Wheel bearing diagnosis with our NVH solution has been documented in a case study here https://www.youtube.com/watch?v=kLcSib7TWpM&t=642s and covered on our forum here topic17901.html. Having multiple sensors attached to 4 corners of the vehicle does save time and reduce variables between road-tests

The image below highlights a typical capture you are likely to acquire when diagnosing a “humming” wheel bearing (i.e. where the noise is present but not excessive)
Note above we have several stand out peaks of interest on both channel A (OSF blue) and channel D (OSR yellow)

Note also how the scale for our connected sensors indicates dB rather than mg as we have configured our connected accelerometers as microphones to enable playback within the software

Be aware that our TA143 accelerometers will not capture high frequency sounds; we can see in the capture above that the peaks appear to halt at approx. 2.7 kHz

Given we have a “humming” wheel bearing the expected frequencies are going to be between 20 and 200 Hz which is the area of the spectrum where we as humans interpret sound and vibration differently. I.e. one Technician may consider the bearing “hum” to be a vibration, another may describe the “hum” as a noise, neither are wrong!

The bearing noise we all interpret as “noise” is that which is present at high road speeds due to an increase in wheel speed/frequency and of course, accompanying most fundamental frequencies are higher harmonics that give the offending noise its signature sound.

More on harmonics here viewtopic.php?p=105115#p105115 (please scroll down to Question 5 in the link above)

Regardless of the science and theory, if we simply listen to the captured data above, I think we can agree which one is our offending wheel bearing. Using the export feature of our NVH software described here viewtopic.php?p=98084#p98084 we can listen to each bearing using any media player as the exported files are in .wav format

Below we export all channels and include only the highlighted recorded data in the signal history (237.5 to 250 seconds)
Below we have all the recordings with each final letter in the title corresponding to the scope channel
As per our original NVH capture (Image 1) we have considerable noise from both offside (OS) bearings with Channel A (OSF) being the main offender

So why are we covering this topic once again?

Before I go on, please watch the first 13 minutes of this video from the Guru’s at the Mobius Institute and re-join the forum post thereafter
https://www.youtube.com/watch?v=67Et4vbKhOM&t=25s

I chose this bearing for analysis as the vehicle belongs to a close friend of mine who knows his car and is more than alert when it comes to noises etc. (2017 Audi A7 3.0 TDI Quattro)

Given the bearing was diagnosed and replaced soon after first detection of noise, I thought it would serve as a good example to put the theories in the Mobius video to the test. Note: where bearings are severely worn (viewtopic.php?p=91701#p91701 ) NVH will locate and diagnose them but to determine the origin of specific frequencies, we are going to be challenged (Note, there was no axial play in this Audi bearing)

With thanks to Ben Martins who highlighted several bearing frequency calculator websites https://www.skf.com/uk/digital-tools/se ... calculator https://medias.schaeffler.co.uk/en/bear ... calculator https://www.gmnbt.com/resources/calcula ... alculator/ if you know the details of the suspect wheel bearing, you can enter the specifications and obtain the following frequencies of interest:

• Ball pass frequency outer race (BPFO)
• Ball pass frequency inner race (BPFI)
• Rolling element frequency (Ball spin frequency)
• Fundamental train frequency (bearing cage frequency)

So how do the above help with diagnosis?

I guess you could argue they don’t! As mentioned in the Mobius video, once a bearing has failed to the point of becoming audible to the human ear (accompanied with peaks in the spectrum) it is game over for the bearing, all we need to know is which bearing to replace.

However, it maybe you are required to analyse NVH data for training, technical reporting, tech support, warranty purposes, vehicle assessments or legal documentation to present where accusations have been made. In such scenarios, the evaluation of the bearing is all that is required and remember, the above data was acquired simply by attaching an accelerometer to the brake caliper securing bolt!

Returning to the bearing frequency calculator websites and the bearing data you are required to input; you may find identification numbers and the bearing manufacturer name stamped on the inner or outer race. If so, “happy days” as you can visit the bearing manufacturer website and enter the ID numbers which are known to the relevant website
Below is an example of the SKF bearing frequency calculator that provides a detailed description against each required entry field
I have to admit, I did struggle here as I could not acquire the bearing measurements from the ID numbers stamped on the inner race of the Audi bearing and so opted to physically measure the old bearing once removed. Below we have entered the required measurements and rotational speed of the inner race which at a steady 52 mph (see the signal history in image 1) equates to:
11.1 Hz (T1) x 60 = 666 rpm
I have highlighted 2 areas of interest in the results above, “Contact angle” (20°) and Frequency of over rolling point on outer ring (101.422 Hz BPFO)

It would appear that a Deep groove ball bearing (DGBB) cannot have a contact angle greater than 20° when using the SKF site, however when using the Schaeffler site, 45° can be entered which appears to match the pictogram highlighted below in the Schaeffler results which indicates BPFO = 105.1667 Hz
So why are we dwelling on these results?

Let’s take a closer look at the peaks from channel A (OSF blue). Prior to analysis, to hide channels and just listen/analyse channel A, right click in the Frequency view and select “Channels in view” to hide the relevant channel

Notice below we have well defined peaks evenly spaced across the spectrum which appear when the bearing noise can be heard inside the cabin. Can we identify the cause of these peaks? (referred to as “unknown vibration orders”)
Using our harmonic rulers (see here viewtopic.php?p=49451#p49451 ) we place our first ruler at position 1 (which appears to be the fundamental frequency of these unknown vibration orders) and note if the proceeding rulers align with the subsequent peaks. As it so happens, they do! (See below)
So, what is this unknown fundamental vibration order creating all this havoc?

Referring back to both the SKF and Schaeffler bearing calculator reports, our BPFO ranges from approx.101 to 105 Hz! Allowing for tolerances in wheel speeds (T1) and my bearing measurements, could the cause of these unknown vibration orders be a single spall (chip or blemish) on the outer race of the bearing? (See below)
Having examined this bearing assembly thoroughly (both inner and outer hub) there was only a single spall on the outer race of the inner hub bearing, located just after the 12 ‘O’ Clock position as viewed from the side of the vehicle looking into the wheel arch. I.e. One position on the outer race where your fingernail could feel and “stick” as you glide your finger around the race.

As for the BPFO (ball pass frequency outer race) the following video indicates just over 9 ball passes of a fixed point on the outer race as we rotate the inner race of this Audi bearing

https://youtu.be/3ty839fzrhM

With these 9.5 ball passes in mind, given we know the inner race frequency at 52 mph is 11.1 Hz (T1) x 9.5 we have a BPFO of 105.45 Hz which I hope we can agree is close enough (allowing for tolerances) to our unknown fundamental frequency of 106 Hz

Had we not removed the bearing and requested to diagnose only; we could conclude:

Based on the audio recordings, the OSF wheel bearing requires replacement; Why?
Based on the spectral analysis, we suspect the offending unknown vibration peaks/harmonics to be linked to BPFO of the OSF wheel bearing

Recommendation:

Replacement of the OSF wheel bearing with further analysis/attention possibly required to the OSR bearing

In keeping with the typical time constraints surrounding repairs etc. the bearing has been replaced and the vehicle is cured. At this stage no post fix capture data has been taken (sorry) but if the opportunity arises, I would like to remeasure both the OSF and OSR. One possible explanation for the peaks a the OSR was the simple location in relation to the OSF (vibration transfer from the OSF)

One final point to add here, the vehicle in question did exhibit a steering wheel vibration at 52 mph which can be seen in the T1 peak at 11.1 Hz which may have contributed to bearing wear (wheels have since been balanced)

This got me thinking about a comment in the Mobius video about “modulation” which I believe we can see in our capture below. Note the overall appearance of data from channel A (OSF blue) appears to take display sinusoidal properties
Given the capture above includes an imbalanced road wheel, we have 2 additional items to consider.

1. Varying load applied to the bearing and rolling elements attributed to forces from our imbalanced wheel

2. Coinciding harmonics (summing) as T1 will also exhibit harmonics that coincide with BPFO

The image below does confirm that periodically we have tyre vibration orders (T19, T38, and T57) coinciding with BPFO orders (2, 4 & 6) resulting in momentary amplification and jointly, modulation
I hope this helps when trying to decipher the spectrum and the multiple peaks that appear as a result of events (some of which are normal) that occur as the vehicle is in motion

Take care…….Steve