Disclaimer: modify your espresso machine at your own risk. My electronics knowledge is virtually nonexistent, and my design should definitely be approached with a healthy dose of skepticism.

This repository contains the source code for an Arduino-powered temperature monitoring and control system for espresso machines with an E61 grouphead. The project is heavily inspired by Howard Smith's setup, which uses a small DC fan to cool the grouphead instead of the usual cooling flush. This approach has the advantage of not wasting water, which is good both environmentally and financially.

I started this project to address an overheating issue on my Sanremo Treviso espresso machine and figured others might also benefit from it, which is why I'm open-sourcing the code and circuit design. This is not meant to be an easy and straightforward build guide, but rather a (hopefully) useful starting point for people interested in building something similar.

• Monitors basket and group temperatures using thermistors. The group thermistor is inserted in an adapter screwed into the M6 closure plug (such as Eric's grouphead thermometer). The basket thermistor is wrapped in foil tape and sandwiched between the portafilter basket and the grouphead.
• Controls a DC fan which cools the grouphead to a target temperature. The target temperature is user-selectable with two push buttons.
• Times the shot using a tilt switch taped to the brew lever.
• Displays basket temperature, group temperature, and shot time on an OLED screen.
• Sends time and temperature logging information over serial. A companion Python script listens to the serial channel and converts the information into JSON files that a companion Jupyter Notebook can then read and display.

• python3
  • pytz
  • pyserial
  • seaborn
• arduino-cli
  • Adafruit_ADS1X15
  • Button (the latest version available from GitHub)
  • TaskScheduler
  • U8g2

The group and basket thermistors' Steinhart-Hart model coefficients ({BASKET,GROUP}_SH_{A,B,C}) and the voltage divider circuits' known resistance ({BASKET,GROUP}_KNOWN_RESISTANCE) are hard-coded for my own setup and need to be calibrated to the specific thermistors and resistances used in a new build:

• The known resistances can be measured empirically using a multimeter or set to the resistances' nominal values.
• The Steinhart-Hart coefficients can be estimated using an online calculator. The thermistors' resistances need to be measured for three known temperatures and input into the calculator. I find that one measurement is enough when using the following temperature-resistance pairs:
  • The thermistor's nominal resistance at 25°C.
  • The resistance at 75°C as predicted by the beta parameter model (using the thermistor's nominal beta).
  • An empirical resistance measurement at around 95°C.

The basket thermistor is meant to be used to characterize the espresso machine's thermal properties. The grouphead thermometer measures temperature a few centimeters above the espresso puck, which means that the temperature it reads is typically slightly higher than water temperature at the puck (assuming that the grouphead acts as a heat sink and the shot has been running for some time).

By pulling throwaway shots using stale coffee beans — I personally would not want to drink a shot that has been in contact with the basket thermistor — we can monitor the temperature at both points and establish a relationship between the temperature read by the grouphead thermistor and the temperature at the espresso puck. I do not have experience with enough HX espresso machines to draw a definitive conclusion, but I suspect that the relationship is machine-dependent.

Once we know the relationship, we can target brew temperatures by setting the corresponding grouphead temperature. I find that placing the DC fan a few centimeters above the E61 grouphead mushroom and having it point downwards works well.

1. Connect the Arduino device to the computer with a USB cable.
2. Run python3 espresso-shot.py --fqbn <YOUR_ARDUINO_FQBN>.
3. Press the spacebar to turn the Python script's recording mode on.
4. Pull a shot.
5. Press the spacebar again to turn recording mode off. This allows to perform a cleaning flush without polluting the data/ directory with spurious measurements.

1. Open Data Analysis.ipynb by running jupyter notebook and opening the notebook with the web interface.
2. Run the Imports and Data loading cells.
3. Select the runs to display from the list that appears below the Data loading cell.
4. Run the Data plotting cell.

1. Position the DC fan.
2. Power on the Arduino device.
3. Adjust the target temperature with the push buttons. The OLED screen will switch from showing the grouphead temperature to showing the target temperature for one second when the buttons are pressed.
4. Wait for the DC fan to cool the grouphead down to the target temperature. The fan turns on when the grouphead is above target temperature and turns off otherwise.