Detection of momentary knocking noises

Ask questions relating specifically to the PicoDiagnostics NVH equipment and software here.
Post Reply
Steve Smith
Pico Staff Member
Pico Staff Member
Posts: 1126
Joined: Sun Aug 25, 2013 7:22 am

Detection of momentary knocking noises

Post by Steve Smith »

Picking up on the following forum post topic13221.html

I thought it would be an idle opportunity to demonstrate the challenges we face when diagnosing momentary knocking noises using the Time Domain feature of our NVH software.

The following video highlights a simple experiment using 4 accelerometers to capture a simulated momentary “knock” induced using a screwdriver handle.

https://drive.google.com/file/d/1uyLMIo ... sp=sharing

The accelerometers are attached to differing sections of the engine bay using the magnetic mounts as follows:
Channel A (Blue) RH Chassis
Channel B (Red) RH Slam Panel (Framework)
Channel C (Green) LH Slam Panel (Bolt on, bonnet locking platform)
Channel D (Yellow) LH Chassis
Accelerometer locations
Accelerometer locations
Momentary knocking noises produce a lot of energy (released as sound) for a very short time period. Mistakenly in the past I have wrongly assumed that the accelerometer closest to the source of the knocking noise will produce the greatest amplitude, this is not the case! (We live and learn together)

Accelerometer amplitude is dependent upon numerous factors such as the properties of the material to which it is mounted, the shape, the formation and other components that may also be attached.

A recent example from a vehicle manufacturer describes a brake clicking noise, where the brake disc “back plate” exhibited the highest amplitude even though the noise developed in the hub assembly. The thin/flexible brake disc “back plate” behaves more like an antenna for the noise where measuring amplitude alone would lead you to believe the “back plate” was the offending component!

Returning now to the experiment in the video above, the NVH Time Domain view captures the deliberate strike/knock adjacent to each accelerometer in the following order, Channel A, B, C, D, C, B & back to A .

During the NVH set up process I have informed the software that the engine speed is fixed at 1000 rpm as this is not relevant given we are looking for a momentary knock noise. (No OBD connector is required)

I have also informed the software that microphones are connected and not accelerometers. See post91701.html#p91701

This allows the added bonus of listening to the momentary knock during playback whilst not affecting amplitude comparison.
NVH Set-up
NVH Set-up
The NVH Time Domain view below reveals the response from each accelerometer when the bodywork is struck adjacent to each accelerometer. (Denoted by the purple boxes)
Time Domain Strike Points
Time Domain Strike Points
It is clear to see how the amplitude of each accelerometer is at its largest when the bodywork receives a strike from the screwdriver handle in the immediate vicinity.

With that said, the first and final strike was adjacent to the accelerometer attached to the RH Chassis (Channel A Blue)
I have drawn a turquoise box around Channels A & B at this point in time as the strike adjacent to the accelerometer attached to the RH Chassis (Blue) is also detected by the accelerometer attached to the RH Slam Panel Framework (Red).
This is hardly surprising given they are in reasonable proximity to one another but could you say which one is the primary offender?

We can “right click” and hide channels so as to view the amplitude from each accelerometer in greater detail. This however reveals how the accelerometer attached to the RH Slam Panel Framework responds with a greater amplitude than the accelerometer attached to the RH Chassis which is receiving the strike from the screwdriver handle!
Right Click Channel In View
Right Click Channel In View
Remember the amplitude is governed by the properties of the material to which it is mounted, the shape, the formation and other components that may also be attached.

In the scenario above the RH Slam Panel Framework is far more flexible/thinner than the rigid RH Chassis member and therefore it is not surprising how a strike to the RH Chassis is immediately transmitted up to the Slam Panel Framework which is far more influenced by the energy contained in the strike. The result of which is a greater response/amplitude recorded by the accelerometer attached to the Slam Panel Framework

Whilst amplitude can mislead you during diagnosis of momentary knocking noise, the response of each accelerometer cannot!

I think it is safe to say that a strike adjacent to the accelerometer attached to the RH Chassis will be detected momentarily by this accelerometer before the remaining accelerometers. Therefore the first responding accelerometer is the closest to the source of the noise.

Using “time” measurements to our advantage, what could be better than NVH when it comes to accurately measuring response time of accelerometers? “The Time Domain in PicoScope.”

With all accelerometers connected as they were during the test above, NVH/Pico Diagnostics was closed and PicoScope opened.

Each channel was set to +- 1 V range and AC coupled at 2 s/div. A repeat of the strikes to the bodywork was again carried out in the following sequence, Channel A, B, C, D, C, B & back to A . (Refer to video)
The image below refers to the first strike adjacent to the accelerometer attached to the RH Chassis (Channel A blue)
PicoScope Capture
PicoScope Capture
No surprises above that the first responding accelerometer is the one located at the RH Chassis (Blue) followed by the accelerometer attached to the RH Slam Panel Framework. This now separates these two accelerometers in terms of “time” (417 micro-seconds) rather than amplitude which we could not distinguish using the Time Domain view in our NVH software.

If this were a knocking noise we were chasing, we could determine the RH Chassis accelerometer to be closest to the source of the knocking noise

Following this through we reveal more information about how the energy contained in the strike to the RH chassis dissipates through the vehicle.

The frequency of the knock/sound is greater at the strike point (RH Chassis) yet the amplitude of the oscillations (post-strike) are greater at the RH Slam panel Framework suggesting there may be some form of bounding in this thinner bodywork.

Ironically the next responding accelerometer after the RH Slam Panel Framework (Red) is the LH Chassis (Yellow) and not the LH Slam Panel Locking platform (Green). I am not sure of the explanations here but could we surmise the chassis assembly is one rigid unit, hence the strike delivered to the RH Chassis is also felt in the LH Chassis? The accelerometer response time between RH and LH Chassis members was 649 micro-seconds!

The final responder was the LH Slam Panel Locking Platform (Green) with minimal oscillations in comparisons to the RH Slam Panel Framework. Again could we surmise this is because of the additional strength gained from the “bolt on” Locking Platform, preventing “flex” and so reducing bounding?

Food for thought.

I have posted both files here for you to listen, view and measure. The psdata file contains multiple strikes where you can review how the energy travels throughout the vehicle dependent upon the strike location
TEST 1 MIC TD.pddata
Accelerometers set as Mic's
(1.92 MiB) Downloaded 271 times
TEST 3 PS TD.psdata
PSdata file
(3.88 MiB) Downloaded 280 times
I hope this helps, take care…….Steve

Post Reply