personal computer as a tachometer

Arborist Forum

Help Support Arborist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

heyduke

Addicted to ArboristSite
AS Supporting Member
Joined
Mar 6, 2009
Messages
1,365
Reaction score
949
Location
in the mountains
thanks to Chris-PA for inspiring this. the idea of using a video or audio file to examine and evaluate performance of a 2stroke engine fired me up. yes, i have a wireless tach that i use in the shop, but its hard to quantify performance of a saw in the field. what rpm does it turn in the cut? this is close to being the under $100 chainsaw dyno.

here's my youtube vid:

 
Good explanatory walk through.

I'll at some point in time try using this for adjusting the idle speed of the saw.

But ideally an app on your smartphone containing a real time acoustic spectrum analyzer would be best.
 
thanks to Chris-PA for inspiring this. the idea of using a video or audio file to examine and evaluate performance of a 2stroke engine fired me up. yes, i have a wireless tach that i use in the shop, but its hard to quantify performance of a saw in the field. what rpm does it turn in the cut? this is close to being the under $100 chainsaw dyno.

here's my youtube vid:


Nice job! A couple of notes:

1. I did not extract the audio tracks separately, as I was able to open the video files directly in Audactiy.
2. I found that you do not need to delete sections of the file you don't want to see - the spectrum plot will show the data you have selected. You can try selecting various section and then click on replot.
3. My camera only had one microphone, so the two stereo tracks were identical and there was no need to combine them.
 
But ideally an app on your smartphone containing a real time acoustic spectrum analyzer would be best.
I have one called "Engine RPM" on my iPhone, but it was fussy to use and I never really got it to work well. Basically, it is vastly easier for a human to look at the graph and decide where the spike of the first harmonic is than it is for the app to figure it out in real time.
 
Nice job! A couple of notes:

1. I did not extract the audio tracks separately, as I was able to open the video files directly in Audactiy.
2. I found that you do not need to delete sections of the file you don't want to see - the spectrum plot will show the data you have selected. You can try selecting various section and then click on replot.
3. My camera only had one microphone, so the two stereo tracks were identical and there was no need to combine them.

thanks chris, i'm sure i tried to load the video file but audacity choked on it. i think it is a .mov file from my nikon.

as far as selecting vs deleting, i do it both ways depending on my whim. i didn't want to complicate and lengthen the vid by showing both methods and went with the delete.

it seems kind of silly to me to put two mic's in a camera. they are only a couple of inches separated, no way to get much of a stereo effect and plenty of opportunity to create noise. i do love my two cameras though, both nikon L series, full 1080hd with cmos sensors.

also, after doing analysis on sounds other than a chainsaw, i'm beginning to think that the shape of the spectrum is actually just the response curve of the mic.

another thing that worked for me, sometimes audacity would screw up and play files back at a high speed. now i invoke audacity with this script:

env PULSE_LATENCY_MSEC=1 audacity
 
Good explanatory walk through.

I'll at some point in time try using this for adjusting the idle speed of the saw.

But ideally an app on your smartphone containing a real time acoustic spectrum analyzer would be best.

I have one called "Engine RPM" on my iPhone, but it was fussy to use and I never really got it to work well. Basically, it is vastly easier for a human to look at the graph and decide where the spike of the first harmonic is than it is for the app to figure it out in real time.

what might work would be to use and adjustable bandpass filter that would knock down everything except freqs between 100hz and 200hz or thereabouts. again the response curve fo a "communications" microphone doesn't lend itself well to spectrum analysis. i doubt there is a mic that is linear for 1hz to 8000hz. limiting it to a single cylinder two stroke would make the problem easier.
 
It's basically a FFT digital signal processing scenario. What you want, is a precise reading in frequency domain to pick a sharp spike of the fundamental, in order to determine eg. the idle speed of the chainsaw .

To do that, pick a homogeneous part of the recorded audio time domain signal. And yes, you can filter this time domain signal, before converting it to frequency.

Also, in order to achieve a good frequency readout, i.e. better frequency resolution readout, requires a longer audio sample sequence in time domain.
 
It's basically a FFT digital signal processing scenario. What you want, is a precise reading in frequency domain to pick a sharp spike of the fundamental, in order to determine eg. the idle speed of the chainsaw .

To do that, pick a homogeneous part of the recorded audio time domain signal. And yes, you can filter this time domain signal, before converting it to frequency.

Also, in order to achieve a good frequency readout, i.e. better frequency resolution readout, requires a longer audio sample sequence in time domain.

geez! fast fourior transformation is over this knuckle dragger's head. but i think your second point makes sense. " pick a homogeneous part of the recorded audio time domain signal," in english, you need to pay attention to what part of a recorded sequence you sample. here we're not really concerned with idle speed, but wfo in the wood. a more powerful saw turns faster under load. we have measured saws turning over 10k rpm in hardwood. so select a part of the audio file that corresponds with a steady load and level audio amplitude.

regarding longer samples, i'm not clear on the significance. a good saw requires ten seconds or less to get thru an 18" diameter cookie. for the sample you want to throw out the rpms at the beginning and end. a skilled operater should be able to maintain a constant rpm in between. does it matter if the duration is five seconds or eight seconds?

again, i think a more linear microphone might be helpful. just sayin'.
 
Good explanatory walk through.

I'll at some point in time try using this for adjusting the idle speed of the saw.

But ideally an app on your smartphone containing a real time acoustic spectrum analyzer would be best.

you don't really need anything elaborate to adjust idle speed. just set it slow enough so the clutch disengages and fast enough so it doesn't die when you set it down.
 
you don't really need anything elaborate to adjust idle speed. just set it slow enough so the clutch disengages and fast enough so it doesn't die when you set it down.
True, I do listen to the engine, give it enough idle to have a steady rpm, without engaging the clutch. Maybe need to have the spark plug cleaned and adjusted in advance though.

But for curiosity, I would like to calibrate my settings to the manufacturers instruction books, whether it's the old Jonsereds, the smaller Stihl MS192 or the larger MS362.
 
It's basically a FFT digital signal processing scenario. What you want, is a precise reading in frequency domain to pick a sharp spike of the fundamental, in order to determine eg. the idle speed of the chainsaw .

To do that, pick a homogeneous part of the recorded audio time domain signal. And yes, you can filter this time domain signal, before converting it to frequency.

Also, in order to achieve a good frequency readout, i.e. better frequency resolution readout, requires a longer audio sample sequence in time domain.
LOL - All correct, but I wasn't going to go there on a saw forum!

A longer sample time would be better, but not if the signal is changing. So it ends up being a compromise where you want to select the longest part of the recording where the rpm is reasonably constant.
 
I have one called "Engine RPM" on my iPhone...
Boy... do I ever feel dumb... I can't even figure out how to take a picture with my "smart" phone :(
"They" tell me it has "apps" on it, my daughter has even installed some on it... I don't even know where to find them. Sometimes I accidentally touch something on it and weird screens come up (must be apps??)... the only way I know to get rid of them is to reboot the phone. I'm "smart phone illiterate"... I've tried, but my old, stiff, callused, arthritis-plagued fingers always touch the wrong thing.

It would be really cool to be able to use the phone as a tachometer... really cool.
*
 
Boy... do I ever feel dumb... I can't even figure out how to take a picture with my "smart" phone :(
"They" tell me it has "apps" on it, my daughter has even installed some on it... I don't even know where to find them. Sometimes I accidentally touch something on it and weird screens come up (must be apps??)... the only way I know to get rid of them is to reboot the phone. I'm "smart phone illiterate"... I've tried, but my old, stiff, callused, arthritis-plagued fingers always touch the wrong thing.

It would be really cool to be able to use the phone as a tachometer... really cool.
*
They're actually a lot easier than you think, it just takes a little time and patience.
 
Back
Top