Noisy frequency analysis

Ask any questions about using PicoScope Automotive software here.
Post Reply
JohnD
Newbie
Posts: 2
Joined: Mon Nov 26, 2018 4:34 pm

Noisy frequency analysis

Post by JohnD » Mon Nov 26, 2018 5:08 pm

20181027 1000rpm.psdata
(256.21 KiB) Downloaded 7 times
Hello, newbie here!

I'm a engineering student studying crankshaft vibration, using a Picoscope 2204. I have a toothed wheel on the crank nose and an inductive sensor, that shows a nice raw trace, but the frequency analysis is as noisy as - a noisy thing! Please see picture.
I have the "Picoscope User Guide" and have read Steve Smith's training guide "Using filtering functions", but the 'low pass filtering' and Math Channel steps don't work on the spectrum trace. Do I need to apply them to the raw signal?
Can I save a run on my test rig, say 5 minutes-worth, or less? One minute would do. The Guides refer to saving the display area, and a "Save on trigger" option but that isn't there under "File". It would be most useful to be able to take my data back to my desk and work there!


Thanks!
John

Steve Smith
Pico Staff Member
Pico Staff Member
Posts: 607
Joined: Sun Aug 25, 2013 7:22 am

Re: Noisy frequency analysis

Post by Steve Smith » Tue Nov 27, 2018 2:40 pm

Hello John thank you for the post.

Whilst a number of math channels can be applied in Spectrum Mode the Filtering Options I have tried do not.

Bandwidth Limit & Low Pass filtering are not applicable whilst Low Pass Math is not applied when created
https://www.picoauto.com/library/training/filtering

You could switch the display mode from Logarithmic to Linear mode but I am not sure if you would lose too much data?
Linear mode.jpg
Linear mode
Here you would need to capture the raw data from the inductive sensor at the crankshaft in the Time Domain.

Now you can apply the relevant math channels for filtering (if applicable) and frequency to determine the acceleration and deceleration of the crank during compression and combustion
post86291.html#p86291

You can capture this data over the relevant time period and save accordingly in order to work at your desk away from the coal face. There is however a concern with the specification of the 2204 Scope you are using as for math to perform efficiently, a high number of samples will be required.
topic21311.html

Measuring the crank signal for 1 minute over a variety of engine speeds (frequencies) is going to be "pushing" the 2204
Sample Rate = Number of Samples across the screen / Time Span across the screen

Filtering will also improve the clarity of the applied math and once again we may struggle here unless we can incorporate filtering into our math formula for frequency

The specifications of the 2204 can be found here
https://www.picotech.com/oscilloscope/2 ... ifications

Where as the forum post you have read and the math options discussed feature the automotive 4425 scope where you can compare scopes with reference to sampling and memory here:
https://www.picoauto.com/products/autom ... ifications

Whilst the 2204 is a Pocket Rocket for sure I am not convinced we can obtain the resolution you require with this scope in your application.

Our NVH software, (which is included with our automotive software) can be downloaded here https://www.picoauto.com/downloads where you can view the pddata file contained in the link below:

Here you can view the engine vibration orders/frequencies of a 4 cylinder diesel engine over 50 seconds from idle to approx. 2000 rpm
https://drive.google.com/file/d/12RDVTe ... sp=sharing

You can find more NVH posts here
forum481.html

N.B. In order to use PicoDiagnostic NVH software you would need an automotive scope, you can however view NVH files without a scope connected to your PC

I hope this helps and please feedback for any clarification. Take care.......Steve

JohnD
Newbie
Posts: 2
Joined: Mon Nov 26, 2018 4:34 pm

Re: Noisy frequency analysis

Post by JohnD » Mon Dec 03, 2018 5:16 pm

Thank you, Steve, for such a detailed and helpful answer!

I'll need to read, think and then have another session on the rig before I can come back, but thank yiou again for the offer!
John

Post Reply