DLP® NIRscan™ Nano Evaluation Module - Review

Table of contents

RoadTest: DLP® NIRscan™ Nano Evaluation Module

Author: loglog

Creation date:

Evaluation Type: Independent Products

Did you receive all parts the manufacturer stated would be included in the package?: True

What other parts do you consider comparable to this product?: null

What were the biggest problems encountered?: The device is essentially field ready but due to the position of the control buttons and the exposed circuits, it is vulnerable to damage if precautions are not taken. However all the documentation required is available to adapt the device for field use.

Detailed Review:

Hello all,

Throughout my latest travels I have been able to complete my review of the DLP® NIRscan™ Nano EVM from Texas Instruments.  I received the device late January 2017 and was given some time to run the tests and review. A million thanks to element14 and Texas Instruments for this opportunity. 

 

From the onset the intent was to use the device for rock and soil engineering field applications. In my trade the assessment of rock and soil properties is performed through various methods of data logging, sampling, as well as field and laboratory testing.  Various portable devices are available for field use enabling the characterization of rock or soil masses whilst in the field. 

 

Spectroscopy

There are several types of spectrometers available for field use.  Portable devices such as X-ray fluorescence (XRF) analyzers have gained in popularity within geosciences (Figure 1) to conduct non-destructive qualitative and quantitative analysis of material composition.

Figure 1Portable XRF Analyser

 

Near-infrared (NIR) spectroscopy is a technique used to identify and characterize materials. NIR spectroscopy is also popular in mining and metallurgy, and new applications are developed with staggering speed for the identification and sorting of minerals during exploration, mining and processing.  The recorded spectra contain a variety of chemical and physical characteristics attributable to the sample and its constituents.  NIR absorption spectra can be complex and require special procedures for data analysis.  This technology relies on multi-variate analysis to correlate the recorded spectral response to known chemical or physical characteristics.

 

The DLP® NIRscan™ Nano EVM is designed to conduct NIR spectroscopy over wavelengths of 900 - 1700 nm using a highly portable assembly.  In simple terms, the sample is exposed to a light source, the light is passed through a diffraction grating, then through a series of small mirrors, and finally the absorbance of the light is measured by a detector (Figure 2).  Two tungsten filament lamps serve as the light source for the device.  The collection lens gathers collimated light from a region of the sample, which is to be placed directly against the sapphire window.  The broadband light goes through a slit, then the individual wavelengths of light are dispersed onto the TI DLP® Digital Micromirror Device (DMD) mirror array using a diffraction grating, allowing subsets of light to be mapped to specific wavelengths. Specific wavelengths of light can then be switched to a single-element InGaAs detector.

 

Figure 2 - NIRscan Nano™ EVM With Cover Removed (from Texas Instruments, Application Report DLPA062–January 2016)

 

Objectives

My objectives for this RoadTest were to:

  1. Evaluate the DLP® NIRscan™ Nano EVM hardware robustness for use in field work applications in rock and soil engineering
  2. Correlate the spectral response from the instrument to known mechanical characteristics for the materials tested.

This required the collection of data using selected standard field investigation practices with the addition of a collection of a population of spectral characteristics measurement samples for the purpose of statistical correlation. 

 

This put the evaluation device to the test, and was useful for my work.  The test was successful as a proof of concept, for a portable device capable of the assessment of material character in the field.  The device was used over several sites during the trial time period.

 

There was little time to prepare, just two days prior to my departure for work in the field.  Thankfully, the comprehensive documentation provided by Texas Instruments on their web site http://www.ti.com/tool/dlpnirnanoevm made it relatively easy to get familiar with the device and bring the necessary items with me for deployment. 

 

Hardware

Figure 3 shows the device out of its box.  It can communicate and be powered directly to a PC by means of a micro-USB cable.  This setup was useful in the laboratory.  I chose to power the DLP® NIRscan™ Nano EVM with a 5V 2600mAh portable backup battery and micro-USB connector.  The battery and thermistor I ordered ahead of the delivery did not arrive on time prior to deployment.  In any case, the device was capable of taking measurements and recording the data as directed.  A photo was taken whilst conducting a scan with the rock sample some distance away from the scanning window for demonstration purposes (Figure 4), this shows the device with the light source turned on.

 

The DLP NIRscan Nano EVM is a complete NIR spectrometer EVM using DLP technology. The EVM package includes

(source: Texas Instruments, Literature Number: DLPU030F, March 2016):

  • Near-infrared optomechanical spectrometer engine optimized for 900 to 1700 nm wavelength range:

– Reflective illumination module with two integrated infrared lamps

– 1.8-mm × 0.025-mm input slit

– Collimating lenses

– 885-nm long wavepass filter

– Reflective diffraction grating

– Focusing lenses

– DLP2010NIR DMD (0.2-inch WVGA, 854 × 480 orthogonal pixel, NIR optimized)

– Collection optics

– 1-mm single-pixel InGaAs non-cooled detector

  • Electronics subsystem with the electronics consisting of four boards:

– Microcontroller board

• Tiva TM4C1297 microprocessor for system control operating at 120 MHz

• 32MB SDRAM for pattern storage

• Power management with Lithium-polymer or Lithium-ion battery charging circuits using bq24250

• CC2564MODN Bluetooth Low Energy module for Bluetooth 4.0 connectivity

• USB micro connector for USB connectivity

• microSD card slot for external data storage

• HDC1000 humidity and temperature sensor

– DLP controller board

• DLPC150 DLP controller

• DLPA2005 integrated power management circuit for DMD and DLP controller supplies

• Constant current lamp driver based on OPA567 and monitored by INA213

– Detector board

• Low-noise differential amplifier circuit

• ADS1255 30 kSPS analog-to-digital converter (ADC) with SPI

• TMP006 thermopile sensor for detector and ambient temperature measurement

• 1-mm non-cooled Hamamatsu G12180-010A InGaAs photodiode

– DMD board

• DLP2010NIR near-infrared digital micromirror device.

 

The compact form factor for the hardware is a real advantage and its power requirement are low. 

 

Effectively the hardware can be used immediately.  The few downsides in the context of my rapid Roadtest trial were:

            -awkward position of interface buttons (Figure 7)

            -exposed circuits.

 

One may readily see the awkward position of the interface buttons as a problem, as well as the exposed circuits. The designers probably knew they were exposed and we can only conjecture that they did not see this as a problem.  In effect, the designers have stacked and dimensioned the PCB boards to ensure the sample may be aligned directly against the sampling window without being obstructed by the boards, which is great.  Furthermore the boards have an irregular hexagonal shape that matches the spectrometer head assembly, which provides an elegant and compact design for the evaluation module.  But the reason I raise the downside items is that substantive modifications must be made to the assembly in order to provide the necessary ingress protection for field use, as well as maintaining access to the Scan/Bluetooth button, as it is recessed below the spectrometer housing (Figure 7).

 

Figure 3 - Unpacking device

 

Figure 4 - Test assembly demonstration

 

Software

Both iOS and PC software were used as GUI.  The PC software NIRscan Nano GUI v2.0.2 was used to push the firmware updates to the DLP® NIRscan™ Nano EVM upon arrival and was used to control the device and interface with the data.  The GUI device status and error status traffic lights make it easy to spot any faults within the assembly.

 

The NanoScan iOS App v2.1 was uploaded to an iPhone running iOS 10.x.  It itself is functioning as intended to communicate via Bluetooth, however a glitch is preventing the correct display of Wavelength for each scan.  The data shows the Wavenumber on the x-axis regardless of the configuration setting chosen.  Nevertheless, the app saves the data with the correct attributes in .csv format, and provides the functionality for sending the .csv data via individual emails through the app.

Figure 5 - NanoScan iOS App v2.1 screen shot

 

Figure 6 - NanoScan iOS App v2.1 .csv data output processed using Microsoft Excel 2013

 

Figure 7 Position of Bluetooth activation button

 

Experiments

The experiments were conducted as planned. In essence Schimdt Hammer and pocket penetrometer (Figure 8) tests were correlated to spectral responses.  The first device provides an estimate of material strength for hard materials.  The tests were carried out in the laboratory (Figure 9) and in an underground mining environment (Figure 10).

The spectral responses were correlated to the field tool measurements using NeuralTools, an excel add-on distributed by Palisade. A large sample population is required for the learning algorithm to run correctly.  The data collected during the expedition was comparatively small and more test results are required. 

A key lesson from the preliminary testing was that the spectral range of the DLP® NIRscan™ Nano EVM does not capture significant spectral response differences for hard rocks, within the 900-1700 nm range.  It is possible those scans would be better suited for a broader NIR spectrum range from 1350 to 2450 nm using the DLP® NIRscan™ Evaluation Module. Significant spectral differences could be noted for porous rocks, soils and weathered rock materials, but those would benefit from a correlation to mechanical properties derived from laboratory testing instead of the simple pocket penetrometer apparatus brought along for the Roadtest experiments.

 

Note that no attempt was made to correlate the scanned spectrum samples to the spectrum library provided by Texas Instruments at this stage, as this was not the objective of this experiment.

 

Figure 8 Penetrometer ST 207 (above) and Schmidt Hammer (middle)

 

Figure 9 Laboratory setting whilst operating device using PC GUI on a weathered rock sample

 

Figure 10 Underground excavation setting, here an engineer is taking a Schmidt Hammer measurement, which was followed by a NIR scan

 

Conclusions

For this specific experiment the DLP® NIRscan™ Nano EVM has passed a proof-of-concept stage for use in field work applications in rock and soil engineering. An insufficiently large data sample in terms of quantity and character was collected so far to conclude on the ability to correlate the spectral response from the instrument to known mechanical characteristics for the materials tested.  However from a qualitative analysis it can be found that the spectral responses of porous and weathered rocks as well as soils were sufficiently distinctive from sample to sample to postulate that correlation to mechanical properties may be practicable.

Pros

-The form factor of the device is a positive factor, the device is packed with technology making it efficient for immediate use in the field

-The pre-compiled PC GUI software provided, which is also inclusive of the source code, is comprehensive and immediately effective to interface with the device

-Similarly the iOS GUI could be uploaded and installed without issues, and was immediately effective to interface with the device

-The cost of the device is comparatively low as compared to other portable spectrometer technologies

-The comprehensive software and documentation library offered by Texas Instruments makes it easy to adapt the device for the intended field use

-The ability to connect a 5V micro-USB power source was a clear advantage to enable immediate field use.

Cons

-The device is essentially field ready but due to the position of the control buttons and the exposed circuits, it is vulnerable to damage if appropriate precautions are not taken

-Protection of the device is difficult due to the position of the buttons, the device would require to be disassembled and re-assembled in a different configuration so it can be mounted in some form of casing and fitted with a battery and thermistor

-The iOS app spectrum graph does not show the correct wavelength from field scans and saved data.

Anonymous
  • Enrico,

    Here is a CSV file showing the CSV outputs of two scans on the same material. One scan done using Android app, the other using Windows 10 TI GUI.

     

    https://www.element14.com/community/docs/DOC-86930/l/comparing-desktop-and-mobile-apps-ti-dlp-nirscanzip

     

    Additionally, here are some images that simply show how the CSV outputs differ.

     

    Android App

     

    The output contains all intensity, absorbance, and reflectance data.

     

     

     

     

    Windows

     

    I can see the spectra for each but the CSV does not contain all 3. Meaning I need to go learn some math and derive the calculations myself. Or just use the Android app

    Note the windows CSV export contains a pretty heft header on top which describes the scan that was performed. I believe most of this data is also shared in the .dict file that is saved via the Android app.

     

  • Hi Jordan, thank you for the clarification

     

    Why don't post a zip with the two samples, just to take a look? I am curious.

    Maybe there is a reason correlating these two apparently so different kind of information.

    Enrico

  • Hi Enrico,

    One thing I didn't communicate well in my prior posts: The two data samples I shared are from different materials. I wasn't expecting to get the same absorbance values from both samples.

     

    What I was expecting was the desktop GUI to output absorbance, reflectance and intensity values, just like the Android app is doing. From briefly playing around in the Windows GUI, I can't find a way to tweak the export.

  • Hello Jordan,

     

    the reason of the difference is in the way you set the Absorbance column. I suppose - if you have used excel or OOO Sheet - that you have not set the number represantation (by default converted to integer) as long value with a decimal number set and the found method. If you have used this import, maybe you should set the column propertiece for Numeric values.

     

    Check the output format in the GUY settings too. Maybe something related to formatting or a tremendous bug

     

    Enrico

  • Hi David,

    When I export to CSV using Android app, I get following:

     

    WavelengthIntensityAbsorbanceReflectance901.707314711.7702310.016973

     

    However when I use the Windows GUI and export to CSV, I get the following:

     

    Wavelength (nm)Absorbance (AU)Reference Signal (unitless)Sample Signal (unitless)901.70730.57570186665

    23022

     

    Was wondering if you get a similar result?

     

    Edit: I should have noted that these are two different materials, so I expected absorbance values to be different. However I lament the fact that I cannot get intensity and reflectance signals saving via Windows GUI.

     

    Jordan

  • Hi!  Thanks for your question.  I did export to CSV from the app to obtain absorbance, intensity and reflectance.  See extract below (only one line):

         

    Serial Number WavelengthIntensityAbsorbanceReflectanceXXXXXX901.1742723110440.7670120.1709968

     

    Using the app, tap on the edit button and send the waveform by email to yourself.

     

    Regards

  • Thanks!  We will continue assembling a base data set for the present time...

    Regards

  • Thanks Gerald, I see there are some applications out there for this, what are you thinking of in particular?

  • Awesome review, and very cool applications!

     

    I am just wrapping up my review (very late, I know) and had a question. Did your export to CSV give you absorbance, intensity, and reflectance when you first ran it on computer with USB cable? Out of the box I am only getting absorbance values. I'm going to try updating firmware but I don't think this should be happening on the stock version.

     

    Thanks!
    Jordan

  • Very interesting. And very interesting approach to document the product and testing it.

     

    Enrico