Quick Question On Pulse Measurements

[NateTheGreat]

I was wondering what the percents are the percent of on the Pulse measurments.vi   Default (10,50,90)%

I understand the low,mid, and high part of the vi. But what are the percents referencing? It it simply  % of your largest pulse?

I know it isn't very clear what I am asking but Im not sure how to ask it more clearly.

Link to help doc

[CCornish]

The percents are the percentage of the amplitude of the signal you are measuring.

So...if you have a signal with amplitude of 0-5V, then:

Low (10%) - .5V

Mid (50%) - 2.5V

High (90%) - 4.5V

You can customize the Low, Mid, High levels to the percentaage you want.

[NateTheGreat]

Well the waveforms I am running through labview are sound recordings of a weld so it is not as simple as that. I need to count the large pulses you can see below, and the entire sound has the charateristic pattern seen below. Noise, followed by a small peak, more noise, followed but the larger spike. and I am currently just using absolutes on the pulse measurement vi, but its missing a few of the smaller pulses, and double counter a few of the larger pulses. And I am not't sure how the % counts since this is a real life example and the amplitude of what I am counting changes.

[CCornish]

You can try to apply a filter to remove the noise if you are able to characterize the noise and the signal.

You can apply windowing to the pulses if the pulses are periodic; this way you can limit the area that you are using for the measurement.

Charactierize the noise and the signal of interest; you might have to "chop up" your recorded data to look at the noise and signal seperately.  Once you have those characterizations you can apply filtering and windowing accordingly.

[NateTheGreat]

Well unfortunately I cannot do any of those things. I already have a basic filter to cut out as much unnecessary noise as possible. I am trying to detect which waveform I am currently receiving  from my mic. There are 3, the first is the above, and the other two are below.  The first pulses are from 200 to 20 Hz, the 3rd pulses similarly but 15 to 1, and the second has occasional random spikes, but they are not the same obvious pulse.

2nd 

3rd

[NateTheGreat]

Oh, and I wasn't trying to get into this much detail originally I was just wondering since it wasn't a perfect waveform how the percents worked. Because if I am measuring a perfect  10Vdc  the 90% is 9v. But since each pulse changes, does it simply take the largest pulse, lets say 1 v for clearness, and take 90% of that even though the rest are around .5 V peak?

[crossrulz]

---

@NateTheGreat wrote:

Oh, and I wasn't trying to get into this much detail originally I was just wondering since it wasn't a perfect waveform how the percents worked. Because if I am measuring a perfect  10Vdc  the 90% is 9v. But since each pulse changes, does it simply take the largest pulse, lets say 1 v for clearness, and take 90% of that even though the rest are around .5 V peak?

---

I'm pretty sure it is based on the entire waveform's min and max.  So in your example, it would use 0.9V (90%) for the high.

[NateTheGreat]

Eh, I ended up getting it working using absolutes anyways. But, thanks for the response.

+kudos for the time.

[johnsold]

Nate...,

Typically the Pulse Measurement VI used on something which looks like a rectangular pulse with a little noise, overshooot, ringing, ...  Your signal is not what one would normally call a pulse by that definition.

In the first image you posted how many pulses do you count? 11? The little spikes (amplitude about 0.1 at 3.318, 3.331, 3.341, and others): Are they to be counted or rejected? Do they alwyas occur before a big spike? Some little ones are not followed by big ones but all the bigones seem to be preceeded by a small one.  Is there a minimum dleay between successive pulses?  Analyzing data like this can be challenging.  Use all the knowledge you have about the process which generates the data to help.

Lynn

[NateTheGreat]

---

@johnsold wrote:

Nate...,

Typically the Pulse Measurement VI used on something which looks like a rectangular pulse with a little noise, overshooot, ringing, ...  Your signal is not what one would normally call a pulse by that definition.

In the first image you posted how many pulses do you count? 11? The little spikes (amplitude about 0.1 at 3.318, 3.331, 3.341, and others): Are they to be counted or rejected? Do they alwyas occur before a big spike? Some little ones are not followed by big ones but all the bigones seem to be preceeded by a small one.  Is there a minimum dleay between successive pulses?  Analyzing data like this can be challenging.  Use all the knowledge you have about the process which generates the data to help.

Lynn

---

Hmm, all I had seen was sine waves before with the pulse. But, you seem to be right, rectangular pulses make a lot more sense.

Yes, I am just counting each period. I just was planning on using the pulse vi to measure the giant spikes to count how many of those occur each second. That first graph is a sound wave of a weld shorting. I want to know how often it shorts a second. I currently just have it detecting an absolute value (looking for that giant spike) and however many times it detects that I average and turn into total shorts and average shorts.

It works but it is not very elegant.

And you are correct about some of the small spikes not being followed by a big spike. That is molten slag splashing about. This short one right before the sharp spike is the sound of the arc igniting (300 Amps worth).

Can you think of any more elegant solutions to detecting the pulse? I am going to do some digging around..

Thanks Lynn!