that is incredibly smart. i never in my life would have gotten the idea to use LEDs for that. In fact i kinda didn't believe it would even work but you managed to prove me wron once more and made it work. congratulations that is just so freaking elegant. also great video as always! also that is the first time i have seen anything other than a diffraction grating/prism and (debayered) cam or linear CCD used in a fairly decently performing absorption or emission spectroscope. incredible
Thank you. There are some advantages: Low power consumption, no heat development by deuterium/halogen lamps, much cheaper, more durable, and almost maintenance-free. Perhaps it could be used in a space probe someday. The main disadvantage is of course the small selection of wavelengths, which I try to compensate by interpolation. Perhaps one day there will be LEDs whose emitting wavelength can be controlled by voltage etc. Oh no, yet another idea... 😅
Excellent project I have been fiddling around spectrophotometer for few years I have few McBeth color eye I’ve attempted to build a solid state spectrophotometer it is work in progress your project gave me ideas , thanks
Thanks for the hint with the background music. In my video editor the background music is much quieter than on UA-cam, which is strange. I will reduce it further in my next video:)
@@blenderbuch Noted for the next video. I think, I completely mute the music when I'm speaking and only use it as a filler when I'm not speaking for a long time.
Really cool to see how you design and built this thing. I have learned a lot but also that it doesn´t make a lot of sense so emit a specific wavelength and measure intensitiy, instead of measure the absorbanse of polychromic light, because of the few data points
The idea is to compensate for the few data points by interpolation. I could also add a lot more LEDs to the ring. No optics means the spectrometer could be used in a very harsh environment.
Fantastic project - I'd love to implement my own with more LEDs for determining caffeine, are there project files for the PCB or other components hosted somewhere?
Thank you. :) That's a cool idea with caffeine. I guess you need more LEDs in the UV range. The project files are not available yet because I have to modify some minor things (for example, clean up the code), and I'm busy with projects ahead. Once the project files are available, I'll post a link in the community section and in the video description.
Thank you. I'm very interested in that. What substances is the blood analyzed for? I am currently doing research on one of my next projects, a blood type analyzer.
@@Marbslab in hematology, hemoglobin is measured by photometry (Green led light absorbance is proportionaly to hemoglobin concentration in a diluented solution with lised red blood cells). In chemistry, the serum react with a colorimetric reactive for some time and then produce a color proportionaly or inversely proportionaly to the analyte. With this metod we can measure creatinine, urea, bilirrubin, glicose. For exemple the serum creatinin react with NaOH and picric acid (Jaffe metod) and produce a red color in the sample, this color is measured by the photometer. The system is calibrated with a know result sample.
Great video, I just have a question - since you can use color controllable leds, why not just use one and change the color of it to change the spectrum?
Thank you:) To answer your question, color-controllable LEDSs are based on the RGB system and can't be used as it's just a mix of the 3 wavelengths. The only way to change the wavelength of an LED is temperature, but the shift is small.
Very nice job. However I don’t think you’ll ever get good results from interpolating. Absorption spectra can be pretty narrow. I doubt you’ll find LEDs to get decent coverage of the spectrum.
Today, there are many wavelengths available for LEDs. You can find a list here: en.wikipedia.org/wiki/Light-emitting_diode#Colors_and_materials As the wavelength of an LED only changes with the temperature, the LEDs could be cooled or heated with a Peltier element. Interpolation is a broad term, for example Gaussians, etc. could also be used.
@ I work on a Wet General Chemistry Analyzer with a Xeon Lamp Module. No cooling, it is incased. The Legacy Analyzer before it was Halogen. :) love your work! :)Best Regards!
ooh i've always wanted to make a mini spectrometer, but I decided on a glass transmission diffraction grating and a tungsten lamp as the light source, which presents its own problems I did not know that mini spectrometer breakboards existed (my intention was to use photodiodes and build an amplifier)
Thanks:) The tubing doesn't matter. The measurement is done with a flow through cell that is connected to the tubing. You can see the flow through cell in the video. I bought it on aliexpress: de.aliexpress.com/item/1005005792925624.html
This project is freaking cool, kudos man. I haven't well grasped how you manage the interpolation. Let's say you're probing a sample that absorbs just a single wavelength, and you don't have any diode that emits such a wavelength, how do you get into the middle? P.s. Music covers your voice too much, in particular at the beginning of the video. I'd suggest you to boost your voice by 4 to 6 dB, since the overall level of sound is not too high. When editing check the level bar and alternatively mute each track to see what contribute the most. In general bground music should be kept at -10 .. -22 dB. For this kind of video style I'd stay at -22dB or so. While the voice (alone) should peak at 0 to +1 dB. Hope to be helpful. Merry Christmas.
Interpolation works better the more data points are available. Interpolation is not possible for one point, as there are an infinite number of functions that contain this point. Cubic spline interpolation for example gives you a set of functions, each of which describes the curve between two points. Since you have now calculated these functions based on the given data points, you can also use other, unmeasured wavelengths and calculate the function value (absorbance) at that unmeasured wavelength. Many thanks for the video editing tips. The background music is set very low in my video editing program. I can only guess, but it might have something to do with UA-cam's new auto-sync feature. I also wish you a Merry Christmas and a Happy New Year.
@Marbslab my point was: if you have no data because the only line absorbed if far from any wavelength emitted by any diode, how can you interpolate in this case? Hope I made my point clear, and it assumes that the diodes make a pretty narrow band around their emitted primary wavelength. From the video I found it not clear if their emission bands overlap.
@@AccidentalScience The following wavelengths are currently featured: 390 nm 470 nm 570 nm 588 nm 600 nm 628 nm Interpolation is only possible in the interval [390 nm, 628 nm]. No interpolation is possible outside this range. However, there is another mathematical method called extrapolation. With this method, statements outside the interval are also possible.
@@Marbslab ok, so for example the bandwidth of both 390 and 470 nm is large enough to overlap a bit so if you have a sample that absorb at, let's say, 410nm you can still detect it.
@@AccidentalScience You can only measure the absorbance of the six mentioned wavelengths. Every not-measured absorbance in the range of 390 to 628 nm can be estimated by interpolation. For example, if you want to know the absorbance at 410 nm, plug in the value in the according interpolation function f(410 nm) and calculate the function value. Here you can play a bit with cubic spline interpolation online: tools.timodenk.com/cubic-spline-interpolation
What is the possible resolution of this? I was wondering if it could be used in agriculture for soil analysis, but the non-homogeneous nature of the soil makes it very difficult, after that I remembered the fact that deficiencies in the soil can be diagnosed based on the coloration of the plant (chlorophyll)
Currently, I only use 6 LEDs ranging from 390 to 628 nm. All intermediate values are interpolated using the cubic spline and other interpolation methods. The resolution is therefore arbitrarily high in this range, but you have to consider the error that the interpolation entails. However, explaining the error estimate in more detail here would go far beyond the scope of this answer. Unfortunately, I have no experience with soil spectroscopy, but it sounds very interesting.
Thanks:) The number of LEDs is limited by the semiconductor material. Here you can find a list of available wavelengths for LEDs based on different materials: en.wikipedia.org/wiki/Light-emitting_diode
I think I made my motivation clear in the video. As mentioned there, you need a deuterium lamp and a tungsten halogen lamp if you want to cover the full UV-VIS spectrum. I wanted to try a different approach and integrate interpolations that give information about absorbances that were not measured at all. Instead of the CCD sensor, you can also use a camera and ready-made software. Everything is fine, but I wanted to try something new :)
Thats a great video man. I have a question as i am not much familiar with this topic as you made a uv/vis spectrophotometer , so if you replace the led's with the leds from uv and ir range and select the sensor accordingly can you achieve the same result?
Thank you:) Each sample absorbs different wavelengths to a greater or lesser extent. So if you use other wavelengths, the results will not be the same, but they will complement the other wavelengths. Interpolation works better the more supporting points you have in an interval of wavelengths.
'Spectrometer' usually refers to 'Spectrophotometer' so I went into this video expecting that. When I saw the description of 'uses several LED colors' I thought they were being used as detectors or something.
I used discrete LEDs, each with a different wavelength, instead of the full spectrum by a prism or a grating. Many more LEDs could be added. I then used the measuring points obtained from the LEDs as supporting points for various interpolations. In the end, my spectrometer contains a graphing calculator. The starting point for the design was not to use any optics such as prisms, gratings, mirrors or lenses.
i dont want to build it, i want to buy it. gia charges 5000 euro for theyr machine. can i buy yours ? or do you know where i can get one i can use for gemstones ... ?
@@QuadDerrick Merry Xmas. I did a bit of research, and you are right: spectrometers for gemstones are also very expensive. A pretty useful machine for gemstone sellers and buyers. Maybe I come back to it in the future. Greetings to Norway from Germany🙂
@QuadDerrick I'm gonna guess he means that his machine is built for fluids, and gemstones typically aren't very fluid haha. I think you'd need to cad out a holder for the stone, with an aperture that the stone would be able to cover (probably would need a rubber nipple on the aperture that the stone would be pressed against, to prevent light bypassing the stone). The neat thing about his design tho is that with a little research you could find the critical wavelengths needed to measure your stones, and put the appropriate leds on the wheel (which I think is the real beauty of his approach, it's superior for specialized applications like that where you know in advance the wavelengths you need to test).
@@xxportalxx. when gem buyers use a refractometer, they use a little drop of refractive oil just on the stone and press it up against the glass to get a reading. i am guessing it is somewhat the same deal but, i am very little scientific despite looking at super advanced "light machines" on youtube. , u can tell my by use of the word "light machine" i think =) i am dumb in many ways, i know that, yet, i ,, feel,, like that 5000 dollar machine to GIA, that keeps the worlds gem marked tied to GIA and they'r certification , is not worth more than 50 dollar.. To me it looks,, and sounds like, by you even, that this is the case too. Tell me if you agree or not and, if that is not a disgraceful thing ? Howcome all the worlds smartest youtube nerds cant beat GIA and put together a machine in the garage for 50$ like GIA's 5000$ machine ?
Lol. I could say that exactly like Schwarzenegger did with his accent. Btw., check out the very end of this video I made a while ago: ua-cam.com/video/Mp5hzbFdCfY/v-deo.html
Interpolation will never work for a precision spectrometer. Your device could be useful as a go/no go quality monitor in a production line for materials of known spectra.
It depends strongly on the monotonicity of the function to be approximated (in this case absorbance as a function of wavelength). In intervals in which the function is strictly monotonically increasing or decreasing, cubic spline interpolation provides very good approximations. I have never claimed that my spectrometer is highly precise. You can adapt it to different applications by changing the LED rings.
Have you ever considered that no optics means that the spectrometer could be used in a very harsh environment (e.g. high vibrations etc.)? Of course I could have used a prism, but I wanted to try something new.
I could also attach a white LED to the LED ring, then the spectrometer could also be used as a colorimeter. I also thought about using a supercontinuum laser.
@@user255 I will definitely look into the subject further, as it would also be very interesting for other applications. You could also use a concave mirror and arrange the LEDs at the edge accordingly. That would eliminate the stepper motor.
@@user255 CTOLEDs would also be worth considering. Or thermochromic substances, or fluorescent substances dissolved in various solvents. So many new ideas...⚗🧪🔬
Yes, interpolation helps. It's a common method to predict unmeasured values. As mentioend in the video, the graph needs to be normalized. It's easily done with a mapping function, changing the sensitivity of the sensor or deluting the chlorophyll solution further.
![Lp. Сердце Вселенной #60 РОЖДЕНИЕ ЛОЛОЛОШКИ [Финал] • Майнкрафт](http://i.ytimg.com/vi/YoR0pAV9FVQ/mqdefault.jpg)
that is incredibly smart. i never in my life would have gotten the idea to use LEDs for that. In fact i kinda didn't believe it would even work but you managed to prove me wron once more and made it work. congratulations that is just so freaking elegant. also great video as always! also that is the first time i have seen anything other than a diffraction grating/prism and (debayered) cam or linear CCD used in a fairly decently performing absorption or emission spectroscope. incredible
Thank you. There are some advantages: Low power consumption, no heat development by deuterium/halogen lamps, much cheaper, more durable, and almost maintenance-free. Perhaps it could be used in a space probe someday. The main disadvantage is of course the small selection of wavelengths, which I try to compensate by interpolation. Perhaps one day there will be LEDs whose emitting wavelength can be controlled by voltage etc. Oh no, yet another idea... 😅
This is absolutely brilliant!!!
Thank you:)
Great project and great video and great result .Congrulation .
Thank you:)
I came to this video and channel through a Microchip post / reshare. Absolutely no regrets. Nicely done!
Thank you:) We share the passion for electronics and have a skull in our channel icons. What could go wrong? Greetings to Greece.
@@Marbslab what could go wrong? Everything, hehehe 😂.
Thanks and many greetings to you too!!!
i love how casual and matter of fact you are about creating something so amazing! what you do is what i aspire to
Thank you:)
I have a motto: a day on which I learn nothing is a day lost. Learn, try, fail, repeat. Ignore nay-sayers. Do what you really enjoy.
Excellent project I have been fiddling around spectrophotometer for few years I have few McBeth color eye I’ve attempted to build a solid state spectrophotometer it is work in progress your project gave me ideas , thanks
Glad to hear that you have gotten some ideas. Good luck with your project:)
Super impressive work, well done. Thanks for sharing
Sorry but the background music is very high compared to your voice. But the end result is amazing !! Well done !!!
Thanks for the hint with the background music. In my video editor the background music is much quieter than on UA-cam, which is strange. I will reduce it further in my next video:)
Also the bg music (which i prefer not to have anyhow) pumps in volume which makes it even more disturbing. Other than that I like your work very much.
@@blenderbuch Noted for the next video. I think, I completely mute the music when I'm speaking and only use it as a filler when I'm not speaking for a long time.
I'd say the music was clearly not whispering, but I actually enjoyed it, and it didn't distract from hearing the voice at all
@@barmalini Thanks for the feedback. Next year I will probably invest some money into a better microphone and a camera slider.
Super Projekt einfach klasse bitte mehr von dem format guten Rutsch ins neue Jahr und viel Erfolg für kommende Projekte
Vielen Dank. Es wird mehr von diesem Format geben. Dir auch einen guten Rutsch!
Really cool video, nice way to implement interpolation
Thank you. For me, the interpolation was the most interesting part of the project, and I already have some ideas to use it in other projects:)
Really cool to see how you design and built this thing. I have learned a lot but also that it doesn´t make a lot of sense so emit a specific wavelength and measure intensitiy, instead of measure the absorbanse of polychromic light, because of the few data points
The idea is to compensate for the few data points by interpolation. I could also add a lot more LEDs to the ring. No optics means the spectrometer could be used in a very harsh environment.
Fantastic project - I'd love to implement my own with more LEDs for determining caffeine, are there project files for the PCB or other components hosted somewhere?
Thank you. :) That's a cool idea with caffeine. I guess you need more LEDs in the UV range. The project files are not available yet because I have to modify some minor things (for example, clean up the code), and I'm busy with projects ahead. Once the project files are available, I'll post a link in the community section and in the video description.
I work with blood automated chemistry analyzer. Same principle, Nice video.
Thank you. I'm very interested in that. What substances is the blood analyzed for? I am currently doing research on one of my next projects, a blood type analyzer.
@@Marbslab in hematology, hemoglobin is measured by photometry (Green led light absorbance is proportionaly to hemoglobin concentration in a diluented solution with lised red blood cells). In chemistry, the serum react with a colorimetric reactive for some time and then produce a color proportionaly or inversely proportionaly to the analyte. With this metod we can measure creatinine, urea, bilirrubin, glicose. For exemple the serum creatinin react with NaOH and picric acid (Jaffe metod) and produce a red color in the sample, this color is measured by the photometer. The system is calibrated with a know result sample.
@@Marbslab blood type abalyzers in general works with a antibody solution, this reacts with the sample and a aglutinação can be observed
@@LuizFernando-ft8qp Thanks a lot for this detailed explanation:)
Great video, I just have a question - since you can use color controllable leds, why not just use one and change the color of it to change the spectrum?
Thank you:) To answer your question, color-controllable LEDSs are based on the RGB system and can't be used as it's just a mix of the 3 wavelengths. The only way to change the wavelength of an LED is temperature, but the shift is small.
Hey love your HIGH QUALITY project!
-> i just have liked and subscribed
Thank you:)
Nice and neat , congratulations.
Thank you. Working already on two new projects...
Very nice job. However I don’t think you’ll ever get good results from interpolating. Absorption spectra can be pretty narrow. I doubt you’ll find LEDs to get decent coverage of the spectrum.
Today, there are many wavelengths available for LEDs. You can find a list here: en.wikipedia.org/wiki/Light-emitting_diode#Colors_and_materials
As the wavelength of an LED only changes with the temperature, the LEDs could be cooled or heated with a Peltier element. Interpolation is a broad term, for example Gaussians, etc. could also be used.
good idea, good craftsmanship, I like it.
Thank you:) Working already on two new projects.
Wow, this is super! Would you share the build files so I could build one? Thanks!
Thank you:) I'll share the files if I have a little bit more time (the 3-D drawing needs some minor changes).
Cool idea. Can you use absorption output as full gaussians instead of just peak intensities? Seems that a lot of information is lost that way
Thanks. Sure you can use Gaussians. If I remember right, somebody else in the comments already suggested it.
solid engineer. fun to watch craft.
Thank you:)
Xeon lamps work well also for the VIS LightRange.
Yeah, but may require convection cooling, and the life time is not great (2000 h).
@ I work on a Wet General Chemistry Analyzer with a Xeon Lamp Module. No cooling, it is incased. The Legacy Analyzer before it was Halogen. :) love your work! :)Best Regards!
@@rmgibbs1861 Thanks for your insights:)
ooh i've always wanted to make a mini spectrometer, but I decided on a glass transmission diffraction grating and a tungsten lamp as the light source, which presents its own problems
I did not know that mini spectrometer breakboards existed (my intention was to use photodiodes and build an amplifier)
something neat is the use of the polynomial interpolation
however i think you may want to switch to a clear tubing (if that exists)
Everything has pros and cons, depending much on for what you want to use the spectrometer.
Thanks:) The tubing doesn't matter. The measurement is done with a flow through cell that is connected to the tubing. You can see the flow through cell in the video. I bought it on aliexpress: de.aliexpress.com/item/1005005792925624.html
@@Marbslab nice
@@Marbslab that's fair
i mostly want to build it because it seems like something neat (The monochromator is sometimes useful)
Nice concept! But no code to share for curiosity? Schematics? Maybe I'm missing a link because of the device I'm viewing this on.
Thanks:) I will share all STL files, schematics, gerbers and source code but it can take a while. Link will be in the video describtion later.
This project is freaking cool, kudos man. I haven't well grasped how you manage the interpolation. Let's say you're probing a sample that absorbs just a single wavelength, and you don't have any diode that emits such a wavelength, how do you get into the middle?
P.s. Music covers your voice too much, in particular at the beginning of the video. I'd suggest you to boost your voice by 4 to 6 dB, since the overall level of sound is not too high. When editing check the level bar and alternatively mute each track to see what contribute the most. In general bground music should be kept at -10 .. -22 dB. For this kind of video style I'd stay at -22dB or so. While the voice (alone) should peak at 0 to +1 dB. Hope to be helpful. Merry Christmas.
Interpolation works better the more data points are available. Interpolation is not possible for one point, as there are an infinite number of functions that contain this point. Cubic spline interpolation for example gives you a set of functions, each of which describes the curve between two points. Since you have now calculated these functions based on the given data points, you can also use other, unmeasured wavelengths and calculate the function value (absorbance) at that unmeasured wavelength.
Many thanks for the video editing tips. The background music is set very low in my video editing program. I can only guess, but it might have something to do with UA-cam's new auto-sync feature.
I also wish you a Merry Christmas and a Happy New Year.
@Marbslab my point was: if you have no data because the only line absorbed if far from any wavelength emitted by any diode, how can you interpolate in this case? Hope I made my point clear, and it assumes that the diodes make a pretty narrow band around their emitted primary wavelength. From the video I found it not clear if their emission bands overlap.
@@AccidentalScience The following wavelengths are currently featured:
390 nm
470 nm
570 nm
588 nm
600 nm
628 nm
Interpolation is only possible in the interval [390 nm, 628 nm]. No interpolation is possible outside this range. However, there is another mathematical method called extrapolation. With this method, statements outside the interval are also possible.
@@Marbslab ok, so for example the bandwidth of both 390 and 470 nm is large enough to overlap a bit so if you have a sample that absorb at, let's say, 410nm you can still detect it.
@@AccidentalScience You can only measure the absorbance of the six mentioned wavelengths. Every not-measured absorbance in the range of 390 to 628 nm can be estimated by interpolation. For example, if you want to know the absorbance at 410 nm, plug in the value in the according interpolation function f(410 nm) and calculate the function value. Here you can play a bit with cubic spline interpolation online: tools.timodenk.com/cubic-spline-interpolation
What is the possible resolution of this? I was wondering if it could be used in agriculture for soil analysis, but the non-homogeneous nature of the soil makes it very difficult, after that I remembered the fact that deficiencies in the soil can be diagnosed based on the coloration of the plant (chlorophyll)
ua-cam.com/video/DofVCfqWi08/v-deo.html&ab_channel=PrecisionAgCoP not completely the same but interesting
Currently, I only use 6 LEDs ranging from 390 to 628 nm. All intermediate values are interpolated using the cubic spline and other interpolation methods. The resolution is therefore arbitrarily high in this range, but you have to consider the error that the interpolation entails. However, explaining the error estimate in more detail here would go far beyond the scope of this answer. Unfortunately, I have no experience with soil spectroscopy, but it sounds very interesting.
Lovely video. You could use 1000 LEDs...
Thanks:) The number of LEDs is limited by the semiconductor material. Here you can find a list of available wavelengths for LEDs based on different materials: en.wikipedia.org/wiki/Light-emitting_diode
Nice, but why not using linear CCD modules (like TCD1304) + diffraction grating? It would give much more accuracy and resolution
I think I made my motivation clear in the video. As mentioned there, you need a deuterium lamp and a tungsten halogen lamp if you want to cover the full UV-VIS spectrum. I wanted to try a different approach and integrate interpolations that give information about absorbances that were not measured at all. Instead of the CCD sensor, you can also use a camera and ready-made software. Everything is fine, but I wanted to try something new :)
Amazing! Thanks for sharing
Thats a great video man. I have a question as i am not much familiar with this topic as you made a uv/vis spectrophotometer , so if you replace the led's with the leds from uv and ir range and select the sensor accordingly can you achieve the same result?
Thank you:) Each sample absorbs different wavelengths to a greater or lesser extent. So if you use other wavelengths, the results will not be the same, but they will complement the other wavelengths. Interpolation works better the more supporting points you have in an interval of wavelengths.
what is the name of the button you are using
It's a mini joystick (5-way tactile switch): de.aliexpress.com/item/4000681535647.html
@@Marbslab thanks
There is no audio it seems! was it claimed music?
Please check your settings. Video audio works. No copyright issues. I only use music from the UA-cam audio media library.
'Spectrometer' usually refers to 'Spectrophotometer' so I went into this video expecting that. When I saw the description of 'uses several LED colors' I thought they were being used as detectors or something.
I used discrete LEDs, each with a different wavelength, instead of the full spectrum by a prism or a grating. Many more LEDs could be added. I then used the measuring points obtained from the LEDs as supporting points for various interpolations. In the end, my spectrometer contains a graphing calculator. The starting point for the design was not to use any optics such as prisms, gratings, mirrors or lenses.
i dont want to build it, i want to buy it. gia charges 5000 euro for theyr machine.
can i buy yours ? or do you know where i can get one i can use for gemstones ... ?
Sorry, you can't buy mine. My spectrometer approach is still under research. It also can't be used for gemstones.
@@Marbslab one can not buy anything, not even for xmas.
if i was smart i would ask why its no good for gemstones but, im not.
happy xmas from norway
@@QuadDerrick Merry Xmas. I did a bit of research, and you are right: spectrometers for gemstones are also very expensive. A pretty useful machine for gemstone sellers and buyers. Maybe I come back to it in the future. Greetings to Norway from Germany🙂
@QuadDerrick I'm gonna guess he means that his machine is built for fluids, and gemstones typically aren't very fluid haha. I think you'd need to cad out a holder for the stone, with an aperture that the stone would be able to cover (probably would need a rubber nipple on the aperture that the stone would be pressed against, to prevent light bypassing the stone). The neat thing about his design tho is that with a little research you could find the critical wavelengths needed to measure your stones, and put the appropriate leds on the wheel (which I think is the real beauty of his approach, it's superior for specialized applications like that where you know in advance the wavelengths you need to test).
@@xxportalxx. when gem buyers use a refractometer, they use a little drop of refractive oil just on the stone and press it up against the glass to get a reading. i am guessing it is somewhat the same deal but, i am very little scientific despite looking at super advanced "light machines" on youtube. , u can tell my by use of the word "light machine" i think =)
i am dumb in many ways, i know that, yet, i ,, feel,, like that 5000 dollar machine to GIA, that keeps the worlds gem marked tied to GIA and they'r certification , is not worth more than 50 dollar..
To me it looks,, and sounds like, by you even, that this is the case too.
Tell me if you agree or not and, if that is not a disgraceful thing ?
Howcome all the worlds smartest youtube nerds cant beat GIA and put together a machine in the garage for 50$ like GIA's 5000$ machine ?
Interesting, I would prefer a written tutorial.#
Great job
Could you end your next video by saying, "I'll be back"? (Also, cool spectrometer.)
Lol. I could say that exactly like Schwarzenegger did with his accent. Btw., check out the very end of this video I made a while ago: ua-cam.com/video/Mp5hzbFdCfY/v-deo.html
Interpolation will never work for a precision spectrometer. Your device could be useful as a go/no go quality monitor in a production line for materials of known spectra.
It depends strongly on the monotonicity of the function to be approximated (in this case absorbance as a function of wavelength). In intervals in which the function is strictly monotonically increasing or decreasing, cubic spline interpolation provides very good approximations. I have never claimed that my spectrometer is highly precise. You can adapt it to different applications by changing the LED rings.
why not just use a prism
Have you ever considered that no optics means that the spectrometer could be used in a very harsh environment (e.g. high vibrations etc.)? Of course I could have used a prism, but I wanted to try something new.
wow
Thanks:) Working already on two new very cool projects...
Instead of expensive special lamps, you could use full spectrum white led, but of course then you would need diffraction grating and still no UV & IR.
I could also attach a white LED to the LED ring, then the spectrometer could also be used as a colorimeter. I also thought about using a supercontinuum laser.
@@Marbslab True! Supercontinuum laser would have been interesting as well, but UV range could be difficult still.
@@user255 I will definitely look into the subject further, as it would also be very interesting for other applications. You could also use a concave mirror and arrange the LEDs at the edge accordingly. That would eliminate the stepper motor.
@@Marbslab Good idea!
@@user255 CTOLEDs would also be worth considering. Or thermochromic substances, or fluorescent substances dissolved in various solvents. So many new ideas...⚗🧪🔬
no. just no. Interpolation doesn't help for this. How can the transmission go above 100% which is the case in your device?
Yes, interpolation helps. It's a common method to predict unmeasured values. As mentioend in the video, the graph needs to be normalized. It's easily done with a mapping function, changing the sensitivity of the sensor or deluting the chlorophyll solution further.