| Vehicle details: | BMW 320D M Sport xDrive Touring (RH-Drive) |
| Engine code: | N47 2.0 Litre 4-Cylinder |
| Year: | 2013 |
| Symptom: | NVH |
| Author: | Steve Smith |
The customer reports that a knocking noise can be felt and heard through the steering column when rocking the steering wheel back and forth. Although the symptom has been present for about six months, it does not affect the performance or handling of the vehicle. Our customer also reports an intermittent knocking noise from the suspension when driving at low speeds over speed humps.
Verifying the customer complaint is an essential step in the diagnostic process and, luckily, the above steering knock could be replicated with a stationary vehicle. An interesting point to note here is that the knocking noise was evident only with the engine off, so without the intervention of Electric Power Steering (EPS)! With the engine running and EPS active, no knocking noise could be produced.
The following video demonstrates the symptom:
Given the vehicle is equipped with Stop/Start, there are times during a typical journey when the EPS is inactive during engine cut (while stationary). Any input into the steering wheel during this standby period would generate a knock before engine restart commanded by an EPS load request signal (prompted by the movement of the steering wheel).
You may well ask: Why is this an issue? You could argue it's simply wear and tear given the mileage of the vehicle. There may be some merit in that question but with complaints of knocking noises surrounding steering, we must err on the side of caution. This may be an early indication of something more sinister such as a worn steering column intermediate shaft universal joint (UJ) or interconnecting spline, both of which have safety implications. Road-testing the vehicle confirmed no drivability issues such as excessive play or tight spots in the steering system and the self-centring action to be functioning as normal!
Regarding the additional customer complaint of an intermittent knocking noise from the suspension over speed humps, occasionally I could produce a similar knock to that felt/heard from the steering column as the right-hand front wheel rides over the speed hump.
Here's a recording of the knock captured using our NVH software:
To confirm my opinion, the best course of action is to resolve the steering knock and retest over speed humps after a steering fix.
With the customer complaint verified, the Vehicle’s ID and Specification were confirmed. No additional or non-genuine accessories were installed that could influence the performance of the steering or suspension
The Customer Interview highlighted typical tire replacements and wheel alignments associated with the mileage covered by the vehicle but no repairs carried out to the steering system.
The Basic Inspection confirmed no visible signs of damage to suspension or steering components including alloy wheels and tires.
Before diving in here, it is essential to take a step back and check for Technical Bulletins (recalls, campaigns and so on). This is where we struck gold, as a swift Google search accompanied by numerous YouTube videos demonstrated our exact steering concern.
This forum post describes both the fault above, parts required and the journey taken to resolve the concern along with a relevant BMW Dealer Bulletin No.22/2018. Ironically, I could not find this information on two major technical information sites!
On this occasion there is no need to list possible causes based on the information above. However, we must bear in mind that we also have a complaint of a suspension knock when driving over speed humps.
The action plan is mostly governed by accessibility, probability and cost.
Once again, no action plan is required as we know the fault condition and the repair required. However, why not gather objective data to qualify the fix suggested in the Dealer Bulletin while also trying to capture our intermittent speed hump knock?
For those lucky enough to own an NVH Advanced kit (with a 4-channel PicoScope) we have the distinct advantage of placing an accelerometer or microphone at strategic points about the vehicle in order to locate the source of our offending knocking noise. This video describes this technique using both our NVH software and, more importantly, the deeper analysis features of the PicoScope time domain.
We cover the same procedure here on our forum. Below we apply the technique of connecting accelerometers at differing points about the front of the vehicle, but we inform the NVH software we have connected microphones! This enables playback within the NVH software to ensure we have indeed captured an offending knock noise rather than exporting the data and listening to the accelerometers in a media player. More information can be found at Listening to Accelerometers or Exporting accelerometers to .wav audio files.
Accelerometers mounted vertically and installed to the following locations measuring the y axis only:
First impression of this capture: It would appear we have an excessive response from the RH and LH front suspension as the vehicle rides over the speed humps. This is to be expected, as here we witness the characteristic behaviour of the suspension in relation to the chassis. (Note: The road wheels will rise and fall within the wheel arch.) I have placed a marker at 193.22 seconds in this capture to denote the point at which a single knock noise was heard inside the cabin while driving over speed humps.
Note the response of the accelerometers when we hide channels A and C and focus on channels B and D:
If we play back the audio recorded by the accelerometers attached to the LH and RH chassis, we hear the knock, which is more prominent in the RH front chassis. The audio recording at the start of this case study is indeed the data we can see in the red accelerometer capture above!
Based on the above, a further road test was carried out to determine the response of the suspension in relation to the chassis when riding over speed humps. Here we use PicoScope.
Note: Accelerometer positions remain in the identical locations as above.
As we can see above, the red and yellow channels (B and D) respond almost identically, confirming the chassis movement to be equal as the vehicle rides over the speed hump. The black and purple waveforms are math channels derived from subtracting the movement of the suspension from the movement of the chassis (A–B and C–D). What I find interesting here is that the relative suspension motion (of the offending RH side of the vehicle) highlights an increase in both amplitude and frequency relative to the LH side! (Could this be poor suspension damping?)
At this stage (if nothing else) we have qualified the following:
With the vehicle stationary (engine off) and the steering wheel rocked back and forth (EPS inactive) as in the video above, we reposition our accelerometers as below and capture the following data using our NVH software.
Without doubt, based on amplitude alone the accelerometer connected to the RH front subframe responded to the highest level of energy generated by the knock, but amplitude can be deceptive! The mass to which the accelerometers are attached will differ from one location to another on the vehicle. Imagine how an accelerometer would respond to energy travelling along a thin section of steel plate compared to a thicker section of the vehicle chassis of greater mass. In the above example, however, all accelerometers are connected to components of similar mass (chassis and subframe) so the amplitude captured here can be considered a realistic indication of the origin of the knock.
A superior way to qualify amplitude is to measure the response time of each accelerometer to the knocking noise, the first responding accelerometer being the closest to the source of the noise. What better tool to measure response time (with a zoom facility) than PicoScope?
Here we conclude that not only was the accelerometer connected to the RH subframe exposed to the highest energy level, it was also the first to respond to the knock that generated this energy and so closer to the offending component.
To summarize response time, we are evaluating the time taken for the energy (generated by the knock) to propagate throughout the vehicle structure.
It should come as no surprise that a knock originating from the RH front subframe should also be detected 460.2 μs later in the LH front subframe as both accelerometers are connected to the same component! Note, however, that it does take time for the energy to travel along the steelwork and dissipate throughout the vehicle. This phenomenon can be seen in the response time of the red accelerometer connected to the RH chassis which is late to respond to the knock even though the RH front subframe is bolted to the RH front chassis.
We must remain mindful of the above characteristics when locating knocking noises; while we may be convinced in our mind that a knock originates from the location where it can be heard most prominently, it does not mean that is the location of the knock!
I think we can agree at this stage the RH front subframe is our area of concern, so the focus now is to zone in further using multiple accelerometers. The subframe provides multiple mounting points and components on which to “plot” the journey of our knocking noise. One such component is the steering rack assembly
Engine off steering wheel rocked
Once again, the results above indicate the RH front subframe / accelerometer to be exposed to the highest level of energy generated by the knock, but is it the first responder?
Let’s take a closer look, with PicoScope:
Clearly from the above capture, the RH front subframe has the greater amplitude during the knock event.
Here we zoom in to reveal activity before the knock event, which could disrupt the theory surrounding the first responding accelerometer. The accelerometers are detecting the momentary steering input into the steering rack before the generation of the knock. We need to focus on the accelerometer response created by the dramatic increase in energy generated by the knock (referred to as the knock signature) and not prior events!
Focusing on the knock events below and measuring response time of each accelerometer, we can now see the response order has changed!
Given we know the root cause of our steering knock (BMW Technical Bulletin) I thought it would be interesting to understand how and why this knock can be felt in the steering wheel. While we know our knock is generated inside the steering gear between the rack and pinion, we also have a steering column intermediate shaft with two universal joints to consider. We can eliminate the intermediate shaft from our diagnosis using accelerometers:
Engine off steering wheel rocked
Here we have the engine off but steering wheel rotated about 20° back and forth:
Here we have to ignore the first response from the red accelerometer as we are detecting rotation of the intermediate shaft.
The knock signature has been complicated by the rotation of the steering shaft. However, the RH steering rack mounting and RH front subframe are once again early responders, suggesting our knock does not originate with the intermediate shaft:
Now we know that the RH steering rack mounting bolt, RH subframe and steering rack pinion housing are all first responders and we can discard the intermediate shaft and LH side of the chassis/subframe. A final measurement was therefore taken at the RH side of the steering gear and the underside of the steering rack pinion housing.
Note: When attaching the magnetic accelerometer to the aluminum steering gear, we bonded a steel washer in place using Bondloc.
Engine off steering wheel rocked
I think first impressions of the above suggest we have found the epicentre of our knock based on the instantaneous and dramatic response of the accelerometer bonded to the user side of the steering rack pinion housing.
Here we zoom in on this capture to reveal the response times:
I have added the response timings above of all accelerometers beginning with the first responding accelerometer (0 seconds) through to the final responder at 204.9 μs. Note how the accelerometers attached to the steering rack mounting bolts responded simultaneously, as they are attached equally about the steering rack pinion housing.
Based on the collective results and response times of our accelerometer tests throughout this study (in relation to the steering knock) we conclude the following response order:
Following the workshop manual, the steering rack thrust piece was replaced and loaded accordingly to ensure minimal backlash between the steering rack and pinion.
Confirmation of repair is often overlooked or sacrificed due to time constraints. Post-fix measurements provide further objective evidence that protects all parties involved and ensures confidence during the customer handover. Never has it been more important to qualify any repair we carry out than with complaints of noise or vibration as, once again, they are a matter of opinion and without scientific measurements we are in a weak position. When carrying out post-fix confirmation, ensure all variables are kept to a minimum and be sure to match the pre- and post-fix test conditions where possible.
Engine off steering wheel rocked fix
Above we can see a totally different response from the accelerometer bonded to the underside of the steering rack pinion housing. The amplitude has reduced dramatically and the response from the accelerometer demonstrates a uniform sine wave (responding to the steering rack rocking motion) rather than the sporadic and instantaneous acceleration in response to the previous knock. Note the red accelerometer has purposely been attached to the steering column intermediate shaft to denote first motion of the steering as there is no longer a knock signature for reference.
Here I have created a reference waveform of the steering knock captured from the accelerometer bonded to the underside of the steering rack pinion housing pre- and post-fix. I think the waveforms speak for themselves and serve as objective data for our customer.
A video of the steering column knock noise post-fix can be seen below:
Regarding our post-fix handover to the customer, when rocking the steering wheel back and forth without EPS intervention there will inevitably be some form of noise due to the loading of the steering gear from the tire contact with the road surface. The wider the cross-sectional area of the tire, the greater the loading and this must not be confused with the backlash noise creating the steering knock. Capturing pre- and post-data will reinforce the integrity of the repair while enabling you to demonstrate to the customer the difference between them!
While the nature of the steering fault was known from the start, I hope the technique described above will assist with knock detection and highlight a number of analysis features both in Pico Diagnostics (NVH) and PicoScope. In most cases the accelerometers were positioned in a near vertical orientation and only the Y axis was captured/measured, however, any axis could have been used to capture the knock.
The energy created by the knock will travel in all directions, think of ripples on a pond after throwing a stone into the water, the accelerometer would detect the knock in any chosen (X, Y or Z axis) but remember for best results, mount all accelerometers in the same orientation and measure the same axis.
Whilst the steering knock is now cured, the intermittent knock over speed humps remains and so onto another case study!
A big thank you to Mr Martin Rubenstein for his invaluable input on a number of NVH studies over recent years and for qualifying numerous measurement techniques that we can all share and benefit
I hope this helps, take care…..Steve
David M Connelly
April 24 2020
Thank You, Steve. Great Case Study, I will be using this again in the future for noise concerns. Also loved the washer glued to the rack!
Ghislain Leclerc
April 24 2020
Hi
too much step and too long as a diagnostic procedure for such a small problem of knocking in the steering.
This kind of noise is something fairly regular in repair shops