"Fragmented" measurement results

[ReBoot]

To be honest, I don't know whether this problem is generated by LabView/DAQmx, but I can't exclude that possibility so I'll ask here hoping for someone to know.

I have an experiment running which contains a laser, a Fabry-Perrot-interferometer and a probe (which is however not avaible yet, so it's the laser and FPI only right now).

I use the attached VI to record the output of a PMT (photomultiplyer tube) behind the FPI. As you can see in the output file/screenshot, the laser peaks consist of sub-peaks instead of clean lorentz/gaussian+ profiles. When using the DAQmx test panel, the fragmentation is not there. However, the sampling rate in the DAQmx test panel seems to be much lower than the 1kHz I use in the VI so I still don't know whether the error is somewhere in the VI or the rest of the experiment.

Can anyone give me any hint on which direction I should think at?

Thanks in advance

[Marian_O]

Hello ReBoot,

unfortunately, I am no physicist, so I am not completly sure about your inquiry:

Analyzing your data (file63.txt) with Excel, I can see four clear "main" peaks with an altitude about 1.3Meg. These are occuring at 500, 1300, 2100 and 3000.

Looking on your Linux-Tool-Screenshot, I can see further (but lower) peaks, which I can not find in your txt-file data.

I think, that your question is: where are the other peaks?
 

Well, it seems to me, that your data is captured correctly by your data aquisition device (DAQ). Is it possible, that you double check your input voltage (AI0 and AI1) by using an external scope? What kind of capture hardware for the voltage signal are you using?

Kind regards,

Marian Olef

[Marian_O]

Hello ReBoot,

did you find out anything new about your inquiry?

Best regards,

MarianMO

[ReBoot]

I have found that the results are fragmented, when using Excel to display them too, but I don't have an external scope to look at the voltages. No real progress here 😞

[Norbert_B]

Looking into the vi shows that there is much improvement possible. Using the clean up functionality from LV 8.6 helps in making the vi more readable. But there are unnecessary structures like the whole sequence structure which can be deleted without changing in the VIs execution too much.

Besides that note, i am not sure what you understand as "fragmentation of the signal".

To make your note "When using the DAQmx test panel, the fragmentation is not there." comparable to your VI, you have to configure the testpanel the same way as your task in the VI. The testpanel is set to "on demand" for sampling by default, so the signal is acquired by interrupts and therefore depending a lot on the system. You have to change that to continuous and configure a sample clock (1kHz is the default in your VI).

Please describe what you suppose the signal to look like.

Norbert 

[ReBoot]

The signal is supposed to have few wide peaks. As you can see in the screenshot, it consists of many narrow peaks which seems to group into fewer wider peaks.

[Norbert_B]

There are several possible reasons for the signal to be the way it is:

a) The "packages" seem to appear with a certain frequency. From the VI i asume that the samples are acquired with 1kHz for 5s. So the frequeny of the peaks seems to be (5 peaks/s) = 5 Hz. Is that a signal you expect?

b) Each peak has (in regard to your suspicion) high noise. That could be created by external noise (e.g. dimmer/transformer) or by grounding issues.

c) Due to your connection of the transducers, the system could include noise of any kind. Please make sure that the connections are properly made using shielded cables.

hope this helps,

Norbert 

[ReBoot]

First I thought that it's a software problem, that some buffer is full shortly after one of the wide peaks start so the y-value drops to 0 until it can record again. So you say that the error is not in the VI but noise?

I don't expect the packages to be at a certain frequency, I just expect them to be few wide peaks from the laser line instead of packages.

Message Edited by ReBoot on 01-16-2009 04:13 AM

[Norbert_B]

If you are not getting error messages, i'd say yes: it is an external issue.