Bearing Defect Frequencies: BPFO, BPFI, BSF and FTF are the four characteristic vibration frequencies a rolling-element bearing generates once a localized fault develops on one of its working surfaces. Each defect frequency points to a specific component, and because they are derived from the internal geometry of the bearing rather than from a whole number of shaft revolutions, they appear at non-integer multiples of running speed. Recognizing them in a spectrum is the core skill behind rolling-element bearing diagnostics.
A defect creates a repetitive impact every time a rolling element passes over it, or every time the defect passes through the load zone. The repetition rate is fixed by ball count, ball and pitch diameter, contact angle and shaft speed. The four frequencies are:
With N rolling elements, ball diameter d, pitch diameter D, contact angle theta, and shaft rotational frequency fr (in Hz), the four frequencies are:
A useful sanity check: BPFO + BPFI = N x fr exactly, and BPFO = N x FTF. If the geometry is unknown, the ratios can be estimated (BPFO is roughly 0.4 x N x RPM and BPFI roughly 0.6 x N x RPM), but published bearing tables or the manufacturer part number give exact values.
A 6205 has N = 9 balls, d = 7.94 mm, D = 39.04 mm and a contact angle of 0 degrees. Running at 1800 RPM (fr = 30 Hz), the calculated frequencies and their multiples of running speed are shown below. These multipliers are constants for the bearing; only the absolute Hz values change with speed.
| Frequency | Component indicated | Multiple of RPM (6205) | Value at 1800 RPM (Hz) |
|---|---|---|---|
| BPFO | Outer race defect | 3.585 x | 107.5 |
| BPFI | Inner race defect | 5.415 x | 162.5 |
| BSF | Rolling element defect | 2.357 x | 70.7 |
| FTF | Cage / train | 0.398 x | 11.9 |
Note that none of these land on 1x, 2x or 3x running speed. That non-synchronous behavior is exactly what separates a bearing fault from imbalance, misalignment or looseness, which sit on integer harmonics.
Because each equation contains the ratio d/D, the multipliers are non-integer values rather than whole numbers. This has two practical consequences. First, defect lines never coincide exactly with the 1x running-speed family, so they can be isolated even when a machine also has imbalance. Second, a small amount of skidding under light load shifts the real frequencies slightly below the calculated values, so analysts allow a tolerance of a few percent when matching a peak. Sidebands spaced at running speed (around BPFI) or at FTF (around BSF) confirm that a defect is moving through the load zone.
Early bearing impacts are low in energy but excite the bearing and structure at their high natural frequencies, ringing in the 1 to 20 kHz range. On a raw velocity spectrum this energy is buried in the noise floor, so the defect frequencies are effectively invisible at first. Envelope detection, also called demodulation or the envelope spectrum, solves this. The signal is band-pass filtered around the resonant ringing zone, then rectified and low-pass filtered to recover the envelope of the impact bursts. An FFT of that envelope reveals BPFO, BPFI, BSF or FTF and their harmonics long before they appear in the ordinary spectrum. Envelope analysis is a standard part of vibration spectrum analysis, and it pairs well with high-frequency techniques such as the shock pulse method for the very earliest detection.
Bearing degradation follows a recognized four-stage progression, and knowing the stage tells you how much warning time remains:
Identifying a Stage 2 outer-race defect is only useful if it drives a work order before the bearing reaches Stage 4. Teams using Fabrico attach the measured defect frequency, stage and trend to the asset record, so the condition-monitoring reading automatically raises a planned replacement task with the correct bearing part number and torque spec. That closes the loop between diagnosis and the shop floor. Understanding calculated service life through L10 bearing life helps set sensible re-inspection intervals once a defect is confirmed. To see the workflow end to end, Book a Fabrico demo.
An inner-race defect rotates in and out of the load zone once per shaft revolution, so its impacts are modulated by running speed and spread energy into sidebands. It also sits deeper in the load path. This often makes inner-race faults harder to detect early but produces the characteristic running-speed sidebands around BPFI once established.
A single defect on a rolling element contacts both the inner and outer race once per element revolution, generating two impacts per turn. That is why the ball spin fault frequently shows up strongest at twice BSF rather than at the fundamental, usually flanked by FTF sidebands.
Yes, approximately. BPFO is close to 0.4 x N x RPM and BPFI close to 0.6 x N x RPM, so even with only the ball count you can search the right region of the spectrum. For confirmation, look up the exact defect multipliers from the manufacturer or a bearing database, since a few percent error can cause you to misread a peak.
It varies with speed, load and lubrication, but envelope detection typically flags a defect in Stage 2, well before the fault is visible in a standard velocity spectrum. That commonly translates into weeks or months of planning time, which is the entire value of catching bearings early rather than reacting at Stage 4.
Anmeldung
Blog
