Ocean Optics Flame Sepctrometer with Labview drivers

[fiskpralin]

I am successfully using the NI drivers "Ocean Optics 4000 2000" to control multiple OO spectrometers. Have recently acquired their new "flame" model but I don't find it in the list of supported spectrmeters. Do anyone know if it can be controlled using the drivers or if there will be a fix soon?

[Eirikur]

Hi fiskpralin,

Is this the driver you are looking for?

[fiskpralin]

Hey Eirikur,

Thanks for your answer, and yes it is exactly the one I am already using. At the time I posted the question, the Flame model was not included in the list of supported spectrometers, now it is; I received prompt reply from NI who informed me that they are going to add it to the list of supported spectrometers. They further told me that it it runs exactly like a USB2000+. In fact, because the drivers have not been updated since august last year, I assume it is not actually an option in the code, but should be used as a USB2000+.

N.B. the Flame-s uses the same sensor as USB2000+ and can thus be used as one, whilst the -t model has a different sensor with another number of pixels and so needs some other settings which I have not looked into.

[Wöcki]

hi,

i just want to add a small footnote to this topic

you migth have a problem if you were already using older ocean optics spectrometers (USB2000, or 2000+ as I did) with an older driver version (before REV 1.4, 11/15/2013) and than update to the actual driver version REV 1.4.1, 08/05/2014, because the measured spectra won´t match

This is due to the reason that in the older versions the "Read Spectra VI" sent the whole measuerd spectra-arrays (from start index 0 to "Number off pixels") to the "Spectrum by Pixels"-output.

The new "Read Optical Spectrum VI" in the new driver versions cut off some parts off this spectrum (dependent on the type of your spectrumeter; e.g. USB2000: 26; USB2000+ (Flame): 20) I don´t know for what reason. It is called "start optical pixel"

As this "cut off" is not done in the VI "Calculate Wavelength" you get a shift off the spectrum, if you match the array "Spectrum by Pixels" with the array "Wavelength (nm)" in a XY-Graph, by these 20/26 measuring points

I solved the problem by removing also the first 20/26 array-elements off the "Wavelength"-array but maybe NI can think about these "start optical pixel"-values they suddenly added in the new "Read Optical Spectrum VI"

best regards

martin

[Albert.Geven]

Hi

I'm working together with NI to update the ocean optic spectrometer driver.

The fact that at the moment it is not OK stemms from a misunderstanding where Ocean Optics starts calculating the wavelength.

In a few weeks a new driver can be born and a compatible mode with the old driver will be available.

PS Im not working for NI but  Philips in Eindhoven, Netherlands.

[James_G1]

Dear all,

I am a senior applications scientist at OOI and have been here for 7 years.  I am also a fan of LabVIEW.  I assisted with the initial driver creation. 

I had not updated my NI drivers for a while until recently and noted this separation of "optical" pixels from "dark" pixels.  I did notice that the wavelengths were off as a result of this change.  I would like to make myself available to try to clear up any questions about our devices and contribute to this ongoing development effort.

I am also have some question to the group.  How are these drivers tested for example?  I actually use them (the old ones) pretty regularly for internal purposes but I certainly have not chased down every use case. Is anyone working on these drivers also making test code to test the drivers?  What would test code look like, how automated could it be?

Many devices cross my desk so I could do some testing on many different hardware as long as the tests are fairly easy to run.

James

[Albert.Geven]

You can use the USB200+ selection for the flame spectrometers.

The next version of the drivers will have the name in place but you still have to select the usb2000+.inf file.

[fiskpralin]

So a pixel shift of ~20 pixels should correspond to a 10 nm shift in the spectrum. Do I get you guys correctly, that a quickfix for this issue would be to set the input number of pixels to 2048 (in the case of USB2000+ or flame), and then crop the first 20 positions of the "Wavelength" array from "Calculate wavelength?

[fiskpralin]

Really good observation above, thanks! But actually, to get the wavelength array correct, I needed to reset the "start optical pixel" in the driver to 0 instead of 20. Thereafter I needed to collect 2048 pixels instead of 2028 (for USB 2000+ or Flame) from the "Get wavelength VI". Finally, I cropped both the first 20 entries in both the "get optical pixels" and the "wavelength" arrays.

That way I could correct the wavelength shift, and verify this by checking so that the peaks from our fluorescent tubes are in exactly the same position as when the spectrometer was run in Spectrasuite. Thanks again for this important notice!

[revealdoc]

Hi,

I am having difficulty with the LabVIEW/Ocean Optics "Read Spectrum.vi" not giving the same spectrum data array on an immediate second request for a spectrum; i.e. the same vi is just repeated.  When I subtract the first array from the second array, the result looks like an attenuated spectrum of the broadband light source, with a maximum difference of about 1000 counts (I am expecting the difference to be a near-zero "spectrum difference").  I am using an Ocean Optics Flame-S spectrometer and LabVIEW 2016.

Thanks,

Guy Hatch

CTO Reveal Biosensors, Inc.

(435) 757-8081