Introduction: Degree of Roast Infrared Analyzer for Coffee Roasters

About: Howdy! I am a Biosystems and Agricultural Engineering Master's Student at Oklahoma State University, graduating December 2018. My research project primarily involved using sensors to detect the degree of roast…

Introduction

Coffee is a beverage consumed around the world for both its sensory and functional properties. Coffee’s taste, aroma, caffeine and antioxidant content are just a few qualities that have made the coffee industry so successful. While the green beans origin, quality, and species all effect the quality of the end product, the roasting of coffee is the most influential factor.

Typically, during roasting, the roast master (a highly trained individual) uses properties of the beans such as temperature, texture, smell, sound, and color to evaluate and adjust the roast accordingly. After roasting, the coffee beans are assessed to ensure bean quality. The Agtron Process Analyzer is an industry standard instrument used to measure thedegree of roastof coffee beans using near-infrared abridged spectrophotometry. The degree of roast is essentially a measurement of the coffee’s quality based on the extent of heat transferred during the roast and categorizes coffee into light, medium and dark roasts.

There has recently been a growth of small roasting companies offering custom in-house roasts. These companies are looking for less expensive alternatives to hiring and training a roast master or using the expensive Agtron Process Analyzer. The Degree of Roast Infrared Analyzer for Coffee Roasters, as described in this document, is meant to be an inexpensive means of measuring the degree of roast of coffee beans. The Degree of Roast Infrared Analyzer uses a tryer, a tool found on coffee roasters used to sample the coffee during roasting, to hold a sample of coffee. The tryer is inserted into the analyzer where the AS7263 NIR Spectral sensor is used to measure 6 different infrared bands (610, 680, 730, 760, 810, and 860nm). The reflectance measurements are transmitted via Bluetooth and can then be correlated to the degree of roast. The analyzer must first be calibrated by pressing a button on the inside of the box in which the PVC is used as a white balance as it has a relatively flat reflectance in the spectral range detected by the sensor.

Step 1: Materials

List of Materials

  1. SparkFun Qwiic Shield (https://www.sparkfun.com/products/14352)
  2. SparkFun Qwiic Connector (https://www.sparkfun.com/products/14427)
  3. SparkFun AS7263 NIR Spectral Sensor (https://www.sparkfun.com/products/14351)
  4. 4 x VCC 6150 Lamps 5V .06A (Incandescent Bulbs) (https://www.mouser.com/)
  5. 2 x Momentary Push buttons
  6. 2 x 10kOhm Resistors
  7. DC Barrel Jack Female (https://www.sparkfun.com/products/10288)
  8. HC-05 Bluetooth Module (https://www.amazon.com/)
  9. Power Switch
  10. Solid State Relay (AD-SSR6M12-DC-200D)(https://www.automationdirect.com/)
  11. 1/2" PVC Cap
  12. 1/2" x 1/2" x 3/4" PVC Tee
  13. Craft Box (Hobby Lobby)
  14. Arduino Uno
  15. Tryer
  16. 5V 2A power supply (https://www.adafruit.com/product/276)
  17. USB Cable - Standard A-B (Programming Cable)

Notes on Materials

VCC 6150 Lamps - These are incandescent bulbs chosen because of their high infrared output. The incandescent bulbs are used instead of the LED light provided on the AS7263 module because the onboard LED does not emit the infrared output needed to reflect off of the coffee beans and to be subsequently measured by the sensor. Additionally, it is important to note that in this design the incandescent bulbs are powered from the 5V 2A power source and controlled by the Arduino via a relay. SparkFun does provide two onboard soldering pins on the AS7263 module for the purpose of powering and controlling an auxiliary light source, however these pins are not used because they do not provide enough voltage or amperage to sufficiently power the chosen incandescent bulbs.

SparkFun Qwiic Shield - This shield is used because of its ability to easily connect to the AS7263 sensor via a Qwicc connector. The shield also provides both 3.3V logic level shifting and a large prototyping area.

Solid State Relay - This kind of relay was chosen because of its fast and quiet switching capabilities, however, it is expensive and unnecessary as a standard electrical relay would work as well. If using a standard electrical relay, the code may need to be modified to slow the sampling and calibration process.

PVC size - The PVC size was chosen because of the diameter of the tryer on hand and should be changed if using a different size tryer.

HC-05 Bluetooth Module - An instructables (https://www.instructables.com/id/How-to-Set-AT-Command-Mode-for-HC-05-Bluetooth-Mod/) was used to change the baud rate of the module from 9600 to 115200 to match the baud rate of the AS7263.

Step 2: Wiring Diagram

S1 -- Power Switch

SSR1 -- Solid State Relay

B1 -- Sampling Button

B2 -- Calibration Button

R1 -- 10kOhm Resistor

R2 -- 10kOhm Resistor

L1, L2, L3, L4 -- Incandescent Bulbs

Step 3: Mounting Incadesent Bulbs to the AS7263

A 3D printed mounting ring (STL provided) was made to hold the lamps around the sensor. The lamps were wired in parallel and hot glue was used to keep the leads of the lamps from touching each other. Liquid rubber insulation can be used instead of hot glue. Next, tiny wires were used to secure the mounting ring to the sensor by tying the wires through the holes provided on the sensor.

Step 4: Assemble the Tryer Port

A hole was drilled into the back of the PVC cap to accommodate the momentary push button. The 3/4" side of the PVC tee was cut off and zip ties were used to secure the sensor to the tryer port. The length of the tee may need to be adjusted to accommodate the size of the tryer. A notch was put into the port side of the PVC tee to align the bean sample in the tryer with the sensor.

Step 5: Wiring the Solid State Relay and Power Switch

The lights from the were wired in series with solid state relay and the DC barrel jack.

The Vin on the Qwiic shield was connected to the DC barrel jack through a power switch.

The Ground on the Qwiic shield was connected to the ground of the DC barrel jack.

Step 6: Wiring the Calibration Button

The calibration button was connected to power, Digital 2, and ground using a resistor.

Step 7: Wiring the Sampling Button

The sampling button was connected to power, Digital 3, and ground using a resistor.

Step 8: Wiring the INPUT to the Solid State Relay

The input side of the solid state relay was wired to Digital 5 and ground.

Step 9: Wiring the Bluetooth Module

The Bluetooth Module was wired in accordance to the wiring diagram provided.

VCC ---> 5V

RXD ---> Digital 11

TXD ---> Digital 10

GND ---> GND

Step 10: Code

Upload the code provided to the Arduino Uno using the programming cable.

As a reference, SparkFun provides a startup guide for the AS726x (https://learn.sparkfun.com/tutorials/as726x-nirvi)

CAUTION!! When testing the code make sure the Arduino is not receiving power from both the 5V power supply AND the programming cable. This will fry the Arduino!

Step 11: Displaying Results Through Bluetooth

To display the Bluetooth results, download Bluetooth Electronics by keuwlsoft from the Google Play Store. Save the DegreeOfRoastInfraRedAnalyzer.kwl file to the keulsoft folder in the Bluetooth device's internal storage. Use the save icon in the app to load the kwl file. Next, connect to the HC-05 Bluetooth Module and run the loaded file.

Step 12: Conclusions

Wavelength Legend:

  • R - 610nm
  • S - 680nm
  • T - 730nm
  • U - 760nm
  • V - 810nm
  • W - 860nm

The AS7263 NIR sensor was used to measure the spectral reflectance of coffee beans at 6 different wavelengths for unroasted coffee as well as light, medium, and dark roasts. The results from the sensor show that infrared reflectance decreases with higher degrees of roast across all tested wavelengths. The wavelength with the greatest variation in accordance to degree of roast was found to be 860nm. This system provides a quick and easy to use basis for the offline measurement of the degree of roast of coffee beans. Data from this sensor will provide coffee roasters an additional method of quality control by ensuring repeatable roasts and reducing human error. Further work needs to be done to correlate the infrared data to industry standards.

Step 13: A Special Thanks To....

  • Dr. Timothy Bowser -- Advisor
  • Dr. Ning Wang -- Committee Member
  • Dr. Paul Weckler -- Committee Member
  • Dan Jolliff -- US Roaster Corp.
  • Connor Cox -- Oklahoma Center for the Advancement of Science and Technology
  • The Department of Biosystems and Agricultural Engineering at Oklahoma State University, Stillwater, OK
  • The Food and Agricultural Products Center at Oklahoma State University, Stillwater, OK