2026-01-02 23:00:00
Thanks to the Raspberry Pi, we have easy access to extremely inexpensive machines running Linux that have all kinds of GPIO as well as various networking protocols. And as the platform has improved over the years, we’ve seen more demanding applications on them as well as applications that use an incredibly small amount of power. This project combines all of these improvements and implements a media streaming server on a Raspberry Pi that uses a tiny amount of energy, something that wouldn’t have been possible on the first generations of Pi.
Part of the reason this server uses such low power, coming in just around two watts, is that it’s based on the Pi Zero 2W. It’s running a piece of software called Mini-Pi Media Server which turns the Pi into a DLNA server capable of streaming media over the network, in this case WiFi. Samba is used to share files and Cockpit is onboard for easy web administration. In testing, the server was capable of streaming video to four different wireless devices simultaneously, all while plugged in to a small USB power supply.
For anyone who wants to try this out, the files for it as well as instructions are also available on a GitHub page. We could think of a number of ways that this would be useful over a more traditional streaming setup, specifically in situations where power demand must remain low such as on a long car trip or while off grid. We also don’t imagine the Pi will be doing much transcoding or streaming of 4K videos with its power and processing limitations, but it would be unreasonable to expect it to do so. For that you’d need something more powerful.
Thanks to [Richard] for the tip!
2026-01-02 20:00:30
There’s folk wisdom in just about every culture that teaches about renewable energy — things like “make hay while the sun shines”. But as an industrial culture, we want to make hay 24/7 and not be at the whims of some capricious weather god! Alas, renewable energy puts a crimp in that. Once again, energy supplies are slowly becoming tied to the sun and the wind.
Since “Make compute while the wind blows” doesn’t have a great ring to it, clearly our civilization needs to come up with some grid-scale storage. Over in Sardinia they’re testing an idea that sounds like hot air, but isn’t — because the working gas is CO2.
The principle is simple: when power is available, carbon dioxide is compressed, cooled, and liquefied into pressure vessels as happens at millions of industrial facilities worldwide every day. When power is required, the compressed CO2 can be run through a turbine to generate sweet, sweet electricity. Since venting tonnes of CO2 into the atmosphere is kind of the thing we’re trying to avoid with this whole rigmarole, the greenhouse gas slash working fluid is stored in a giant bag. It sits, waiting for the next charge cycle, like the world’s heaviest and saddest dirigible. In the test project in Sardinia — backed by Google, amongst others — the gas bag holds 2000 tonnes and can produce 20 megawatts of power for up-to 10 hours.
That’s not exactly astounding. It gets you through the night, but leaves you hanging if the next day is cloudy. But it’s scalable. The turbine is 20 megawatts, sure, but all you need is land to add extra energy capacity. The 200 MWh pilot plant is a five hectare facility, which is only about 12.3 acres, or roughly 1/10th the size of the Mall of America. It seems like increasing capacity would be fairly trivial; unlike, say, pumped hydro storage, no special topography is required. Ten hours of storage is also notably longer than the six to eight hours grid-scale battery farms usually aim for.
As of this writing, there’s only one of these plants in operation, but expect that to change rapidly. In 2026 the company behind the Sardinia project, Energy Dome, plans on putting in grid-scale storage based on its technology in India and Wisconsin, and that’s before Google gets into it. They’re hoping to roll this technology out at a number of data centers worldwide, though the exact details of the deal aren’t public.
We’ve talked about grid-scale energy storage before, using everything from liquid tin to electric car batteries and big piles of gravel. This methodology has a lot to recommend it over those others in comparison, and should worst come to worst, at least it won’t burn for days like certain batteries we could name. Releasing 2000 tonnes of CO2 might not be as benign as a failure from a liquid air battery, but storing liquid CO2 under pressure is a lot easier holding onto cryogenic air.
All images credited to Luigi Avantaggiato.
2026-01-02 17:00:39
Hardware hackers tend to have loads of hookup wire, and that led [firstgizmo] to design a 3D printable wire and cable spool storage system. As a bonus, it’s Gridfinity-compatible!
There are a lot of little design touches we love. For example, we like the little notch into which the wire ends are held, which provides a way to secure the loose ends without any moving parts. Also, while at first glance these holders look like something that goes together with a few screws, they actually require no additional hardware and can be assembled entirely with printed parts. But should one wish to do so, [firstgizmo] has an alternate design that goes together with some M3 bolts instead.
Want to adjust something? The STEP files are included, which we always love to see because it makes modifications to the models so much more accessible. One thing that hasn’t changed over the years is that making engineering-type adjustments to STL files is awful, at best.
If there is one gotcha, it is that one must remove wire from their old spools and re-wind onto the new to use this system. However, [firstgizmo] tries to make that as easy as possible by providing two tools to make re-spooling easier: one for hand-cranking, and one for using a hand drill to do the work for you.
It’s a very thoughtful design, and as mentioned, can also be used with the Gridfinity system, which seems to open organizational floodgates in most people’s minds. Most of us are pinched for storage space, and small improvements in space-saving really, really add up.
2026-01-02 14:00:43
Just because LEGO Technic is technically a toy doesn’t mean that you cannot do solid engineering with it, like building air-powered engines. After first building a simple air-powered piston engine, this time around [Jamie’s Brick Jams] sought to not only optimize the engine, but also build a clutch and something to power with said engine.
The piston head is one of the handful of 3D printed parts, with the new design featuring twin rubber o-rings as a seal instead of a single big one as in the old design. This incidentally matches the multiple seal rings on an internal combustion engine’s pistons, probably for similar blow-by related reasons. The air hose diameter was also increased from 2 to 3 mm to give the engine a larger volume of air to work with, which along with a new flywheel gave a lot more torque. Next the piston rod length was optimized.
The final radial 4-piston engine turns out to work pretty well, with the clutch engaging smoothly. This was used to drive a DIY generator that turned out to produce about 3 Watt of usable power in its final configuration at 6 V, though it’s admittedly a rather crude generator that could be further optimized. When trying a twin-piston configuration with the highest air pressure before air hoses began to pop off, it hit a dizzying 14,600 RPM.
These aren’t half bad results for some LEGO Technic together with some 3D printed bits, rubber o-rings and some lube.
2026-01-02 11:00:50
Programmers hold to a wide spectrum of positions on software complexity, from the rare command-line purists to the much more common web app developers, and the two extremes rarely meet. One point of contact, though, might be [Jan Antos]’s Brow6el, which uses sixel graphics to display a fully graphical web browser within a terminal.
Behind the scenes, the Chromium Embedded Framework renders webpages headless, then Brow6el uses libsixel to convert the rendered output image to sixels, a simple kind of console-based graphics representation, which it then outputs to the terminal. It regularly re-renders the page to catch page updates and display them in real time, and it can send mouse or keyboard input back to the webpage. For more advanced work, it also has a JavaScript development console, and it’s possibly to manually inject scripts into rendered webpages, or inject them automatically using URL match patterns.
Some other convenient features include a bookmark system, a download manager, terminal-based popup dialog support, support for multiple simultaneous open windows, and a private mode, all of these features being controllable through the keyboard alone. The mouse input can be taken from a real mouse or from a keyboard-controlled virtual mouse, which lets the user click and scroll through websites even on fully text-based systems. [Jan] provides an impressive video demonstration (and we’re not just saying that because of the demo website), which is embedded below.
Brow6el takes inspiration from a few other terminal-based web browsers, such as Carbonyl, though it improves on their graphics. Experienced readers, however, might already know that with some Wayland tricks, it’s possible to turn any application into a terminal app.
2026-01-02 08:00:37
These days ready-to-use DC-DC converters are everywhere, with some of the cheaper ones even being safe to use without an immediate risk to life and limb(s). This piques one’s curiosity when browsing various online shopping platforms that are quite literally flooded with e.g. QS-4884CCCV-1800W clones of a DC-DC boost converter. Do they really manage 1800 Watt even without active cooling? Are they perhaps a good deal? These were some of the questions that [Josh] over at the [Signal Drift] channel set out to answer.
The only real ‘datasheet’ for this module seems to come courtesy of a Floridian company who also calls it the 36843-PS, but it features specifications that are repeated across store listings so it might as well by the official ‘datasheet’. This module is marketed as being designed for the charging of lead-acid and similar batteries, including the boosting of PV solar panel outputs, though you’d really want to use an MPPT charger for that.
With this use case in mind, it’s probably no surprise to see on the oscilloscope shots under load that it has a tragic 100 kHz switching frequency and a peak-to-peak noise on the output of somewhere between 1-7 VDC depending on the load. Clearly this output voltage was not meant for delicate electronics.
Looking closer at the board, we can see that it features a TI TL494C as the PWM controller IC, which drives the MOSFETs that form the boost circuit. There’s also an XLSemi XL7005A buck converter that is used for the low-voltage supply on the board. Meanwhile an LM358 dual opamp seems to be used in the voltage monitoring circuit, which also completes the analysis minus the passives, the MOSFETs for the buck (IRFB3206) and boost (IRFP4468) circuits, and a 100V-rated Power Schottky rectifier.
While the board does implement some basic voltage- and current-related safeties and limits, even the documentation tells you to not leave it powered on for too long. As for pushing it to the full 1,800 Watt output, this would require at least 48 VDC input, enabling e.g. 90 VDC output at 20A. Since the input terminal is only rated for 300V at 30A, the input for the subsequent stress test was limited to 48V at 30A for a total of 1,440 Watt from three 48V PSUs.
Using two resistive heating elements as a ~1,800 Watt load the output of the module was measured to see how far the module can be pushed. This turned out to be 1,200 Watt with the 48VDC input proving to be the limit. With the maximum 60VDC input you may be able to provide the current required to hit the full 1,800 Watt, but at that point you’re pretty close to the output voltage anyway. This makes a total of 500-1,000 Watt more reasonable.
Considering the overall performance, the original listed application as a battery charger seems to be about right, with a very barebones design. Its output switching noise and lack of safeties, as well as inability to fully turn off, mean that it should not be used by itself for anything that will be powered for extended periods of time, nor should anything sensitive to switching noise be exposed to its output voltage. For the $18 or so that this module goes for on certain popular platforms one could do much worse if you know what you’re doing.
