I am having difficulty coming to terms with the Rexnord ZA-2115 bearing technical specifications. This bearing is used in the NASA Bearing Dataset where the Rexnord bearings are tested at 2000 RPM under a 6000 pound load. My questions in particular are:
(A) The Vibration Frequency Fundamental Train, Inner Ring Defect, Outer Ring Defect, and Roller Spin values are specified as 0.0072, 0.1617, 0.1217, 0.0559 cps, respectively. I might assume these are somehow related to Fundamental Train Frequency (FTF), Ballpass frequency outer race (BPFO), Ballpass frequency inner race (BPFI), and Ball spin frequency (BSF).
$\begin{array}{lcl|}
FTF &=&\frac{R}{2}\left(1-\frac{d}{D}\cos\phi\right)\\
BPFO&=&\frac{nR}{2}\left(1-\frac{d}{D}\cos\phi\right)\\
BPFI&=&\frac{nR}{2}\left(1+\frac{d}{D}\cos\phi\right)\\
BSF &=&\frac{DR}{2d}\left[1-\left(\frac{d}{D}\cos\phi\right)^2\right]
\end{array}$
where $n$ is the number of balls, $R$ is rotational speed, and $\phi$ is the contact angle.
But how do I get (and/or should I be able to get) from the Rexnord values to the (FTF, BPFO, BPFI, BSF) 4-tuple? Do I need to multiply Rexnord values by shaft rotational frequency? For example, is 0.0072 x RPM/60 equal to FTF in Hertz?
(B) How can I determine the number of rolling elements in ZA-2115? Since BPFO and FTF only differ by a factor $n$, then I might think that $n$ should equal BPFO / FTF (regardless of RPM). However, when I calculate 0.1217 / 0.0072, I get approximately 16.9 ... shouldn't I get an integer? Or are the values derived empirically using actual test data and a PSD?
Some PSD-Welch calculations suggests peaks at the following locations for $x$ axis
$\begin{array}{cc}
492.95775 \;Hz& 2.1302751\\
985.91549 \;Hz& 3.6660246\\
4201.87793 \;Hz& 1.4599044\\
5164.31925 \;Hz& 1.5721200\\
7018.77934 \;Hz& 1.3078587\\
9061.03286 \;Hz& 1.6893742
\end{array}$
and $y$ axis
$\begin{array}{cc}
492.957746 \;Hz& 1.497\\
985.915493 \;Hz& 1.596\\
5164.319249 \;Hz& 1.024\\
7136.150235 \;Hz& 0.979\\
8990.610329 \;Hz& 2.115
\end{array}$