Problems with QE65000 spectrometer and Ocena Optics Driver

[Diegin]

Hi, a few months ago I installed the driver to use the USB4000, HR4000 and QE65000 spectrometers with LabVIEW 2010 with Windows 7-x64. Recently, I measured the spectrum of the HG-1 Mercury Argon Calibration Source with QE65000 spectrometer using either SpectraSuite Software and a LabVIEW program ("Ocean Optics 2000 4000 Acquire Continuous Waveform.vi, example of library, Fig1"). When I obtained both spectra I observed that with LabVIEW program there is a shift to infrared region in the spectrum. The calibration coefficients acquired are according with the "Wavelength Calibration Data" which comes with the spectrometer (Fig2). So, I would like to know if there are problems with the example program of the library or what is the problem?. Attached to the this message I send the image of the spectra obtained. The black graphic is the spectrum obtained with example program of LabVIEW including in ocean optics library. The red graphic is the spectrum obtained with SpectraSuite and which coincides with the real spectrum of the HG-1 Lamp. Regards, Diego

[Diegin]

Fig1

[Diegin]

Here, I post the figures because previously I had some problems with the internet.

[James_G1]

This is a cross post and I replied to the other post at http://forums.ni.com/t5/Instrument-Control-GPIB-Serial/Trubble-initiating-USB-2000-Ocean-optics-spec...

However, since this seems like the more proper post fot a solution I will repeat, (please forgive me forum gods)

There is a discrepancy between the LabVIEW drivers and the OOI drivers.  NI did not know that some 10 pixels on the Qe65000 are moved from the front to the back in our Omnidriver software.  This matters only because our manufacturing uses our driver internally to wavelength calibrate our devices, which means the calibration will be off by this amount when using the NI drivers.  The simple fix is to move the first 10 (I think its 10, please check) pixels to the back of the array in the "read spectra.vi".  I have put a quick fix together but need someone to verify it.  Everything from there onward should work fine.

Attachments:

Read Spectra.vi 57 KB

[James_G1]

There is a discrepancy between the LabVIEW drivers and the OOI drivers.  NI did not know that some 10 pixels on the Qe65000 are moved from the front to the back in our Omnidriver software.  This matters only because our manufacturing uses our driver internally to wavelength calibrate our devices, which means the calibration will be off by this amount when using the NI drivers.  The simple fix is to move the first 10 (I think its 10, please check) pixels to the back of the array in the "read spectra.vi".  I have put a quick fix together but need someone to verify it.  Everything from there onward should work fine.

[Albert.Geven]

Thanks James for updating this.

(I also replied to the other message first)

You indeed moved the pixels to the back, but with rotate (-10) pixels I get the same result, without rebuilding the array, this makes it 10 times faster.

Furthermore there is a problem with the merge error in this vi that is already solved on the NI webpage.

In fact the order inwhich the errors are merged is the wrong way around.

When not wiring a visa resource (and that happens sometimes) you will get the wrong errormessage (power needs to be cycled) instead of resource unknown.

So please update the internal used NI software.

PS I'm not an NI employee , just using ocean optics spectrometers.

[James_G1]

The rotate is faster than the rebuild, that makes sense but didn't occur to me.

As far as the error stuff I guess I modified an out of date copy in my rush to get a quick fix.  Thank you for a looking at these issues.  PS I am also not an NI employee.  I am an Ocean Optics employee but also a LV enthusiast.

James

[Diegin]

Thanks for your solution. In fact, I was trying to find what the problem was and as you mentioned it is neccesary to move 10 pixels. This information appears in the page 12 of the datasheet of QE65000 (http://www.oceanoptics.com/technical/engineering/OEM%20Data%20Sheet%20--%20QE65000.pdf). Here it is describe how the data should be reorganized. I think that the Sub.vi called "Read Spectra.vi" should be modified before to send the data to the variable "Spectrum by pixles", and check if it's neccesary to modify in a similar way the spectrum  for other kind of spectrometers.

Thanks so much for your help.

Best Regards

[Shawn_S. · Accepted solution]

Version 1.3.2 should be live in a few hours and this latest version implements the changes suggested previously.  Would someone mind trying these out and let us know how it goes?

Ocean Optics 2000 4000 Driver Page

http://sine.ni.com/apps/utf8/niid_web_display.download_page?p_id_guid=7833BD4A31DA1274E04400144FB7D2...

[Albert.Geven]

Hi Shawn

Fast response. Only the solution you implemented was not mine but a good catch from James_G from ocean optics.

My version used the rotate instead of the cut and build.

I can test this end next week, and will supply also some rewritten examples.

I simply hate the "wait until next multiple ms" that NI seems to use in all examples.

I will try to find my arguments on that again, but the most irritating behaviour is the undefined first wait, followed by : why should a multiple be necessary, a simple wait is good enough to keep the machine quiet!

I know that you did not change this but maybe you can help in the fight against using this wait mult in all examples.