This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DLP2010NIR: Does the instrument change the detection sensitivity during acquisition?

Part Number: DLP2010NIR

Hello everybody

I have a DLP NIRscan nano EVM purchased from Texas Instruments.

I made some scans with solutions. My goal is to get a quantitative analysis. My solutions are sucrose dissolved in h2o by different concentrations (1, 2, 3, 4, 5, 10, 15%).

Scanning first h2o (in trasmission mode, I purchased from Optiks the transmissive head) and then the solutions I get the spectra. I'm referring to a scientific article where the researcher did the same thing.

here the link of scientific article: link.springer.com/.../s12596-012-0102-0

Here is the issue: scanning the solutions, increasing the concentration of sucrose in water, I don't get a coherence. In fact increasing sucrose concentration the absorbance peaks should decrease in intensity (like described in paper) but I get a strange effect: increasing the sucrose concentration some peaks are decreased, other ones increased and other ones decreased but in a non-linear way.

My question is: is it possible that the instrument could change detector's sensitivity during a scan (maybe because of different light intensity received)? Is there some electronic control inside the instrument that manage the mathematical values of the spectra in some way?

I have to know this because I'm pretty sure that I'm not making mistakes in measurement setup, nor in cleaning the cuvette.

Hope you could give me a response soon

Thank you

  • Marco - we regularly use mixtures of dextrose and water (from 1% w/w dextrose to 40% w/w dextrose) as a convenient and cheap way of gathering spectral data from our DLP NIRscan-based spectrometer, so that we can experiment with different algorithms and data-preprocessing routines. Although we're using transflectance rather than transmittance, you should get good results either way. But I would strongly recommend that you use suitable multivariate maths to build your models, rather than relying on being able to pick peaks yourself. We usually find that the parts of the spectra that the computer ends up using, are often not what we (as humans) expect it to use. It shouldn't matter whether peaks increase or decrease, should it, if the end result is an accurate and precise model? kind regards, George

  • Dear George
    In my case I followed the same steps that are described in the paper whose I linked above.
    The issue consists not in my non-usage of maths because neither in the paper a preprocessing has been exploited, but in non-coherence present in the spectra.
    In the paper the researcher didn't use math and they got linear peak decrease with increase of sucrose. In my case this doesn't happen.
    My question to TI is if there is some electronic control of detector's sensitivity in acquiring light.
    I know the power of mathematical spectral treatment, but in my case preprocessing (like MSC) is useless because scattering is few. Moreover I'm focusing only on a wavelength (seeing the intensity peaks) so I want to make a simple linear regression on that wavelength and build up a correlation line.
    Thank's
  • Marco,

    The NIRscan Nano has a ADC PGA gain that is automatically set by the software. If the absorption of two different solution lies between PGA settings, this might cause a difference in the intensities reported. Also, lamp warm up and aging might account for difference in spectra between scans. The next version of the GUI allows setting the PGA to a fixed value as well as leaving the lamp on during scan to minimize these issues.

    As far as changing detector sensitivity, the software only controls the PGA setting of the ADC. The detector signal is amplified with a transimpedance amplifier and gained by the on-board PGA of the ADC. Refer to the application note Signal Chain Performance Optimizations in the TI DLP® Technology-Based Spectrometer (DLPA072) for more information on the signal chain.The detector sensitivity is dependent on the light reaching the detector. If the path length of the solution is too long, then water will absorb significant amount of light. If the path length of the solution is too short, then not enough of the sample is being scanned and light just goes right through it with little being absorbed. Thus, Sampling Techniques needs to be considered for repeatable and stable data.

    Best regards,
    Pedro