Geothermal Cooling For New York’s Subways

September 29, 2026 by 21 Comments

You might think that if you dug some tunnels down into the ground where you could be out of the sunshine, that you might find them to be relatively cool and comfortable inside. Yet, fill those tunnels with trains and supporting equipment and millions of commuters, and you’ll instead find yourself sweltering in the heat.

New York’s subways are too hot for comfort, and have been for some time. Unfortunately, unlike any old regular building, you can’t just air condition a subway and call it done. Instead, authorities are looking for other solutions to beat the heat on the famous underground transit network.

Continue reading “Geothermal Cooling For New York’s Subways” →

Testing Coulomb’s Law And Similar Fundamentals Yourself Remains Tricky

September 29, 2026 by 12 Comments

How easy is it to build a simple demonstration at home that proves fundamental law of physics, such as Coulomb’s inverse square law, or Maxwell’s equations? This is one of those questions that always nagged at [Casual Physics Enjoyer] as they sought to find ways to develop a more intuitive understanding of these fundamentals with simple DIY-at-home experiments. Unfortunately this turned out to be rather tricky for Coulomb’s law.

Us H. sapiens – courtesy of a few bright individuals – have figured out many of the fundamentals that underlie this physical world over the past few thousands of years. Even if we still got some pretty massive fundamentals left today for equally bright-minded folk to bash their collective heads against, basics like Coulomb’s law or Maxwell’s equations ought to be a snap now for the average person to replicate since they were after all performed in an era before quantum mechanics, computers or even electrification.

To the dismay of the average physics student, experiments that demonstrate these fundamental laws of physics are still pretty hard to replicate without some solid time and monetary investment. This is demonstrated quite succinctly in the linked article, where the question of how to demonstrate the force between two charged particles, as in Coulomb’s law, and that it is an inverse square relationship is attempted.

Whereas the Wikipedia entry claims it to be a ‘simple experiment’ with just two identical spheres, in a DIY setting this is somewhat tricky, including the part where you have to charge up the spheres and measure the force. Using some aluminium foil and wires you can get the two ends to separate as you’d expect, but quantifying the force is a whole other matter.

The many polygons of Solid Snake on an ESP-driven LCD

Metal Gear Solid Moves From PlayStation To ESP32

September 29, 2026 by 5 Comments

If you haven’t heard of Metal Gear, you probably haven’t played video games these last few decades. The original Metal Gear dates back to 1987 and launched on the MSX, but the real rise to fame probably started with Metal Gear Solid on the original PlayStation in 1998. What needed a hefty console in 1998 can comfortably fit on a microcontroller in 2026, though, as [David Montero Crespo] demonstrates with his port of MGS to the ESP32-S3.

[David] is — as we always are — standing on the shoulders of giants with this hack. Most specifically, the project relies entirely on the [FoxdieTeam] MGS Reversing decompliation project. Of course what one team decompiles, another can recompile, and in this case [David] chose to recompile the game for Expressif’s exceptional ESP32. It wasn’t quite as easy as just forking the repo and compiling with the ESP32 as a target though, as the blog post explains.

We won’t spoil it, but [David] did have to make some changes to account for the different quirks the MIPS processor in the PlayStation has compared to the Xtensa cores on the microcontroller. Then to make it playable, he put the ESP32-S3 module onto some perfboard with an analog stick from a drone controller, an ILI9341 LCD panel, and a resistor ladder to run the buttons for a barebones handheld.

Fans of MSG and its many sequels may appreciate this replica of the codex communicator. Fans of the ESP32 may remember the Wipeout clone we featured earlier, which already proved the popular microcontroller could handle PlayStation-level graphics. If you were hoping for actual solid metal gears, you can cut your own at home.

Pitting A CFD-Optimized Toroidal Propeller Against A Conventional One

September 28, 2026 by 23 Comments

Although we often think that we got certain aspects of aerodynamics pretty much licked at this point, details like the optimal shape of a propeller remains hotly debated, both in- and outside of academia. This also includes wilder designs like toroidal propellers that even after more than a hundred years are still mostly just being evaluated. Recently [Neuronautics] took a shot at figuring out whether toroidal propellers even make sense.

In order to do this, first an efficiency baseline was established using a conventional and highly optimized propeller. After scanning it in to get its exact geometry and running it through a computational fluid dynamics (CFD) simulation, the software spat out a number of about 73%.

This left figuring out an optimized shape for the toroidal propeller to pit against it. While you can absolutely brute-force the seventeen shape parameters being considered here and test them in CFD, this would take insanely long. The hack here is to use multiple reference frame (MRF) to drastically speed up the selection process, though even then it still took two months. An example of using MRF with regular propellers is discussed  in a 2019 paper by [Randi Franzke] et al. in Energies.

Although MRF saves a lot of simulation time, you still end up with a lot of data that has to be analyzed for interesting patterns. For this [Neuronautics] trained a artificial neural network to automate filtering the many options for the most optimal ones, until converging onto a single design.

This G1401 design was the lucky winner, though with only a simulated 62.9% efficiency. Subsequently the one aspect that had been left unchanged was also iterated through, in the form of many different airfoil shapes until the final design appeared.

This design was then 3D printed in resin, which showed the first hurdle with the selection process, in that the printed versions were too thin and flexible to be usable as propellers. Cue many hours of manual tweaking of the design to make it actually printable.

Although the final design didn’t exceed 62% efficiency in a final test, the comment section to the video rightfully points out that the comparison was between a commercially made propeller and a DIY resin-printed one, which adds a whole other batch of variables. That said, it’s unlikely that there’d have been an obvious improvement either way, otherwise we’d already have seen toroidal propellers pop up everywhere on drones and aircraft.

Continue reading “Pitting A CFD-Optimized Toroidal Propeller Against A Conventional One” →

Two-Dimensional Material Now Easier To Manufacture

September 28, 2026 by 2 Comments

We often see scientific breakthroughs in journals or other media that’s reported on as if it’s a revolutionary technology guaranteed to reshape human existence, only to never hear about it again. A more cynical reason for this phenomenon is a certain amount of clickbait or engagement farming, but the real culprit often tends to be the discovery of a process that can’t produce the new material or effect at a scale that makes sense for mass production. One of those are MXenes (“max-enes”), a two-dimensional material first produced over a decade ago, but new research into them has developed a much more efficient way of producing them.

Before this discovery, these materials were produced in a convoluted process involving the MXene precursor materials, combining them with an etchant, washing them off, and then repeatedly spinning them in a centrifuge to separate out unwanted byproducts. As one can imagine, this doesn’t produce material in an industrial quantity. But the new method uses a vapor deposition process which simplifies the precursor steps using less expensive materials, as well as skips the etching step. After that, the researchers have found that the MXenes can grow in a much more controlled way, allowing for greater production amounts and higher quality.

In the intervening years since their discovery and first synthesis MXenes have shown potential for wide-ranging applications, from battery production to antennas to water purification. Another interesting application is as a switchable Faraday cage, as we’ve covered in the past. Hopefully the slow plod of scientific discovery continues and we can start seeing more of these incremental gains improving our lives, even if we don’t get a sudden technological revolution from it.

a beauty shot of the 3D-printed version

This PICO-8 Handheld Is No Fantasy

September 28, 2026 by 7 Comments

The PICO-8 is what is known as a ‘fantasy console’– you can program it, and there are oodles of games, but it’s not actually the 8-bit console it pretends to be. It’s a virtual machine, one that can run on top of any modern OS. When [UncleStem] got into PICO-8, he decided that he liked the PICO-8’s portability, he wanted a different kind of portability– so he took the ‘fantasy’ out of ‘fantasy console’ with his GameBoy-esqe handheld build, the build/dev video of which is embedded below.

At its core, it’s a Raspberry Pi Zero2W. That’s rather overkill for an 8-bit virtual machine, but it’s what he had on hand, and it makes for an easy build. The real hero of the build is the 720 by 720 pixel IPS display that matches the PICO-8’s square aspect ratio and takes up most of the real estate on the handheld. Even better, it has HDMI in and sound support, so with some buttons and a battery the hardware is all there after [UncleStem] designs a handsome 3D printed case, and also has a go milling it out of aluminum before giving up and having a pro do it.

Since it’s so easy to do these days, everything lives on a custom PCB from his sponsor that gives the handheld good mechanical stability. He’s also using the full-colour silkscreen option to make a very pretty PCB. If that part of the project appeals, we’ve covered guides to PCB art before. He’s shared the PCB and everything else he can on a google drive, though note that the fantasy console itself is not open source, and would need purchased to finish the project.

Oddly enough, this is not the first handheld we’ve seen based on the PICO-8. That would be this project that uses a single-button camcorder form-factor. As unique as that is, this project has the benefit of working with the whole library of PICO-8 games.

Continue reading “This PICO-8 Handheld Is No Fantasy” →

A drone is shown, carrying underneath it a white plastic box. On the side of the box are two patch antennas. A camera extends from one end of the box, and a large GPS antenna from the other end.

Synthetic Aperture Radar Drone Gets Interferometric Imaging

September 28, 2026 by 9 Comments

It’s been more than a year since [Henrik Forstén] built the first iteration of his synthetic-aperture radar (SAR) imaging drone, and he’s certainly been productive in the meantime. Not only did he develop a much more powerful autofocus algorithm to clean up the radar images, but he also extended the software to create high-resolution interferometric images.

The main limitation of the original radar system was the GPS, which only had a resolution of about one meter; the autofocus algorithm owed much of its improved clarity to an improved estimation of the drone’s position. A simpler, though more expensive, solution was to add an RTK-capable GPS receiver. RTK (Real-Time Kinematic) receivers use a fixed ground station to constantly transmit a correction signal, letting them reach a couple centimeters of accuracy. Since the drone doesn’t actually need to know its position in real time, it can also use PPK (Post-Processing Kinematic) positioning, which compares recorded GPS signals after the flight to obtain similarly accurate positions.

[Henrik] also implemented a few other hardware improvements, including stabilizing the phase-locked loop used to generate the radar’s frequency sweep. The controller FPGA’s SD card interface had too low a bandwidth to record data in real time, so [Henrik] also implemented a simple, fast compression algorithm to speed that up. Most significantly, he also developed a program for interferometric imaging. The drone flies the same path twice at different altitudes; by comparing phase information from different passes, it’s possible to detect a target’s elevation. Normally, the radar program assumes constant elevation, making tall objects seem to lean toward the radar source; an interferogram, on the other hand, allowed [Henrik] to generate a detailed elevation map.

[Henrik] is no stranger to synthetic aperture radar; we’ve previously covered a bike-mounted iteration and a budget SAR system. If the concepts behind this are still a bit fuzzy, we’ve also covered a guide to making your own SAR setup.