Maker Chris Hackett, of Zoid Expertise, has demonstrated dramatically enhance the performance of a $150 spectrometer — with out having to spend a penny extra.
“Priced at below $150, it claims a measurement vary of 340 to 1,000nm, a decision of 7nm, and an accuracy of 1nm,” Hackett explains of the Torch Bearer spectrometer which is the main target of his work. “However, most significantly for me, it returns information in absolute models, with an marketed illumination accuracy of simply 4 p.c.”
With professional-grade gadgets costing upwards of a thousand {dollars}, Hackett was eager to place the Torch Bearer gadget to the check — however bumped into some early difficulties. First, its USB cable was really a USB to TTL serial cable; second, the corporate’s software program proved limiting with a set window dimension and solely primary performance. Worse, whereas the spectrometer would possibly output absolute models they are not simple to get out of the software program in any format however PDF.
The answer: ditching the producer’s software program in favor of some homebrew Python. “Torch Bearer gives protocol documentation for a few of their different fashions,” Hackett explains, “[and] whereas my mannequin had a number of variations, it was shut sufficient to place collectively a easy Python library and extract some information.”
Just a little hacking with the Android app delivered a approach to unlock extra superior options — together with absolute spectral information. (📷: Zoid Expertise)
That pulled information samples out with a 16-bit decision, decrease than the marketed 20 bits, however the spectrum information itself remained inaccessibly encrypted. For this, Hackett turned to the corporate’s Android app — disassembling and patching the software program to unlock options usually reserved for high-end fashions within the firm’s spectrometer line-up, together with CSV export, and to determine the encryption scheme to unlock the spectral information.
Hackett’s challenge is detailed in full within the video above and on the Zoid Expertise YouTube channel; the software program has been launched on GitHub below the reciprocal GNU Common Public License 3.
