Accelerating Innovation with 3D Printing

The rise of low-cost 3D printers has enabled fabrication and rapid prototyping of all sorts of structures that once required either great skill and lots of time, or a large budget and access to industrial equipment, to produce. While they are extremely versatile, these printers are perhaps best known for their ability to produce plastic replacement parts, cases for electronic devices, and decorative items. Some people have certainly pushed these machines to go further than that, but even still, you would probably not expect a 3D printer to produce a sensor in one pass.

Yet that is exactly what a team of researchers at Aarhus University and MIT has done. And not just any sensor either — they have 3D-printed an accelerometer, which is traditionally a fairly complex instrument. Using their newly-developed approach called AcceloPrint and a multi-material 3D printer, it was shown that a functional accelerometer can be printed in less than an hour, and at a cost of around $1.50.

AcceloPrint works by leveraging capacitive sensing to detect acceleration or angular tilt. The design consists of three key components: a stationary frame, a stationary sensing plate, and a swinging mass. As the object moves, the swinging mass — attached via a cantilever beam — deflects toward or away from the sensing plate. This movement changes the capacitance between the two, and by tracking that variation, the sensor determines the applied acceleration.

The sensor is fabricated entirely with conductive PLA (cPLA) and standard PLA plastics, requiring no post-processing or manual assembly. cPLA is used for both the sensing plate and the swinging mass, while the frame is printed from non-conductive PLA. This integration of mechanical and electrical components means that, once printed, the sensor is ready for immediate use — there are not even any support structures to remove.

Users can choose between a static design, with a fixed beam length determined by a computational model, or an adjustable design, where a mechanical slider allows for tuning the sensitivity range after printing. Longer beams detect finer movements, while shorter ones respond better to larger accelerations.

To support these customizations, the team developed a Blender plugin that automates sensor generation. It uses the Euler-Bernoulli beam theory to calculate the optimal beam length for a given acceleration range. This lets users quickly design sensors suited to tasks ranging from gentle tilt detection to high-impact motion, such as monitoring a golf swing.

An ESP32-based sensing board powers the system. It includes an AC inverter that supplies a 64V, 7kHz sinusoidal signal to the swinging mass. Capacitance changes are read by the ESP32 via an ADC, and are processed using a combination of a peak filter and a Kalman filter to reduce noise. Communication is handled wirelessly via ESP-NOW. To further combat interference, the team uses EM-shielded coaxial cables between the sensor and the board.

In tests, the sensors detected acceleration up to 50 m/s² with less than 4% error in lower ranges, and up to 10.6% error in higher ones. Calibration is required for more accurate readings, especially when the sensor is used on the body or in noisy environments. This performance may not be good enough to replace traditional accelerometers, but for rapid prototyping and other special purposes, AcceloPrint may be the ideal solution.