2025-11-03 02:00:15
Not every robot has to be big. Sometimes, you can build something fun that’s better sized for exploring your tabletop rather than the wastelands of Mars. To that end, [philosiraptor] built the diminutive PITANK rover.
As you might guess from the name, the rover is based on the Raspberry Pi Zero 2. It uses the GPIO pins to output PWM signals, commanding a pair of servos that drive the tracks on either side of the ‘bot. The drivetrain and chassis are made from 3D-printed components. Controlling the robot is handled via a web interface, which [philosiraptor] coded in C# to be as responsive as possible. So you can see where you’re driving, the ‘bot is also kitted out with a camera to provide a live video feed.
Given its low ground clearance and diminutive size, you’re not going to go on big outdoor adventures with PITANK. However, if you wish to explore a nice flat indoor environment, its simple tracked drivetrain should do nicely. We’ve featured a great many rovers over the years; if you’ve got a particularly special one, don’t hesitate to notify the tipsline!
2025-11-02 23:00:32
If you’re really good, it’s possible to solve a Rubik’s Cube in under 10 seconds. For the rest of us, though, it can be an exceedingly tedious task. For that reason, you might like a Rubik’s Cube that can solve itself, like the one [zeroshot] is trying to build.
What [zeroshot] built is essentially a very small robotic platform inside the center section of an existing Rubik’s Cube. It uses five gear motors that are assembled into the cube’s core, which have enough torque to rotate the individual faces quite easily. While six motors would allow more efficient solves in fewer moves, it was easier to fit just five motors inside the cube, and they’d still get the job done. The motors are controlled by an ESP32, hooked up to a bank of DRV8833 motor drivers. For now, the cube is still a work in progress. While the core can move the faces, [zeroshot] is trying to figure out how to best tackle the problem of feedback in the limited space available. After all, the ESP32 needs to know where the faces are if it’s to make the right moves to reach a solved state. Soldering wires between individual modules can be quite space inefficient; this is one build that might benefit from being integrated onto a single tiny PCB.
We’re used to seeing robots that grab a Rubik’s cube and solve it for you; we haven’t seen a lot of cubes that solve themselves. Regardless, this feat has been achieved before. Video after the break.
2025-11-02 20:00:26
This project gives a whole new meaning to DIY PC. We don’t know how capable you were as a teenager, but could you have designed your own Ryzen-based mini PC?
Whilst making repairs to laptop internals, [Dominik Baroński] was busy taking notes. Modern super-integrated laptop PCs have reached the point where all the functions of a complete PC are embedded in a single chip. But it’s a big, complicated chip with very specific feeding and care needs. Once you’ve figured out what it needs, it ‘merely’ remains to supply it power, hook up some DDR4 RAM, PCIe storage, and some USB ports, and you’re away. It sounds easy when you say it like that, but do not underestimate how difficult it is to create such a board—or even to populate it by hand—yet that’s precisely what [Dominik] has achieved.
The first video is a time-lapse of the soldering process, which isn’t very interesting beyond the fact that they didn’t even waste time making a solder paste stencil and just ran with manual tinning and hot-air reflow. Well, we guess it works, but you wouldn’t want to build a whole batch this way! Anyway, the second video, produced by YouTuber [Coleslav], is originally in Polish, but auto-dubbed to English for the rest of us, and whilst a bit long-winded, does give a flavor of how [Dominik] approached this project. There are quite a few interesting little technical details that [Coleslav] has teased out of [Dominik] when interviewing them for the video, such as they noticed that certain laptop manufacturers were reusing older PMU circuits designed to power DDR2 RAMs by tricking the controller into operating at lower DDR4 voltages by tweaking resistor values, rather than specifying a ‘proper’ (i.e. more expensive) DDR4 compliant device and redesigning the circuit. [Dominik] relied heavily on the Saturn PCB toolkit for calculating differential pairs and other physical PCB aspects to make it possible to design the circuit in KiCAD with just six layers on a minuscule 100 mm x 100 mm outline. Quite a feat!
There were issues with using certain chips that were available to buy, but the documentation was leaked, so the seller was likely not authorized. But the biggest problem is the BIOS, which was duplicated from a similar laptop. [Dominik] hopes to find help to get coreboot running on this board, at which point the archaic keyboard and system controller (now called the EC) can be junked in favor of a more hacker-friendly STM32 setup.
No PCB footprint for the Ryzen chip was available either. [Dominik] created it using a Python script that read the SVG view of the ball-out downloaded from the WikiChip site. The pad positions were known, but the names still needed to be entered manually. All 1140 of them. Once the mappings were entered, schematic symbols could be generated to complete the schematic. Next, they created a 3D model using ChatGPT to write a Python script that read in the ball positions and spat out an STL file that could be molded into a complete footprint! [Dominik] has made a short write-up on Hackaday.io with a few images, and hopefully, more details will appear in the coming months. We’ll be keeping an eye on this young maker over the next few years; we have a feeling great things are coming.
Whilst we’re on the subject of building PCs, here’s a DIY gaming laptop with a twist. At the other end of the complexity scale, here’s a neat DIY computer built from scratch.
Thanks to [JM] for the tip!
2025-11-02 16:00:18
Before a spectrometer can do any useful work, it needs to be calibrated to identify wavelengths correctly. This is usually done by detecting several characteristic peaks or dips in a well-known light source and using these as a reference to identify other wavelengths. The most common reference for hobbyists is the pair of peaks produced by a mercury-vapor fluorescent light, but a more versatile option is a xenon-bulb light source, such as [Markus Bindhammer] made in his latest video.
A xenon gas discharge produces a wide band of wavelengths, which makes it a useful illumination source for absorbance spectroscopy. Even better, Xenon also has several characteristic spikes in the infrared region. For his light source, [Markus] used an H7 xenon bulb meant for a vehicle headlight. The bulb sits in the center of the source, with a concave mirror behind it and a pair of converging lenses in front of it. The converging lenses focus the light onto the end of an optical cable made of PMMA to better transmit UV. A few aluminum brackets hold all the parts in place. The concave mirror is made out of a cut-open section of aluminum pipe. The entire setup is mounted inside an aluminum case, with a fan on one end for cooling. To keep stray light out of the case, a light trap covers the fan’s outlet.
[Markus] hadn’t yet tested the light source with his unique spectrometer, but it looks as though it should work nicely. We’ve seen a wide variety of amateur spectrometers here, but it’s also illuminating to take a look at commercial scientific light sources.
2025-11-02 13:00:07
Multitasking is something we take for granted these days. Just about every computer we use, from our desktops to our phones, is capable of multitasking. It might sound silly to implement multitasking on lower-spec machines from many decades ago, given their limited resources, but it can be done, as [bchiha] demonstrates on a Z80-based machine.
[bchiha] has achieved pre-emptive multitasking on the TEC-1G Z80 computer, a modern reimagining of the classic Talking Electronics TEC-1 from the 1980s. The proof of concept code allows running up to eight separate tasks at once. Task switching runs on interrupts, triggered at approximately 50 Hz. When an interrupt fires, the CPU registers are transferred onto that task’s stack, and the next task’s stack is swapped to the stack pointer to allow execution of the new task to proceed. There is an overhead, of course, with [bchiha] noting that the task swapping routine itself takes about 430 clock cycles to run in between tasks.
Multitasking took some time to appear on home computers for good reason—it’s not very useful unless you have a machine with enough power to practically run multiple tasks at once. While a Z80 machine like this can do multitasking, you’d better hope each task is pretty tiny to avoid each individual task taking forever to run.
[bchiha] has made the simple multitasking code available on Github for the curious. We’ve featured multitasking work on other unconventional platforms before, too, like the Arduino Uno. Video after the break.
[Thanks to Stephen Walters for the tip!]
2025-11-02 10:00:44
If you’ve ever wanted to pump sound to all the rooms of your house, you might use any one of a number of commercial solutions. Or, you could go the more DIY route and whip up something like the Esparagus Audio Brick built by [Andriy].
The concept is simple—it’s a small unit, roughly the size of a brick, which streams high-quality audio. It’s based around an ESP32, which pulls in digital audio over Wi-Fi or Ethernet. The microcontroller is hooked up to a TAS5825M DAC, which comes with a built-in amplifier for convenience. The Esparagus is designed for integration with Home Assistant, allowing for easy control as part of a smart home setup. It’s also compatible with Spotify Connect, AirPlay, and Snapcast—the latter of which provides excellent sync when using multiple units across several rooms.
Design files are available on Github for the curious. We’ve seen other neat projects in this space, before, too—like the charmingly-named OtterCast. Video after the break.
