Crystal Instruments Spider - FFT Spectrum or Constant Bandwidth Spectrum

Spider DSA User’s Manual
268
contrast, if the rotating speed is at 6000 RPM, the user probably won’t care if the
instrument can discriminate the measurement between 100Hz (order 1) and
101Hz.
With the digital resampling technique, the order tracks and order spectrum are
extracted based on a filter with equal ΔOrder instead of equal ΔFrequency. The
concept is illustrated in the following figure:
RPM
Frequency
of Order
Frequency Band
used to extract
the Orders
RPM
Frequency
of Order
Frequency Band
used to extract
the Orders
Constant Band Tracking using
Regular FFT method
Order Tracking using Digital
Resample method
Figure 178: Comparison of constant band tracking and digital re-sampling method.
The left figure shows when the order tracks are extracted using conventional FFT
method with fixed resolution, the ΔFrequency of the tracking filter will be fixed;
the right figure illustrates that if the order tracks are extracted using digital
resampling, the ΔFrequency tracking filter will be increased proportionally with
the RPM. Obviously, the method of digital resampling is more desirable in
extracting the measurement of orders.
FFT Spectrum or Constant Bandwidth Spectrum
Constant bandwidth spectra (FFT Spectra) remain an important tool for the
diagnosis of mechanical problems. While an order spectrum is ideal to track the
behavior of a machine’s forcing functions, and FFT is ideal for tracking its
resonances. Further, various 3-D visualization tools make it easier to understand
the interplay of these two machine characteristics. For example, a series of FFT
Spectra can be cascaded (by time or by RPM) in a 3D waterfall as shown below: