How does NVH kit Capture a Periodical Knock?

Ask questions relating specifically to the PicoDiagnostics NVH equipment and software here.
Post Reply
User avatar
Posts: 28
Joined: Mon Jun 01, 2015 1:29 am

How does NVH kit Capture a Periodical Knock?

Post by Belet » Sat May 07, 2016 4:03 am

Good morning, all!
Mothers' Day is coming! Enjoy it! 8)
Well, let's get to the point. We know that it's inevitable that knock occurs under certain conditions when engine works. To monitor the knock, a knock sensor is fitted there.

According to the help page popping up when selecting Automotive guided tests/Knock sensor on the software, the frequency of the vibration caused by knocking is about 15kHz.

Supposing knock occurs periodically per 0.2s with 30mg.
Knock occurs periodically per 0.2s, the frequency is 5Hz. But the frequency of knock itself is 15kHz.

What if we capture it by NVH kit ?

Does the frequency chart display it that 30mg at 5Hz or 30mg at 15kHz or both?

Let's look at another example. If a bolt is nailed into the tire surface and the tire rotates at 20Hz, it is no doubt that a vibration of 20Hz will occur due to unsmooth tire. By capturing it with NVH kit, we can see a higher amplitude at 20Hz. This is similar to periodical knock above.

Which is the real frequency of knock? 5Hz of 15kHz.

How does the NVH kit distinguish the frequency of a vibration?

I get confuzed at this question. :roll:

Any post is appreciated! :D
Best regards

Steve Smith
Pico Staff Member
Pico Staff Member
Posts: 633
Joined: Sun Aug 25, 2013 7:22 am

Re: How does NVH kit Capture a Periodical Knock?

Post by Steve Smith » Sun May 15, 2016 9:56 pm

Hello Belet and thank you for the enquiry
I think we have a number of questions here that we need to answer in stages.
Your initial question “How does NVH kit capture a Periodical Knock?”
This answer is taken care of here:

The issue around a periodic knock is the lack of time the knock is present. (If referring to knocks such as suspension etc.)

We require a minimum 2.5 seconds of continuous disturbance in order to process and display such events using the NVH software?

Please forgive me here Belet as I am a little confused with your calculation surrounding engine knock.
An event that occurs every 0.2 seconds equates to 5 Hz (1 sec / 0.2 sec = 5)
5 Hz would be the fundamental frequency of the knock, however 15 kHz could be a 3000 harmonic (order) of the fundamental knock. The frequency display would show both the fundamental knock at 5 Hz in mg or dB (assuming the mic can detect 5 Hz) and the high frequency knock at 15 kHz in dB only

Just thinking this through, a 4 cylinder 4 stroke engine running at 3000 rpm is also rotating at 50 Hz (3000 / 60 = 50)
50Hz or 50 crankshaft revolutions per second will produce 100 combustion events per 50 revolutions (2 combustion events per revolution)

This would be displayed as a second order engine vibration in our NVH software (E2) at 100 Hz in the frequency display.
Engine knock created as result of premature detonation of the fuel will occur at a high frequency proportional to the fundamental engine speed (E1)

Engine knock (Pinking) is generated by numerous factors including engine load, temperature, fuel quality, vaporisation, compression and cylinder condition (The frequency of such a knock is predominately unknown but measurable if present for sufficient time)

Often, pinking will be evident for a very short period of time at high frequencies detected by the NVH microphone. The concern here once again is the sporadic nature of pinking and time span this type of noise would be evident.
Here we are faced with a challenging measurement using the microphone in an exceptional noisy environment (engine bay)

To capture pinking the microphone would be selected where the NVH frequency spectrum/display would be extended to 20 kHz. Pinking/knocking at high frequency would be detected by the mic and displayed at the relevant frequency, the amplitude would therefore be measured in dB and not mg

Remember that sounds and vibrations are one and the same thing. From 0-20 Hz we feel, 20-200 Hz we can feel or hear and above 200 Hz we can hear only (hence the use of the mic)

I have never tried to capture pinking via the NVH mic but I would imagine that the frequency could be traced back to the initial engine speed/frequency? (This sounds like a future experiment for sure)

Looking at your nail in the tyre analogy, we have one disturbance or shake per revolution of the tyre and so a high amplitude of vibration would be present (mg) at 20 Hz (The fundamental speed of the tyre-T1)

With regards to how does the NVH kit distinguish the frequency of a vibration?

This is predominately dependent on the information entered by the user.
For example, entering tyre size, engine cylinder count and differential ratio (as a minimum) enables the software to identify vibrations relevant to road or engine speed obtained via the ODB connector.

Remember all vibrations are fed into the software via the accelerometer placed initially at the driver’s seat bolt. If we take wheel frequency as an example, given we have entered the tyre size, the software will calculate the tyre circumference, then calculate the relevant wheel frequency required to match the acquired road speed
Multiply wheel frequency by the differential ratio and then you have components rotating at propshaft or transmission output shaft frequency.

Not all vibrations can be identified, but unknown vibration levels will be indicated at their specific frequencies allowing technicians to identify their source by calculating components rotating at speeds indicted by the unknown frequency.
If we can identify the frequency of an offending vibration or sound we are half way to a diagnosis

I hope this information is of some help.
Take care………Steve

Post Reply