Combining a Series of Sensors for a Cheap Spectrometer to Identify Everything

Need a reflectance spectrometer, but don’t want to pay a lot of money? You can either buy AMS’s AS7265X demo kit for about $100, or you can build your own like Kris Winer, for a much cheaper price tag of $25. Winer designed his own as part of this year’s Hackaday Prize, where contestants vie for cash prizes by showing ways on how technology can enrich humanity.

Reflectance spectrometers work using a pair of lights to measure the difference in reflectance on surfaces with different linear polarizations to identify materials. To put it more simply- different materials produce different wavelengths in varied amounts. Those wavelengths give materials their own specific fingerprints, making them easy to identify. Usually, reflectance spectrometers can cost an insane amount of money, but Winer is on his way to making them incredibly cheap, using just $25 worth of hardware.

Winer’s build is similar to the AMS AS7265X board, which uses three different optical sensors to accurately identify materials using a master-slave-slave configuration, giving it a total of 18 different channels in the near IR spectrum. Winer’s version, on the other hand, was designed using various AMS spectrometer sensors taken from the company’s other offerings.

To make his spectrometer as small as possible, he designed a custom PCB to house the different sensors taken from AMS’s AS7262 and AS7263 boards. Each offers a 6-channel near IR sensor but with different resolutions — 40nm FWHM resolution for the former and 20nm FWHM resolution for the latter. Those slave sensors will feed into an AMS AS72651 master sensor, which offers another 6-channels with a resolution of 20nm FWHM resolution.

For illumination, Winer chose a 5700 K 90 CRI LED and a pair of IR LEDs, giving him one broad 18-channel spectrum to identify and distinguish between organic compounds. Winer is still in the process of developing his spectrometer and needs to tackle the issue of what firmware he needs or whether the master bus requires pullup resistors. If he can overcome these issues, Winer states his design could become the base for a functional modern tricorder.