DIAdem

cancel
Showing results for 
Search instead for 
Did you mean: 

LabVIEW equivallent order tracking functions in DIAdem

Till now we have used LabVIEW for online as well as offline order tarcking of some rotating machineries. In all this application, we acquire vibration signals with DSA cards and tachometer signals with counter card. For frequency measurement with counter card we use the buffered edge counting technique.So, our logged files contain waveform data type for vibration channels and 1D array of U32 (output of counter) for the  speed channel. 

 

 We are thinking of moving the offline analysis part to DIAdem. For this we have decided to Log the data in TDMS (as some of our applications run on RT target) format with one (waveform) channel for Vibration and one (numeric) channel for speed.

 

Now the problem is DIAdem order analysis functions has this configuration panel for Order analysis in time domain.

 

Here ther are three channels, the time channel, signal channel( which will be the vibration channel) and rpm channel (which will be my speed channel).

 

1. Since I am logging counter data in the raw form, I will have to convert this channel to RPM channel. is there any function available in DIAdem which can do this?

 

2. Else I can speed data after convert to RPM. But then I will have two arrays for one speed, the speed array and corresponding time array (returned by the OAT Digital tacho process.vi). So my TDMS file will have one RPM channel and a time channel for any speed channel. Now can I use this time and RPM channel in the order analysis configuration and the (waveform) vibration channel in the signal field?

 

3. Else is it advisable to convert the acquired digital tacho to an analog tacho signal using the OAT Convert Speed XY graph to Waveform.vi in labview befroe writing to the TDMS file. In this case I will have one waveform channel for vibration and one waveform channel for speed. But speed channel will be in the form of pulses and not RPM, so will it work?

 

4. Or is there any other format in which I should store speed data?

 

I was not able to see and example of Order analysis implementation in DIAdem examples. May be it is just that I am not abne to find it. Is there any examples available?

 

Thanks!!!

 

 

 

 

Message Edited by kikiduu on 12-23-2008 12:36 AM
------

"A VI inside a Class is worth hundreds in the bush"
യവന്‍ പുലിയാണു കേട്ടാ!!!
0 Kudos
Message 1 of 7
(4,808 Views)

Hi kikiduu,

 

You will need to create an explicit RPM channel out of your tachometer channel in order to use the DIAdem order analysis functions.  I recommend downloading the "DIAdem Joint Time Frequency Order Analysis" example, which contains a method for doing the tachometer to RPM conversion.

 

http://zone.ni.com/devzone/cda/tut/p/id/3549 

 

Brad Turpin
DIAdm Product Support Engineer

Naitonal Instruments

Message 2 of 7
(4,787 Views)

Hi Brad,

 

 I download the example that you specified. There the tachometer to RPM conversion needs the speed channel to be in the form of pulses. In my case, since I am using counter (edge counting method) to acquire tachometer signal, I will not get the speed signal as pulses but as counts.

 

I have a feeling that the best way will be to do analysis by calling LabVIEW VIs in DIAdem.

 

Thanks

------

"A VI inside a Class is worth hundreds in the bush"
യവന്‍ പുലിയാണു കേട്ടാ!!!
0 Kudos
Message 3 of 7
(4,779 Views)

Hi kikiduu,

 

The Tachomet to RPM conversion in that example counts rising edges, so it doesn't matter if you have spikes o square waves.  If there are a certain number of rising edges from your counter per revolution, this example should convert your counter channel to RPM channel.

 

Brad Turpin
DIAdem Product Support Engineer
National Instruments

0 Kudos
Message 4 of 7
(4,762 Views)

Hello Brad,

 

 The data that I am using is attached here. Actually it is not a waveform data that I get from my counter. It gives me counts (from the 32-bit counter which is incremeneted by a known clock) between two successive edges of my speed signal.

------

"A VI inside a Class is worth hundreds in the bush"
യവന്‍ പുലിയാണു കേട്ടാ!!!
0 Kudos
Message 5 of 7
(4,754 Views)

Hi kikiduu,

 

I see now that the counter is counting up the total number of edge events.  We should be able to convert this to a series of edges by taking the derivative of the counter total over time.  But your "rpm" channel has 222 data points, whereas your "Dev2/ai0" channel has 180224 data points.  You say that it is from a know clock-- what is the sampling rate of that clock?

 

Brad Turpin

DIAdem Product Support Engineer
National Instruments

0 Kudos
Message 6 of 7
(4,746 Views)

Hi Brad,

 

The counter timebase in this case is 16777216Hz. This is included as a property of the rpm channel in the file.

 

Normally when we acquire vibration and speed data together, to make sure both data are synchronized, we use the oversampling clock of the DSA card as timebase (the clock which increments the 32-bit counter) for the counter.

 

The value of the 32-bit counter will be passed on to the read buffer, only when the rising/falling edge of the speed signal comes. This means that, if your speed signal is of 1000Hz, then the read buffer will have only 1000 data values in one second. The number of points in the read buffer keeps changing depending on the speed signal's frequency.

 

So if the DSA sampling rate is 10kS/s, at the end of 1 second, you will have 10k samples from the DSA card but may or may not have that many samples from the counter card. The counter card can give you either less, equal or more than 10k samples in 1 second depending on the speed signal. This is the reason why they call the speed signal as the 'sampling clock' in bufferededge counting method.

 

The known clock in this case (which increments the 32-bit counter) will be the oversampling clock of the DSA card.

 


Brad Turpin wrote:

 

I see now that the counter is counting up the total number of edge events.  We should be able to convert this to a series of edges by taking the derivative of the counter total over time. 


 

 Actually the counter in not counting the total number of edge events, but it puts the value of the counter to the read buffer at every edge event.

Message Edited by kikiduu on 12-26-2008 09:24 PM
------

"A VI inside a Class is worth hundreds in the bush"
യവന്‍ പുലിയാണു കേട്ടാ!!!
0 Kudos
Message 7 of 7
(4,742 Views)