Classic Colorimeter Clone Calibrates Cuvettes’ Contents

A home-made colorimeter

For anyone dabbling in home chemistry, having access to accurate measurement equipment can mean the difference between success and failure. But with many instruments expensive and hard to find, what’s a home chemist to do? Build their own equipment, naturally. [Abizar] went ahead and built himself a colorimeter out of wood and spare electronic components.

A colorimeter (in a chemistry context) is an instrument that determines the concentration of a solution by measuring how much light of a certain wavelength is absorbed. [Abizar]’s design was inspired by the classic Klett-Summerson colorimeter from the 1950s, which uses a light bulb and color filters to select a wavelength, plus a photoresistor to measure the amount of light absorbed by the sample. Of course, a more modern solution would be to use LEDs of various colors, which is exactly what [Abizar] did, although he did give it a retro touch by using an analog meter as the readout device.

The body of the colorimeter is made from laser-cut pieces of wood, which form a rigid enclosure when stacked together. The color wheel holds eleven different LEDs and is made with a clever ratchet mechanism to keep it aligned to the cuvette, as well as a sliding contact to drive current into the selected LED. All parts are painted black to prevent stray reflections inside the instrument, but also make it look cool enough to fit in any evil genius’s lab. In the video embedded below, [Abizar] demonstrates the instrument and shows how it was put together.

While we haven’t seen anyone make their own colorimeter before, we have seen DIY spectrophotometers (which measure the entire absorption spectrum of a solution) and even building blocks to make a complete biochemistry lab.

12 thoughts on “Classic Colorimeter Clone Calibrates Cuvettes’ Contents

  1. Some sanding and some spray paint could have made it look much nicer, but it’s a great design. I wonder if those color sensors that adafruit has are sensitive enough to do this with digital output..

    1. I had a quick look (I don’t buy there). They have:
      https://www.adafruit.com/product/4698
      Which claims to be a spectrometer but in reality it’s only 10 Channel.

      It depends on what or how accurately you want to detect. Fixed band spectrometery is good for some things where accuracy isn’t important and the widths of the band isn’t important.

      To do better involves complex and or expensive optics, kind of like the image for Pink Floyd Dark Side of the Moon and a linear array sensor like from a document scanner.

      Color and the human perception of color are entirely different. In fact color doesn’t actual exists except for human (and some other animals) perception. Kind of like Schrödinger’s cat that doesn’t exist until it’s observed.

      So RGBW or CMYK schemes give the humans the perception of color without conforming to the complete physics of color.

      The most amusing thing about color is that the less your able to perceive color the more likely you are to understand the physics behind it (in the general population).

  2. The biggest thing with this will be calibration.
    The LEDs have different efficiencies and the lrd has different sensitivities at different wavelengths.
    Lots of known samples and lots of data collection, but still a lot cheaper and more accessible than lab hardware and it’s always nice to learn something by creating your own equipment.

    1. Calibration wouldn’t be stable as the current sink is highly temperature sensitive. Both the diodes and the base-emitter junction will vary the forward voltage with temperature. Fortunately one of the diodes will cancel most of the variance of Vbe. The weakness of the circuit is that most of the voltage is dropped across the transistor and not the reference resistor. This gives the variation of the diode Vf a greater influence of the current. He should replace one of the diodes with a 5.6Volt zener (the most temperature stable), leave the other to roughly compensate for the transistor Vbe and change the reference resistor to suit the desired current. Then this circuit would be far less temperature sensitive.

      That leaves the temperature vs light output of the diodes and I am not sure about that. You could go a long way even having different currents for each of the LEDs but that may be far more than he needs. At least the suggestion above is simple.

      If wanted, I cold do the math for the reference resistor to be used with one 1N4148 and a 5.6 Volt Zener (and teh same transistor). The Minimum Vcc would be higher but the current Vcc is enough as it stands.

      Another thing that might be worth experimenting with is to put the LED in with the reference resistor so that LEDs with a higher Vf were supplied proportionally less current.

      1. “Another thing that might be worth experimenting with is to put the LED in”…” with the reference resistor so that LEDs with a higher Vf were supplied proportionally less current.”

        LED in [series/parallel] with the reference resistor?

    1. This reminds me, I found an Ocean Optics Spectrometer in the recycle bin at work. I don’t know if it works (it is an older model). Someday I hope to test it out.

Leave a Reply

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.