Convert to square to detect phase shift

[Tubi]

Hi,

Im quiet new in labview, using version 8.6.

I need to develop a system to detect a phase shift of vibration between a case and the rotor. (of a washmachine)

I simulated the signals in labview and the upper "pulse measurements.vi" quotes:

"Error -20308 occurred at an unidentified location

Possible reason(s):

Analysis: The waveform did not cross the mid reference level enough times to perform this measurement.

Check the signal length, reference levels, and ref level units.Analysis:

The waveform did not cross the mid reference level enough times to perform this measurement.

Check the signal length, reference levels, and ref level units."

If i attach a signal directly to the upper pulse measurement vi then it works fine.

if i run with the converted signal he gives the error, but i can continue

But i need to measure from 3Hz, now it's still 20

If i place it on 3Hz he refuses everything.

Hope someone know's a solution.

Download File:post-15503-1242203275.vi

[ShaunR]

You don't have enough data points above and below the threshold to complete the "Peak Measurement" analysis.

Either

1.Change the "percent level settings" method from "Auto Select" to "histogram" to force that method

or

2. Change the threshold levels to something like 40%,50% and 60%

[Tubi]

QUOTE (ShaunR @ May 13 2009, 07:29 PM)

You don't have enough data points above and below the threshold to complete the "Peak Measurement" analysis.

Either

1.Change the "percent level settings" method from "Auto Select" to "histogram" to force that method

or

2. Change the threshold levels to something like 40%,50% and 60%

Thx,

Tried the percent levels and also setted the signal in resetting mode.

it al works fine. then i tried it on the machine himself but it didnt worked.

Figured out that it was of cause of the low frequency. I also simulated it, but it didn't work.

I already tried to changhe the samples and the rate. But didn't make any sense.

Tried to change my filters. but all without succes.

http://lavag.org/old_files/post-15503-1242391709.vi'>Download File:post-15503-1242391709.vi

[ShaunR]

QUOTE (Tubi @ May 15 2009, 01:49 PM)

Thx,

Tried the percent levels and also setted the signal in resetting mode.

it al works fine. then i tried it on the machine himself but it didnt worked.

Figured out that it was of cause of the low frequency. I also simulated it, but it didn't work.

I already tried to changhe the samples and the rate. But didn't make any sense.

Tried to change my filters. but all without succes.

http://lavag.org/old_files/post-15503-1242391709.vi'>Download File:post-15503-1242391709.vi

Well the latest example you supplied seems to work fine. I changed the variables on the front panel and it didn't complain. Are you saying that it works in simulation but not on the hardware?

If this is the case. It might be useful to capture the data from the machine to a file and run that through the simulation. Then w can all see and investigate the differences.

[Tubi]

QUOTE (ShaunR @ May 15 2009, 03:18 PM)

Well the latest example you supplied seems to work fine. I changed the variables on the front panel and it didn't complain. Are you saying that it works in simulation but not on the hardware?

If this is the case. It might be useful to capture the data from the machine to a file and run that through the simulation. Then w can all see and investigate the differences.

I didn't runned it on the machine yet, but just simulated it.

If i lower the frequency then he gives the error again about the reference levels.

Don't understand why it works on 20 Hz but not on 3Hz

[ShaunR]

QUOTE (Tubi @ May 17 2009, 10:25 PM)

I didn't runned it on the machine yet, but just simulated it.

If i lower the frequency then he gives the error again about the reference levels.

Don't understand why it works on 20 Hz but not on 3Hz

This time, it's because you don't have enough samples to begin with to do the measurement.

When you run at 20Hz, you generate enough data for about 2 cycles. With the limits set tightly, this is just enough threshold transitions to do the measurement. At 3Hz you are only generating about 1/4 cycle to begin with. So no matter what you set the thresholds to be, you will never have enough transitions to make the analysis.

I have adjusted your signals to give enough data for the measurement @ 3Hz. See attached.

http://lavag.org/old_files/post-15232-1242599891.vi'>Download File:post-15232-1242599891.vi

[Tubi]

QUOTE (ShaunR @ May 18 2009, 12:40 AM)

This time, it's because you don't have enough samples to begin with to do the measurement.

When you run at 20Hz, you generate enough data for about 2 cycles. With the limits set tightly, this is just enough threshold transitions to do the measurement. At 3Hz you are only generating about 1/4 cycle to begin with. So no matter what you set the thresholds to be, you will never have enough transitions to make the analysis.

I have adjusted your signals to give enough data for the measurement @ 3Hz. See attached.

http://lavag.org/old_files/post-15232-1242599891.vi'>Download File:post-15232-1242599891.vi

Thx for the quick reply,

This is a huge step forward for us,

We can now start to measure on the machine.

Thx again for your help.

[JasonXCX]

Hi,

I am also interested in vibration measurments. I use also the free version of

http://www.20sim.com/

to perform simulations ( 1 hour of use allowed )

JJ

[Tubi]

QUOTE (JasonXCX @ May 18 2009, 03:07 PM)

Hi,

I am also interested in vibration measurments. I use also the free version of

http://www.20sim.com/

to perform simulations ( 1 hour of use allowed )

JJ

Intresting,

I use a accelerometer of analog device (ADXL103)

a 1,7g single axis accelerometer.

Is very usefull.

I read the signal with a daq to labview.

Then i simulate the signal identical on the measured one.

And when the simulation works i test the real signal.

works fine now.

thx for the link, defenetly will try it today.

[Tubi]

QUOTE (ShaunR @ May 18 2009, 12:40 AM)

This time, it's because you don't have enough samples to begin with to do the measurement.

When you run at 20Hz, you generate enough data for about 2 cycles. With the limits set tightly, this is just enough threshold transitions to do the measurement. At 3Hz you are only generating about 1/4 cycle to begin with. So no matter what you set the thresholds to be, you will never have enough transitions to make the analysis.

I have adjusted your signals to give enough data for the measurement @ 3Hz. See attached.

http://lavag.org/old_files/post-15232-1242599891.vi'>Download File:post-15232-1242599891.vi

Tried it on the machine,

everything worked fine

except the phase shift sometimes gives me value's of 227° or -201° for example.

These phase shifts are imposible cause you always stay between the -180° to 180°.

How can i prevent this?

by triggering or something like that...?

[Neville D]

I'm not sure what you mean but if you want to "unwrap" the phase so there are not discontinuities at +/-180 deg, search for Unwrap Phase.vi.

N.

[Tubi]

QUOTE (Neville D @ May 20 2009, 07:22 PM)

I'm not sure what you mean but if you want to "unwrap" the phase so there are not discontinuities at +/-180 deg, search for Unwrap Phase.vi.

N.

I know that vi exists but it doesn't work for a square signal.

So i can not use that vi.

Now i measure a phase shift that sometimes is more than +/-180 deg.

not errors but more like ==> (340° = -20°) it needs to stay between the +/-180°.

because now i can not put this in a graph in excell or something like that.

[Neville D]

QUOTE (Tubi @ May 21 2009, 03:15 PM)

I know that vi exists but it doesn't work for a square signal.

So i can not use that vi.

Now i measure a phase shift that sometimes is more than +/-180 deg.

not errors but more like ==> (340° = -20°) it needs to stay between the +/-180°.

because now i can not put this in a graph in excell or something like that.

Well you can easily write a VI to stay within whatever bounds you desire +/-180 or +/-360.

Alternately, download the MGI freeware VI's using the VI Package Manager.

Install the MGI VI's then use Coerce to Signed Degrees.vi for +/-180 or Coerce degrees.vi for +/-360.

Found under following pallets: 7.x functions>MGI All VI's>MGI Numeric Tools>MGI Conversions.

But it might be quicker to write it yourself:

http://lavag.org/old_files/monthly_05_2009/post-2680-1243011626.jpg' target="_blank">

N.

[Tubi]

QUOTE (Neville D @ May 22 2009, 07:01 PM)

Well you can easily write a VI to stay within whatever bounds you desire +/-180 or +/-360.

Alternately, download the MGI freeware VI's using the VI Package Manager.

Install the MGI VI's then use Coerce to Signed Degrees.vi for +/-180 or Coerce degrees.vi for +/-360.

Found under following pallets: 7.x functions>MGI All VI's>MGI Numeric Tools>MGI Conversions.

But it might be quicker to write it yourself:

http://lavag.org/old_files/monthly_05_2009/post-2680-1243011626.jpg' target="_blank">

N.

I didn't exactly did the same like you but i used the case structure and now it works.

Finaly can start measuring.

thx.