and look up its pitch diameter, ball diameter, and number of balls from a manufacturer datasheet or a bearing dimension table.
2. Measure or estimate the shaft speed. If the machine runs at variable speed, use the speed at which the vibration was recorded.
3. Enter the contact angle. For most radial ball bearings, this is 0°. For angular contact bearings, it may be 15°, 25°, or 40°.
4. Run the calculator and note all four frequencies.
5. Compare these frequencies to peaks in your vibration spectrum. Look for exact matches or near-matches with harmonics.

If you do not have the exact geometry, you can often find standard values online. For example, a 6205 bearing has 9 balls, ball diameter ≈ 7.94 mm, pitch diameter ≈ 39.04 mm. Using those values changes the outputs noticeably.

Common Mistakes and Misconceptions

Several errors can lead to misdiagnosis:

• Unit mix-ups: Using ball diameter in inches and pitch diameter in millimeters will produce nonsense. Keep both in the same unit.
• Confusing RPM and Hz: The formulas require shaft speed in revolutions per second, not per minute. Forgetting to divide by 60 is the most common mistake.
• Ignoring contact angle: For angular contact bearings, the contact angle significantly changes the frequencies. Using 0° when the actual angle is 25° can shift BPFO by 5–10%.
• Assuming the calculated frequencies are exact: Real bearings have slip (typically 1–2%), so the measured peak may be slightly lower than calculated. This is normal.
• Using the wrong number of balls: Some bearings have balls in a cage with a different count than the table says. Double-check.

A common misconception is that BPFO is always the largest peak. In practice, the amplitude depends on load and sensor location, not just the frequency. A small peak at BPFO can still indicate a defect.

Limits of the Formula

The formulas assume ideal geometry and pure rolling motion. In reality:

• Bearings have sliding between balls and races, especially under light loads.
• Defects can be distributed (e.g., waviness) rather than a single spall, producing broadband noise instead of discrete peaks.
• At very high speeds, centrifugal forces change the contact angle.
• The formulas do not account for bearing clearance, preload, or lubrication effects.

Therefore, the calculated frequencies are a starting point. For a reliable diagnosis, use them alongside other techniques like envelope analysis (demodulation), time waveform analysis, and temperature monitoring. The calculator is most useful for planning a vibration analysis or for quickly checking whether a suspected peak matches a bearing fault.

Frequently Asked Questions

Q: What is the difference between BPFO and BPFI?
BPFO is the frequency at which balls pass a defect on the outer race, while BPFI is for a defect on the inner race. BPFI is always higher because the inner race rotates with the shaft, adding to the relative speed between the ball and the defect.

Q: Do I need to know the contact angle?
For most deep groove ball bearings, the contact angle is effectively 0°, so you can enter 0. For angular contact bearings, you must enter the manufacturer’s contact angle (typically 15°, 25°, or 40°) because it affects all four frequencies.

Q: Can I use this calculator for roller bearings?
Yes, the formulas apply to cylindrical roller bearings as well, provided you use the roller diameter and the number of rollers. The contact angle is usually 0° for cylindrical rollers.

Q: Why is my measured vibration frequency slightly lower than the calculated value?
Bearing slip is the main reason. Balls do not roll perfectly without sliding; under normal operating conditions, slip of 1–2% is typical. This shifts the actual fault frequencies slightly downward.

Q: What if I don’t know the bearing geometry?
You can find the geometry from the bearing number in manufacturer catalogs or online bearing databases. If you cannot find it, you can measure the pitch diameter and ball diameter using calipers, but this requires disassembly.

Q: Does the calculator work for variable speed machines?
Yes, but you must enter the shaft speed at the moment the vibration was measured. If the speed changes, the fault frequencies change proportionally. Use order tracking (normalizing to shaft speed) for variable speed machines.

Related

• Calculator Voltage: Typical Battery and Solar Power
• Partial Pressure: Formula and Calculation Example
• 3 Phase Power Calculation: Formula and Examples

in n | Shaft speed | RPM | 1800 | 0
in bd | Ball diameter | mm | 8 | 0
in pd | Pitch diameter | mm | 40 | 0
in nb | Number of balls | - | 8 | 0
in ca | Contact angle | degrees | 0 | 0
out bpf | Ball pass frequency outer (BPFO) | Hz | (n/60) * (nb/2) * (1 - (bd/pd) * cosd(ca)) | 1
out bpfi | Ball pass frequency inner (BPFI) | Hz | (n/60) * (nb/2) * (1 + (bd/pd) * cosd(ca)) | 1
out bsf | Ball spin frequency (BSF) | Hz | (n/60) * (pd/(2*bd)) * (1 - ((bd/pd) * cosd(ca))^2) | 1
out ftf | Fundamental train frequency (FTF) | Hz | (n/60) * 0.5 * (1 - (bd/pd) * cosd(ca)) | 2