Author: Aaron Beckendorf

83 Articles

A 3D-printed assembly standing on short legs is visible. A portion extends upward with the word "Nord" sunk into it. Cables extend from one side of the upright portion, and a side view of a circuit board is visible at the front of the assembly.

Measuring Earth’s Rotation With Two Gyroscopes

November 23, 2025 by 35 Comments

We’ve probably all had a few conversations with people who hold eccentric scientific ideas, and most of the time they yield nothing more than frustration and perhaps a headache. In [Bertrand Selva]’s case, however, a conversation with a flat-earth believer yielded a device that uses a pair of gyroscopes to detect earth’s rotation, demonstrating that rotation exists without the bulkiness of a Foucalt pendulum.

[Bertrand] built his apparatus around a pair of BMI160 MEMS gyroscopes, which have a least significant bit for angular velocity corresponding to 0.0038 degrees per second, while the earth rotates at 0.00416 degrees per second. To extract such a small signal from all the noise in the measurements, the device makes measurements with the sensors in four different positions to detect and eliminate the bias of the sensors and the influence of the gravitational field. Before running a test, [Bertrand] oriented the sensors toward true north, then had a stepper motor cycle the sensors through the four positions, while a Raspberry Pi Pico records 128 measurements at each position. It might run the cycle as many as 200 times, with error tending to decrease as the number of cycles increases.

A Kalman filter processes the raw data and extracts the signal, which came within two percent of the true rotational velocity. [Bertrand] found that the accuracy was strongly dependent on how well the system was aligned to true north. Indeed, the alignment effect was so strong that he could use it as a compass.

In the end, the system didn’t convince [Bertrand]’s neighbor, but it’s an impressive demonstration nonetheless. This system is a bit simpler, but it’s also possible to measure the earth’s rotation using a PlayStation. For higher precision, check out how the standards organizations manage these measurements.

A circular metal vessel is shown, with a symmetrical rotor of four vanes standing inside. At the bottom of the vessel are four loudspeakers.

Building An Acoustic Radiometer

November 22, 2025 by 9 Comments

A Crookes radiometer, despite what many explanations claim, does not work because of radiation pressure. When light strikes the vanes inside the near-vacuum chamber, it heats the vanes, which then impart some extra energy to gas molecules bouncing off of them, causing the vanes to be pushed in the opposite direction. On the other hand, however, it is possible to build a radiometer that spins because of radiation pressure differences, but it’s easier to use acoustic radiation than light.

[Ben Krasnow] built two sets of vanes out of laser-cut aluminium with sound-absorbing foam attached to one side, and mounted the vanes around a jewel bearing taken from an analog voltmeter. He positioned the rotor above four speakers in an acoustically well-sealed chamber, then played 130-decibel white noise on the speakers. The aluminium side of the vanes, which reflected more sound, experienced more pressure than the foam side, causing them to spin. [Ben] tested both sets of vanes, which had the foam mounted on opposite sides, and they spun in opposite directions, which suggests that the pressure difference really was causing them to spin, and not some acoustic streaming effect.

The process of creating such loud sounds burned out a number of speakers, so to prevent this, [Ben] monitored the temperature of a speaker coil at varying amounts of power. He realized that the resistance of the coil increased as it heated up, so by measuring its resistance, he could calculate the coil’s temperature and keep it from getting too hot. [Ben] also tested the radiometer’s performance when the chamber contained other gasses, including hydrogen, helium, carbon dioxide, and sulfur hexafluoride, but none worked as well as air did. It’s a bit counterintuitive that none of these widely-varying gasses worked better than air did, but it makes sense when one considers that speakers are designed to efficiently transfer energy to air.

It’s far from an efficient way to convert electrical power into motion, but we’ve also seen several engines powered by acoustic resonance. If you’d like to hear more about the original Crookes radiometers, [Ben]’s also explained those before.

A microscope objective is sitting on a spool of solder in a metal tin, in front of a circuit board which has wires running away from it.

Watching Radioactive Decay With A Homemade Spinthariscope

November 17, 2025 by 2 Comments

Among the many science toys that have fallen out of fashion since we started getting nervous around things like mercury, chlorinated hydrocarbons, and radiation is the spinthariscope, which let people watch the flashes of light on a phosphor screen as a radioactive material decayed behind it. In fact, they hardly expose their viewers to any radiation, which makes [stoppi]’s homemade spinthariscope much safer than it might first seem.

[Stoppi] built the spinthariscope out of the eyepiece of a telescope, a silver-doped zinc sulfide phosphor screen, and the americium-241 capsule from a smoke detector. A bit of epoxy holds the phosphor screen in the lens’s focal plane, and the americium capsule is mounted on a light filter and screwed onto the eyepiece. Since americium is mainly an alpha emitter, almost all of the radiation is contained within the device.

After sitting in a dark room for a few minutes to let one’s eyes adjust, it’s possible to see small flashes of light as alpha particles hit the phosphor screen. The flashes were too faint for a smartphone camera to pick up, so [stoppi] mounted it in a light-tight metal box with a photomultiplier and viewed the signal on an oscilloscope, which revealed many small pulses.

Continue reading “Watching Radioactive Decay With A Homemade Spinthariscope” →

A circuit board in the shape of a business card is shown. The circuitry is confined to the left side of the board, and the rest is used for text.

(Neural) Networking With A Business Card

November 15, 2025 by 13 Comments

A PCB business card is a great way for electrical engineers to impress employers with their design skills, but the software they run can be just as impressive as the card itself. As a programmer with an interest in embedded machine learning, [Dave McKinnon] wanted a card that showcased his skills, so he designed one that runs voice recognition.

[Dave] specifically wanted to run a neural network on his card, but needed to make it small enough to run on a microcontroller. Voice recognition looked like a good fit for this, since audio can be represented with relatively little data, a microphone is cheap and easy to add to a circuit board, and there was already an example of someone running such a voice recognition network on an Arduino. To fit the neural network into 46 kB, it only distinguishes the words “one” through “nine,” and displays its guess on an LED seven-segment display. [Dave] first prototyped the system with an Arduino, then designed the circuit board around an RP2040.

Continue reading “(Neural) Networking With A Business Card” →

A man's hands are shown holding a microphone capsule with a 3D-printed part on top of it, with a flared metal tube protruding from the plastic.

2025 Component Abuse Challenge: Playing Audio On A Microphone

November 10, 2025 by 5 Comments

Using a speaker as a microphone is a trick old enough to have become common knowledge, but how often do you see the hack reversed? As part of a larger project to measure the acoustic power of a subwoofer, [DeepSOIC] needed to characterize the phase shift of a microphone, and to do that, he needed a test speaker. A normal speaker’s resonance was throwing off measurements, but an electret microphone worked perfectly.

For a test apparatus, [DeepSOIC] had sealed the face of the microphone under test against the membrane of a speaker, and then measured the microphone’s phase shift as the speaker played a range of frequencies. The speaker membrane he started with had several resonance spikes at higher frequencies, however, which made it impossible to take accurate measurements. To shift the resonance to higher frequencies beyond the test range, the membrane needed to be more rigid, and the driver needed to apply force evenly across the membrane, not just in the center. [DeepSOIC] realized that an electret microphone does basically this, but in reverse: it has a thin membrane which can be uniformly attracted and repelled from the electret. After taking a large capsule electret microphone, adding more vent holes behind the diaphragm, and removing the metal mesh from the front, it could play recognizable music.

Replacing the speaker with another microphone gave good test results, with much better frequency stability than the electromagnetic speaker could provide, and let the final project work out (the video below goes over the full project with English subtitles, and the calibration is from minutes 17 to 34). The smooth frequency response of electret microphones also makes them good for high-quality recording, and at least once, we’ve seen someone build his own electrets. Continue reading “2025 Component Abuse Challenge: Playing Audio On A Microphone” →

A central circular element is releasing steel ball bearings into a complex nest of eight intertwined plastic paths.

Mesmerizing Marble Runs From Procedural Generation

November 9, 2025 by 10 Comments

There are few things that can keep a certain kind of mechanically-inclined mind entranced as well as a marble run, and few structures that look as interestingly organic as procedurally-generated designs – combine the two and you get [Will Morrison]’s Marble Fountain.

[Will]’s first approach to generating a marble run was to have a script randomly place some points, generate a path following a spline through those points, and give that path a constant slope. This worked, but the paths it generated were a bit too simple to take full advantage of a 3D printer’s capabilities, so he next wrote a path solver to generate more complicated runs. The solver starts by generating a series of random line segments connecting the top and bottom of the run, then iteratively moves the segments into position. Each segment has to stay within the print volume, be evenly spaced with the others, maintain a constant slope, avoid segments from other tracks, and avoid distant segments of its own track. The result is a complicated network of tracks that keeps the marbles in motion without letting them fly out in fast sections. Continue reading “Mesmerizing Marble Runs From Procedural Generation” →

A drone is shown hovering in the sky, with two bright lights shining from its underside.

2025 Component Abuse Challenge: Overdriven LEDs Outshine The Sun

November 6, 2025 by 13 Comments

Tagging wildlife is never straightforward in the best of times, but it becomes a great deal more complicated when you’re trying to track flying insects. Instead of trying to use a sensor package, [DeepSOIC] attached tiny, light retroreflectors to bees and hornets, then used a pulsed infrared light mounted on a drone to illuminate them. Two infrared cameras on the drone track the bright dot that indicates the insect, letting the drone follow it. To get a spot bright enough to track in full sunlight, though, [DeepSOIC] had to drive some infrared LEDs well above their rated tolerances.

The LEDs manage to survive because they only fire in 15-µs pulses at 100 Hz, in synchrony with the frame rate of the cameras, rather like some welding cameras. The driver circuit is very simple, just a MOSFET switch driven by an external pulse source, a capacitor to steady the supply voltage, and a current-limiting resistor doing so little limiting that it could probably be removed. LEDs can indeed survive high-current pulses, so this might not really seem like component abuse, but the 5-6 amps used here are well beyond the rated pulse current of 3 amps for the original SFH4715AS LEDs. After proving the concept, [DeepSOIC] switched to 940 nm LEDs, which provide more contrast because the atmosphere absorbs more sunlight around this wavelength. These new LEDs were rated for 5A, so they weren’t being driven so far out of spec, but in tests they did survive current up to 10A.

We’ve seen a similar principle used to drive laser diodes in very high-power pulses a few times before. For an opposite approach to putting every last bit of current through an LED, check out this low-power safety light.

Continue reading “2025 Component Abuse Challenge: Overdriven LEDs Outshine The Sun” →