NVH analysis Help

Ask questions relating specifically to the PicoDiagnostics NVH equipment and software here.

NVH analysis Help

Postby Mark Dalton » Tue Sep 01, 2015 3:39 am

Hi Guys,
Still trying to get my head around noise diagnosis using the NVH tool. So I've got a V6 with an engine noise, that I suspect might be from the Camshaft. I wanted to use the software to analyse the noise and try to tie it to the engine frequency.
I suspect the noise is showing up at around 650-700 Hz.
If the engine is idling at 800 RPM and the Camshaft at 400 RPM, this gives the Camshaft a frequency of 6.6 Hz.
If I multiplied the Camshaft frequency by 100 then I am in the ball park for the suspected frequency of the noise.
So is it just coincedence that my noise is approx a 100 times my Camshaft frequency or is there a logical reason?
Attachments
XXR991.jpg
XXR991 Mic.pddata
(3.36 MiB) Downloaded 72 times
User avatar
Mark Dalton
Advanced User
Advanced User
 
Posts: 54
Joined: Mon Dec 01, 2014 9:35 pm
Location: Melbourne

Re: NVH analysis Help

Postby Steve Smith » Tue Sep 01, 2015 10:06 am

Hello Mark and thank you for the enquiry

It is interesting you have highlighted engine noise analysis using the NVH microphone as this can be challenging given the amount of background noise present in the engine bay.

From my experience it helps to "home in" on the frequency of interest and once again this can be a challenge.

If the engine noise is present continually and conditions allow, can you disconnect the auxiliary drive belt to eliminate all belt driven components? (Ensure cooling fans do not operate during the capture of the noise of interest)

Where possible, ensure environmental noise is kept to a minimum too (Which I know is difficult in the workshop)

Could you post the captured data here (PD file) so I can listen to the noise and apply settings that may help reveal the frequency of interest?

The formula of half engine speed / 60 X 100 may have some relevance but would require some study to confirm what possible components could generate such a noise.

Depending on the camshaft arrangement and number of valves in contact we have the potential for multiple noises (harmonics) per camshaft revolution.

Take into consideration any wear in bearing or lobe faces, and we have more potential for noise generation that we cannot factor into our test results.

What we can do is highlight the frequency of interest, measure the amplitude, eliminate components not rotating at the frequencies we have identified and then form an action plan on how bets to proceed, keeping the customer in the loop.

Could you let me know how this unfolds and post the file if possible?

Take care......Steve
Steve Smith
Pico Staff Member
Pico Staff Member
 
Posts: 426
Joined: Sun Aug 25, 2013 7:22 am

Re: NVH analysis Help

Postby Mark Dalton » Wed Sep 02, 2015 3:42 am

Hi Steve,
I did attach that file in my original post and The pddata file appears to be there in the original post as a 3.36Mb attachment. I'll attach it here again anyway.
The pddata file was captured with the accessory drive belt removed, the noise is similar to a squeaky drive belt so that was our first step to eliminate the belt and associated pullies and accessory components.
Some of the back story to this engine is that it was a change over assembly supplied to us by a 3rd. party supplier. So all I was hoping to achieve was to tie the noise to either the engine speed or an order of the engine speed. So I could narrow down the search area and work out how far I go with my diagnosis and work out if it will be a warranty issue or not.
Hopefully what I've said here and my attachments will help you understand what I'm trying to get to.
Attachments
Publication1.pdf
(86.43 KiB) Downloaded 90 times
XXR991 Mic.pddata
(3.36 MiB) Downloaded 74 times
User avatar
Mark Dalton
Advanced User
Advanced User
 
Posts: 54
Joined: Mon Dec 01, 2014 9:35 pm
Location: Melbourne

Re: NVH analysis Help

Postby Steve Smith » Wed Sep 02, 2015 2:52 pm

Hello Mark thank you for the feedback and sorry, I should look more closely at “Attachments” before asking for files that are already included.

I have downloaded the pddata file and listened very closely to the audio. I have noticed an intermittent squeak type noise similar to drive belt (as you have mentioned).
I can also hear a tapping noises that’s comes into the audio but not sure it’s relevant?

From the data I have, the engine speed appears static fixed at 750 rpm and the engine speed signal also appears lost part way through the capture.

Would it be possible to retake the capture with the rpm captured from either the OBD or the crankshaft position sensor into channel A?
Here the rpm would be accurate for analysis throughout the capture.

Can I ask if the position of the microphone was altered during the capture?
The signal history displays a higher amplitude at the beginning of the capture. (The squeak noise appears later on through the signal highlighter captured in the image below.)

DATA CAPTURED.jpg


If we are to focus on the squeak noise (which sounds intermittent) the area of interest you have highlighted (600 Hz) is present throughout the capture. (I would expect that peak to rise and fall)

Looking and listening to the data very closely, when the squeak develops, a peak is present around 113.16 Hz. Multiply this value by 6 and were in the range you have highlighted 678.96 Hz.

Could this be due to the fact that the microphone can detect something in this frequency range not audible by ear due to back ground noise?

Could the peak at 113.16 Hz be our fundamental frequency and the peak at 678.96 Hz be 6th harmonic frequency of the fundamental frequency. To be totally honest, I do not know!

What I do know is 113.16 Hz is too low an audible frequency to generate a squeak noise (Squeaks are a higher frequency)

They do however coincide very nicely and display some of the highest amplitudes (60-70 dB) with 6 cylinders being the connection to our 6th Harmonic.

I don’t believe these peaks can be ignored but I would recapture the data with a fixed microphone and true engine speed input throughout the capture.

A good tip hear would be to down load “audio frequency generator” software as it can often help you to home in on noises in order to locate the frequency of interest.

This will help you to identify the frequencies of noise as you become more experienced.
You can play the audio captured in your NVH software through headphones whilst introducing matching frequencies from the audio generator software.

Here you can be sure the peak you see in the frequency chart of the NVH software during playback matches the frequency generated by the audio frequency generator.

I know exactly where you are Mark and what you are trying to get at with regards to any potential warranty on this engine.

Moving onto your “Publication 1” attachment.

PUBLICATION 1.jpg


The area you have highlighted around 699 Hz in the frequency chart may well be the sound we hear, but this could be inaudible to the ear as a result of background noise.

The frequency chart indicates the level of sound (dB) in direct relation to the frequency of rotating parts within the engine bay (If they are linked) and the frequencies that are applied to the microphone (if external and unrelated)

The frequency spectrum of 0-1000 Hz (our frequency chart in your capture) will display all audio frequencies present at the microphone.

For example:

Noise levels (dB) generated as a result of engine speed alone will appear at a frequency of 12.5 Hz at 750 rpm (E1) at the relevant sound level (dB)

Noise levels (dB) generated as a result of an electric cooling fan will appear at a frequency somewhere within the 0-1000 Hz spectrum at the relevant sound level (dB)

Both are visible in the frequency chart (0-1000 HZ spectrum) but neither are directly related.

The scales you have mentioned in the signal history and the frequency chart are not the same and this will be rectified as what you have identified is a bug.

The signal history graph is best utilised as a visual indictor of activity detected by the microphone.

The dB value however in the frequency chart is correct and is an absolute measurement of the signals captured by the microphone

The NVH software requires vibrations and noise to be present for a minimum of 2-4 seconds before it can process and display the frequencies of interest associated with the signals measured by the accelerometer or microphone.

Unfortunately the spikes you have highlighted are not audible for long enough.

With that said, the spikes will still be incorporated into the calculations and all frequencies will be displayed in the frequency chart.

This would be an average of all signals contained inside the signal highlighter, including the spike in the calculation but not displaying a frequency value specific to the spike alone.

I hope this helps, take care……..Steve
Steve Smith
Pico Staff Member
Pico Staff Member
 
Posts: 426
Joined: Sun Aug 25, 2013 7:22 am

Re: NVH analysis Help

Postby Mark Dalton » Thu Sep 03, 2015 4:13 am

Hi Steve,
Thanks for the help, every little bit of knowlege gained helps.
I did move the mic during the capture, just to see if there was any improvement in the signal.
I didn't have an RPM signal as it was this car that I started experimenting with the Bluetooth ELM327 Adapter.
I'm not sure I'll get a chance to get another capture of the noise, it's gone to another workshop for assessment for the 3rd. party that supplied the engine to us.
I'll definately see if I can down load a audio frquency generator, that could be a really usefull tool.
Thanks Again.
Mark.
User avatar
Mark Dalton
Advanced User
Advanced User
 
Posts: 54
Joined: Mon Dec 01, 2014 9:35 pm
Location: Melbourne

Re: NVH analysis Help

Postby csasziszerviz » Fri May 27, 2016 9:13 am

Hi Guys,

I would like to ask something about the program. I take a measure by NVH kit, which is a long drive (620s). Now I would like to search what happened in the middle of the trip (eg. in 250s-280s), but I can't find the button wich take the window (and the time) back.

Is it possible to go back in the file, or I saw only the last 60s in the window? If somebody knows the solution, please write it to me.

Thanks.
csasziszerviz
User
User
 
Posts: 10
Joined: Sun Dec 14, 2014 1:11 pm

Re: NVH analysis Help

Postby Steve Smith » Sat Jun 04, 2016 1:13 pm

Hello and thank you for the post.

The maximum record time of the NVH software is 500 seconds, but the default setting is 50 seconds.

If you are recording data for 620 seconds (extended road test) with your signal history set to 50 seconds, you will only, ever record the last 50 second of your road tests and your earlier data (previous 570 seconds will be lost)

The signal history can be set to a maximum of 500 seconds at the "Set-Up" screen. (See screen shot below)

INCREASE NVH SIGNAL HISTORY.jpg


Remember that you can pause the NVH capture (when the noise/vibration is present) using the space bar and save this data when it is safe to do so. This will keep your file size down whilst containing the detail of interest

I hope this helps, take care.........Steve
Steve Smith
Pico Staff Member
Pico Staff Member
 
Posts: 426
Joined: Sun Aug 25, 2013 7:22 am

Re: NVH analysis Help

Postby csasziszerviz » Sun Jun 05, 2016 8:50 pm

Helo Steve,

Thanks the help. I have to pay attention to the correct settings after that. I will try to measure one more time what I couldn't last time.

Thank you the picture and the help one more time.

Have a nice day!
csasziszerviz
User
User
 
Posts: 10
Joined: Sun Dec 14, 2014 1:11 pm


Return to PicoDiagnostics NVH

Who is online

Users browsing this forum: No registered users and 0 guests