Categories
Uncategorized

Agtron Gourmet Conversion

Please find our first conversion from Roast Vision to Agtron Gourmet here!

Protocol:

Sample roasting was carried out on 11/27/2020 using an Aillio Bullet R1 V1. Each batch was 100g, for a total of 9 batches, indicated with red hashmarks on the following image indicating the drop times for each of the batches. The temperatures are NOT accurate, as 100g is too small of a batch size for this roaster’s probe position. While the temperatures are not meaningful, the profile of each roast was reproduced with minimal variation, and each batch was simply stopped at a different development time to simulate different roast levels. Images of each roast follow the profile below.

The roasts are generally even, with some variation due to the challenge of roasting an extremely small sample in a 1kg capacity roaster evenly without scorching or causing other roast defects.

After roasting, the samples were cooled to room temperature, and then placed in jars, loosely sealed for 6 hours. The beans were then measured in a Lighttells CM100+, a Roast Vision, and a third experimental sensor. The grind size for the Roast Vision and experimental sensor were set for Turkish (extremely fine), and the grind size for the CM100+ was medium, to medium fine. Future tests will include the use of coffee sieves to accurately communicate grind size for testing purposes.

Results:

After measuring each of the sample roasts, as well as many other coffees that were currently on hand, the following data was plotted, a linear best fit line was applied, and an equation was extracted to create our first Agtron Gourmet conversion. We will continue to revise the conversion as we gather more data, ensuring grind size is accurate, and accounting for any other variation.

Categories
Uncategorized

Roast Vision Public Dataset

By popular request we’ve released a Google Form and accompanying Spreadsheet where users can enter their Roast Vision measurements of their favorite coffees, and view the public results!

Google Form

Results Spreadsheet

Categories
Uncategorized

Coffee Tech Talk Tuesday

Featured in Episode 12 of Coffee Tech Talk Tuesday! Watch the video here!

Categories
Uncategorized

Hackaday Post

Now featured at Hackaday! Check out the post here.

Categories
Uncategorized

SparkFun Electronics Blog Post

Now featured in SparkFun’s Blog! Check out the post here.

Categories
Uncategorized

Daily Coffee News Feature

Now featured in Daily Coffee News, by Roast Magazine! Check out the entry here.

Categories
Uncategorized

New in Coffee

Now featured in New in Coffee, by the Specialty Coffee Association! Check out the entry here.

Categories
Uncategorized

Spark! at Boston University Profile

Click here to view the full story!

Categories
Uncategorized

Featured at Boston University’s Build Lab

Check out the full post at the following link – here!

Categories
Uncategorized

Origin Story

This project initially began in the Boston University Spark! Innovation Fellowship Program, with early development done on a Flair Classic with a Pressure Kit.

The Problem – Pressure Profiling

Great tasting espresso is hard to make, especially for a home user without access to expensive equipment.  Among the many variables that go into making espresso, the applied pressure to the coffee is an important one for ensuring consistency and reproducibility between each shot. And while there are some current solutions for tracking pressure, they are cost prohibitive for the majority of users.

The Idea – Computer Vision Tool

To use a smartphone selfie camera to take a video of the pressure gauge while making espresso, and then utilize a computer vision algorithm to extract the pressure values from the video.

Project Status

Our team has developed a computer vision algorithm that converts a video of an analog pressure gauge to digital values by using the selfie camera on a smartphone without the need for additional equipment.

We also began the development of a companion journaling app so that users can track each shot and more easily reproduce their favorite shots. This app is intended to help users decrease waste, reduce frustration, and have a more enjoyable experience when making espresso at home.

Future Features

Future features include the ability to provide recommendations to users based on the qualitative and quantitative factors that they are tracking for each shot of espresso, as well as provide a platform and educational system that would help guide new espresso makers through the detailed process of making high quality espresso. This would be achieved by incorporating existing content from the espresso community, in addition to providing a set of guiding principles for new espresso makers to follow and experiment with.

Additional features include the ability for users to share their recipes with others, as well as allowing cafes and roasters to provide recommended recipes and profiles for their own beans.