[Nagy Krisztián] had an Intel 286 CPU, only… There was no motherboard to install it in. Perhaps not wanting the processor to be lonely, [Nagy] built a simulated system to bring the chip back to life.
Okay, 68 pins does look like a lot when you arrange them like that.
The concept is simple enough. [Nagy] merely intended to wire the 286 up to a Raspberry Pi Pico that could emulate other parts of a computer that it would normally expect to talk to. This isn’t so hard with an ancient CPU like the 286, which has just 68 pins compared to the 1000+ pins on modern CPUs. All it took was a PLCC-68 socket, an adapter PCB, a breadboard, and some MCP23s17 logic expanders to give the diminutive microcontroller enough I/O. With a bit of work, [Nagy] was able to get the Pi Pico running the 286, allowing it to execute a simple program that retrieves numbers from “memory” and writes them back in turn.

Notably, this setup won’t run the 286 at its full clock speed of 12 MHz, and it’s a long way off from doing anything complex like talking to peripherals or booting an OS. Still, it’s neat to see the old metal live again, even if it’s just rattling through a few simple machine instructions that don’t mean a whole lot. [Nagy] equates this project to The Matrix; you might also think of it as a brain in a jar. The 286 is not in a real computer; it’s just hooked up to a microcontroller stimulating its various pins in a way that is indistinguishable from its own perspective.

We’ve seen similar retro projects before, such as this FPGA rig that helped a NEC V20 get back on its feet. If you’re doing your own tinkering on the platforms of yesteryear, we probably want to know about it on the tips line.

hackaday.com/2026/03/01/making…

We’ll start this week off with a bit of controversy from Linux Land. Anyone who’s ever used the sudo command knows that you don’t see any kind of visual feedback while entering your password. This was intended as a security feature, as it was believed that an on-screen indicator of how many characters had been entered would allow somebody snooping over your shoulder to figure out the length of your password. But in Ubuntu 26.04, that’s no longer the case. The traditional sudo binary has been replaced with a one written in Rust, which Canonical has recently patched to follow the modern convention of showing asterisks on the password prompt.

As you might expect, this prompted an immediate reaction from Linux greybeards. A bug report was filed just a few days ago demanding that the change be reverted, arguing that breaking a decades-old expectation with no warning could be confusing for users. The official response from a Canonical dev was that they see it the other way around, and that the change was made to improve the user experience. It was also pointed out that those who want to revert to the old style of prompt can do so with a config change. The issue was immediately marked as “Won’t Fix”, but the discussion is ongoing.

Speaking of unexpected changes, multiple reports are coming in that the February security update for Samsung Galaxy devices, which is currently rolling out, removes several functions from the Android recovery menu. After the update is applied to phones such as the S25 and Fold 7, long-standing features, such as the ability to wipe the device’s cache partition or install updates via Android Debug Bridge (ADB), disappear.

Just like with the change to sudo, this is the sort of thing that will aggravate veteran users the most. There’s been no official explanation for these changes, and it’s not immediately obvious why Samsung would fiddle with the recovery menu that’s remain largely unchanged since Android’s introduction. As 9to5Google mentions, it could be an attempt to prevent users from installing leaked firmware builds — a practice that’s gotten the attention of the electronic giant’s legal department.

These days, software updates are just one of the things you need to keep track of. Add in emails, RSS feeds, and incoming chat messages, and keeping up with the notifications on your computer or smartphone can be a challenge. But that’s nothing compared to the 800,000 alerts fired off earlier this week by the Vera Rubin Observatory. The observatory uses a 3.2 gigapixel camera to take long exposure images of the night sky, which are then compared with earlier shots to detect visual changes. Astronomers create filters to narrow down what they’re after, and can be notified when the automated system detects a match. A preview image is available in just seconds, while the full-resolution imagery takes around 80 hours to process. It’s still early days, but once the VRO gets up to speed, it’s expected that as many as seven million alerts will be generated each night.

While on the subject of large-scale engineering projects, this week, Google announced that its new data center in Minnesota will be hooked up to the world’s largest battery. The 300 megawatt array built by Form Energy will use iron-air technology, which essentially uses a reversible rusting process to store energy produced by renewable sources such as wind and solar. When those sources aren’t available, the data center can run off of battery power for up to 100 hours.

While heavier and less efficient than lithium-ion, iron-air batteries have the advantage of being substantially cheaper to produce. So while it’s unlikely you’ll see the technology in smartphones anytime soon, it’s perfect for static installations like this.

Finally, some sad news from the world of retro computing/games: a very rare copy of Tsukihime Trial Edition was apparently destroyed while in transit from one collector to another. It might not look like much — the game was distributed by the indie developers on unbranded floppies at a Japanese convention in 1999 — but it represents one of only 50 copies known to exist. While the occasional damaged package is all but unavoidable, this one is particularly egregious as it appears that someone at US Customs intentionally ripped the disk to pieces. The purchaser has filed a complaint with Customs, and we’re interested in hearing what their version of the story sounds like.

See something interesting that you think would be a good fit for our weekly Links column? Drop us a line, we’d love to hear about it.

hackaday.com/2026/03/01/hackad…

Bottom of the cast-iron radiator gaming PC during plumbing. (Credit: Billet Labs, YouTube)
Water-cooled PCs generally have in common that there’s a radiator somewhere in the loop, yet nobody said that you can’t build the PCB into the radiator. Something like a genuine Victorian-era cast-iron radiator, for example. For the folk over at [Billet Labs], this is just your typical project, of course, even if it took a solid three months to make it all work.

Their previous project was also a water-cooled PC, but in the form of a steampunk-esque wall-mounted installation. What differentiates this new build is that it’s trying to be more of a sleeper PC, as long as you ignore some copper tubing and the like running around the outside of this vintage radiator.

Of course, by using a vintage cast-iron radiator like this, you’re also dealing with all the disadvantages of cast-iron, such as the countless impurities in the metal and the immense weight. With water in the loop, the entire build comes in at about 99 kilograms, and cleaning the radiator of particulates released inside it — including rust — was a challenge.

With the amount of water inside the loop, it was little surprise that even a computer stress test only raised the water temperature by two degrees, but the main takeaway from the project was that cast-iron in a water loop is a pain, even with a galvanic corrosion inhibitor. For this reason, the video’s comment section mentioned [gsuberland], for example, mentioning chemical passivation, referencing steel sanitary pipes, and the formation of a nitride layer.

Maybe using a more modern, pure steel radiator would be easier here for that reason, but we can only admire [Billet Labs] for persevering with this project. As a bonus, this is also pretty much a guaranteed theft-proof PC, as even his massive new TV and sound setup weigh less combined, and are probably infinitely more portable. Or, you could choose to go ultra-modern for a futuristic look instead.

youtube.com/embed/xxtZ9DnQJVk?…

hackaday.com/2026/03/01/turnin…

Fish are popular animals to keep as pets, and for good reason. They’re relatively low maintenance, relaxing to watch, and have a high aesthetic appeal. But for all their upsides, they aren’t quite as companionable as a dog or a cat. Although some fish can do limited walking or flying, these aren’t generally kept as pets and would still need considerable help navigating the terrestrial world. To that end, [Everything is Hacked] built a fish tank that allows his fish to move around on their own. We presume he’s heard the old joke about two fish in a tank. One says, “Do you know how to drive this thing?”

The first prototype of this “fish tank” is actually built on a tracked vehicle with differential steering, on which the fish tank would sit. But after building a basic, driveable machine, the realities of fish ownership set in. The fish with the smallest tank needs is a betta fish, but even that sort of fish needs much more space than would easily fit on a robotics platform. So [Everything is Hacked] set up a complete ecosystem for his new pet, making the passenger vehicle a secondary tank.

The new fish’s name is [Carrot], named after the carrots that [Everything is Hacked] used to test the computer vision system that would track the fish’s movements and use them to control the mobile fish tank. There was some configuration needed to ensure that when this feisty fish swam in circles, the tank didn’t spin around uncontrollably, but eventually he was able to get it working in an “arena” where [Carrot] could drive towards some favorite items he might like to interact with. Mostly, though, he drove his tank to investigate the other fish in the area.

The ultimate goal was for [Everything is Hacked] to take his fish on a walk, though, so he set about training [Carrot] to respond to visual cues and swim towards them. In theory, this would have allowed him to be followed by his fish tank, but a test at a local grocery did not go as smoothly as hoped. Still, it’s an interesting project that pushes the boundaries of pet ownership much like other fish-driving projects we’ve seen.

youtube.com/embed/ayWqjUOTNQ8?…

hackaday.com/2026/03/01/fish-d…

For a late-1990s engineer with good soldering skills, many a free pint of beer could be earned by installing modchips on the game consoles of the day. Modchips were usually a small microcontroller connected with a few wires to selected pins on the chips or pads on the board that masked or overrode the copy protection and region locking. This scene was brought back for us by a recent [Modern vintage gamer] video looking at the history of console hardware mods, and it’s worth a watch (see the video, below).

The story starts in 1996 with the original PlayStation, largely the source of those free pints for a nascent Hackaday scribe back in the day. Along the way, as he expands the story, we find other memories, for example, the LPC bus-based hijacks of the first XBox console, and the huge modding scenes on both that machine and Sony’s PS2. The conclusion is that this community left its mark on today’s consoles even though the easy hardware hacks may be a thing of the past on the latest hardware, and as past Hackaday articles can attest, jailbreaking older consoles still has a way to go.

In the early days, our recollection is that the PlayStation modchips were driven by the region locking rather than piracy, for the simple reason that Sony used 80-minute ISOs which wouldn’t fit on the then-available consumer 74-minute CD-R. We also remember them being used by people who couldn’t afford a blue debuugging PlayStation,. or the rare black developer model.

youtube.com/embed/oWSTEJEgFSE?…

hackaday.com/2026/03/01/the-lo…

One of the problems with being a graffiti artist is that you have to carry around a different spray can for each color you intend to use. [Sandesh Manik] decided to solve this problem by building a rig that can produce a wider range of colors by mixing the paint from several cans at once. Check it out in the video below.

The project is called Spectrum. It uses four off-the-shelf spray paint cans—colored red, blue, yellow, and white—and mixes them to create a wider range of colors. All four cans are hooked up to a single output nozzle via a nest of tubing and a four-to-one tube manifold. Key to controlling the flow of paint is a custom device which [Sandesh] calls the “rotary pinch valve,” with one fitted to the feed line coming from each spray can. These valves use a motor-driven lever to pinch a plastic tube shut, allowing them to control the paint flow. This design keeps the mechanism and paint completely separate, which was important to stop paint from fouling the valves in short order. It also prevents backflow, which keeps the paint going towards the outlet and prevents ugly messes. By quickly actuating the valve, the paint flow from each can is modulated to mix various colors as desired.

The mixing valves are under the command of an Arduino Nano. The microcontroller reads a series of knobs to select the amount of each component color to mix, and displays relevant information on a screen. Then, when a pushbutton is pressed, the valves are actuated to spit out the right amount of each paint from the atomizer nozzle. [Sandesh] went so far as to include an advanced “gradient” mode, where a force-sensitive button allows the device to transition smoothly from one color to another depending on how hard the button is pushed.

It’s a neat concept which we’d love to see explored further, perhaps with a more traditional selection of CMYK paints rather than the more unusual red, yellow, blue, and white. We’ve also seen some fun spray paint projects before, like this neat wall-mount plotter. Video after the break.

youtube.com/embed/IicmGjPu4J0?…

hackaday.com/2026/03/01/color-…

Not so long ago, most computer users didn’t own their own machines. Instead, they shared time on mainframes or servers, interacting with this new technology through remote terminals. While the rise of cloud computing and AI might feel like a modern, more dystopian echo of that era, some look back on those early days with genuine fondness. If you agree, check out this 70s-era terminal replica from [David Green].

The inspiration for this build was a Lear Siegler ADM-3A terminal seen at a local computer festival. These machines had no local computing resources and were only connected to their host computer via a serial connection. The new enclosure, modeled on this design, was 3D-printed and then assembled and finished for the classic 70s look. There are a few deviations from a 70s terminal, though: notably, a flat LCD panel and a Raspberry Pi 3, which, despite being a bit limited by today’s standards, still offers orders of magnitude more computing power than the average user in the 70s would have had access to.

On the software side, there are a few modifications to allow the Pi 3 to emulate a CRT-style display. It also runs the i3 windows manager, which was the easiest way to replicate the feel of an old terminal without going command-line-only. With the Pi’s computing power available, though, it’s easier to run emulators for older computer systems, and there’s perhaps no better way to get a sense of how these systems behaved than to use a replica from the era. Another excellent way is to completely reimagine what these computers could have been like in an alternate past.

hackaday.com/2026/03/01/comput…

The timer mechanism of the Vend-o-Vision. (Credit: SpaceTime Junction, YouTube)
There was a time before portable TVs and personal media players when the idea of putting coin-operated TVs everywhere, from restaurants to airports and laundromats, would have seemed like a solid business model. Thus was born the Vend-o-Vision by Mini-TV USA, which presented itself as a cash earner for businesses and a way to make their customers even happier. One of these new-in-box units recently made its way over to [Mark] of the SpaceTime Junction YouTube channel.

This unit is very simple, with what appears to be an off-the-shelf Panasonic black-and-white TV with UHF and VHF reception capability, inside a metal box that contains the timer mechanism, which is linked to the coin mechanism. Depending on a physical slider with three positions, you get anywhere from 10 to 20 minutes per quarter, with the customer having to tune into the station themselves using the TV’s controls. A counter mechanism is provided as an option.

Time to enjoy your favorite TV shows. (Credit: SpaceTime Junction, YouTube)
As would be expected from a new-in-box unit, after chiseling off the 30-odd-year-old Styrofoam packaging, it fires right up and works fine. Of course, it’s a small black-and-white TV, so it’s not incredibly useful, and clearly wasn’t even back in 1989 when the Vend-o-Vision first appeared.

After some finagling with adapters, [Mark] gets everyone’s favorite movie playing on the tiny screen, giving us the first glimpse of what it would have been like to gaze at this miracle of technology back around the early 1990s in a noisy laundromat or restaurant. One can hardly imagine why it didn’t catch on.

We can see a patent for this appear in a 1990 scan of the USPTO’s gazette, where it’s listed as being first in commercial operation on the 29th of November 1989. The system was short-lived, however, with in 1995 the FTC settling with the company for deceptive practices, as the company had overinflated the projected earnings per TV when it started flaunting it at tradeshows in 1990. A few years prior, Mini TV USA appears to have already ceased operations, making these remaining Vend-o-Vision quite rare indeed. These types of coin-operated TVs were usually in public places or hotels. But we’ve seen coin-operated TVs that briefly appeared in homes, too.

youtube.com/embed/bRFbgluNduQ?…

hackaday.com/2026/02/28/vend-o…

At first glance, it may look like [Rybitski]’s 7-segment RGB LED clock is something that’s been done before, but look past the beautiful mounting. It’s not just stylishly framed; the back end is just as attentively executed. It’s got a built-in web UI, MQTT automation, so Home Assistant integration is a snap, and allows remote OTA updates, so software changes don’t require taking the thing down and plugging in a cable.

A slick web interface allows configuring which LEDs belong to which segments without code changes.
Pixel Clock is code for the Wemos D1 Mini microcontroller board and WS2812/WS2812B RGB LED strips, but it’s made to be flexible enough to support different implementations. For example, altering which LEDs in the strip belong to which segments on which digits can be configured entirely from the web interface. Naturally, one could build an LED strip clock using the same layout [Rybitski] did and require no changes at all — but it’s very nice to see that different wiring layouts are supported without needing to edit any code. There’s even automatic brightness adjustment if one adds an LDR (light-dependent resistor), which is a nice touch.

[Rybitski]’s enclosure is CNC-routed MDF, framed and given a marble finish. The number segments are capped with laser-cut frosted white acrylic, which serve as both diffuser for the LEDs and an attractive fit with the marble finish at the front. MDF is dense and opaque enough that no additional baffles or louvers are needed between segments.

With this code and an RGB LED strip, you can implement your own 7-segment clock any way you like, focusing on an artful presentation instead of re-inventing the wheel in software. Of course, there’s nothing that says one must use 7-segment numerals; some say your LED clock need not display numbers at all.

hackaday.com/2026/02/28/this-l…

Picture this: you have an irregular opening you need to fabricate a piece to fill. Maybe it’s the stonework of a fireplace; maybe it’s the curved bulkhead of a ship. How do you get that shape? The most “Hackaday” answer would be to 3D scan the area, create a CAD model based on the point cloud, and route the shape with CNC. Of course, none of those were options for the entirety of human history. So how do you do it if you don’t have such high-tech toys? With a stick, as [Essential Craftsman] takes great pains to show us in the video below.

It’s not just any stick, of course. Call it a “tick stick”, a “speil stick”, or a “joggle stick” — whatever you call it, it’s just an irregularly shaped piece of wood. The irregular shape is key to the whole process. How you use it is simple: get some kind of storyboard — cardboard, MDF, whatever — that fits inside your irregular void. Thanks to the magic of the stick, it need not fit flush to the edges of the hole. You put the tick stick on the storyboard, press the pointy end against a reference point on the side of the hole, and trace the stick. The irregular shape means you’re going to be able to get that reference point back exactly later. Number the outline you just made, and rinse and repeat until you’ve got a single-plane “point cloud” made of tick stick outlines.

Your storyboard is probably going to look mighty confusing, but that’s what the numbers are for. Bring your storyboard and your tick stick onto the workbench and whatever you want to cut out– plywood, cardboard, 1/4″ steel armor plate, you name it–and simply repeat the process. Put the tick stick inside outline #1 and mark where the pointy end lands on the material. Then do it again for the other outlines, reproducing the points you measured on the original piece. After that, it’s just a game of ‘connect the dots’ and cutting with whatever methodology works for your substrate. A sharp knife will work for cardboard, but you’ll probably want something more substantial for steel plate.

It’s not often you’re going to need the tick stick– the [Craftsman] reports only needing it a few times over the course of a decades-long career, but when you need it, there’s not much else that will do the job. Well, unless you have a 3D scanner handy, that is.

youtube.com/embed/p1eiruU6v5I?…

hackaday.com/2026/02/28/have-y…

Everyone is probably familiar with the concept of battery-powered devices, but generally, this involves a laptop with a beefy battery pack and hardware optimized for low power draw. You could also do the complete opposite and try to run a desktop PC off alkaline AA cells, as [ScuffedBits] recently did out of morbid curiosity. Exactly how many alkaline cells does it take to run a desktop PC for any reasonable amount of time?

One nice thing about using batteries with a desktop PC is that you can ditch the entire AC-DC power conversion step and instead use a DC-DC adapter like the well-known PicoATX and its many clones. These just take in 12 VDC and tend to have a fairly wide input voltage range, which is useful when your batteries begin to run out of juice. In this case, just above 10 VDC seemed to be the cut-off point for the used DC-DC adapter.

In the end, [ScuffedBits] used what looks like 56 alkaline AA cells connected in both parallel and series, along with two series-connected 6,800 µF, 40V electrolytic capacitors to buffer the spikes in power demand, after early experiments showed that the cells just cannot provide power that quickly. Although admittedly, the initial thin wiring didn’t help either. With alkaline rather than carbon AA cells, improved wiring, and some buffer capacitors, it turns out that you can indeed run a desktop PC off AA cells, if only just about long enough for a small game of Minesweeper.

Amusingly, the small LCD monitor used in the experiment drew so little power that it happily ran on eight NiMH cells for much longer, highlighting just how important power conservation is for battery-powered devices. We wonder if you could marry this project to a battery project we saw and end up with something practically portable?

youtube.com/embed/U5lskFXDbWs?…

hackaday.com/2026/02/28/runnin…

As temperatures warm up in the Northern Hemisphere, one’s mind naturally turns to the outdoors and the garden — even if some of our gardens are still snow-covered. One secret to good gardening is that many of the plants we love take too long to grow if started from seed outside, at least in relatively temperate climes. There are a myriad of ways to grow seedlings indoors, and this new hack highlighted by [GrowVeg] looks like a great way to get started.

The idea apparently comes from the seedier side of Instagram, where [Farida Sober] has been popularizing it as a “seed snail”, a name they seem to have coined. The technique is very simple: take a sheet of something cheap that won’t disintegrate when moist like bubble wrap or cardboard, layer it with soil — up to 5 cm depending on your seed size — and you roll the whole thing up like a piece of sushi to produce the spiral shape that gives the hack its name. With a piece of tape to hold the roll, it’s just a matter of planting your seeds according to the packet directions. If that’s clear as mud, check out the video embedded below.

Once the seedlings have grown, it looks like it will be very easy to unroll the spiral and pluck them out to plant in the ground or bigger pots without overly traumatizing their roots, like we always do starting in flats. If it weren’t for those delicate roots, it certainly looks like the snail might save some space compared to, say, peat pots. Just remember that starting under the proper LEDs can make a huge difference to how quickly your seeds grow. No dirt? No problem — once sprouted, your plants can be made to grow hydroponically. For the really adventurous, there’s even aeroponics.

youtube.com/embed/UfDU3wDmrdo?…

hackaday.com/2026/02/28/the-on…

Al and I were talking on the podcast today about a sweet 3D printed wide-format camera build, and we got to musing on why we 3D-print.

For Al, it’s an opportunity to experiment with 3D printing itself: tweaking his machines to get the best performance out of them. Other people make small, functional objects that they need in their daily life, like bag clips or spare parts for broken appliances. Some folks go for the ornamental or the aesthetic. The kids in my son’s class all seem obsessed with sci-fi props and fidget toys. The initial RepRap ideal was to replace all commercial fabrication with machines owned by the individual, rather than by companies – it was going to be Marxist revolutionary.

But there’s another group of 3D printer enthusiasts that I think doesn’t get enough coverage, and I’m going to call them the hobbyist industrial designers. These are the people who design a custom dog-poop-bag holder that exactly fits their extra-wide dog leash, not because they couldn’t find one that fit in the pet store, but because it’s simply fun to design and fabricate things. (OK, that’s literally me.)

It’s fun to learn CAD tools, to learn about how things are designed, how they work, and how to manufacture them at least in quantity one. Dreaming, designing, fabricating, failing, and repeating until you get it right is a great joy. And then you get to use the poop-bag holder every day for a few years, until you decide to refine the design and incorporate the lessons learned on the tough streets of practical use.

Of course none of this is exclusive to 3D printing. There were always people who designed-and-built things in the metal machine shop, or made their creations out of wood. In that sense, the 3D printer is just another tool, and the real fun isn’t in using the 3D printer, but rather in the process of bringing things out of your mind and into the world. So maybe there is nothing new here, but the latitude that 3D printing affords the hobby designer is amazing, and that makes it all the more fun, and challenging.

So do you 3D print for necessity, to stick it to the man, to pimp your printer, for the mini-figs, or simply for the joy of the process of making things? It’s all good. 3D printing is a big tent.

This article is part of the Hackaday.com newsletter, delivered every seven days for each of the last 200+ weeks. It also includes our favorite articles from the last seven days that you can see on the web version of the newsletter. Want this type of article to hit your inbox every Friday morning? You should sign up!

hackaday.com/2026/02/28/the-jo…

MRI machines come with a variety of safety warnings. Perhaps most importantly, you have to be very careful not to take ferrous metal objects anywhere near them, since strong magnetic fields can send them flying, causing damage and injuries. To that end, you might find yourself in need of magnetically-safe tools when working on such machines. [Sam Schmitz] recently whipped up a nifty example of an MRI-safe torque wrench himself.
The torque wrench mechanism, which operates in one direction only.
It’s a 3D printed design which can be produced on a Formlabs Fuse 1+ as a single piece in nylon using a selective laser sintering process. The torque wrench works in a deceptively simple manner. As the handle is rotated, a flap mates with the flat side of a fin on the shaft. This allows the shaft to turn. However, apply more than 0.6 Nm of torque, and the fin will eventually give in, snapping over the lip and stopping any further rotation that would over-tighten the fastener. [Sam] suggests these printed torque wrenches largely come out to the correct torque spec when printed, and can survive a thousand cycles or more while remaining in a usable spec.

The wrench does have one drawback though—it is apparently painfully loud to use. When the handle snaps past the detent, the “click” is quite piercing. [Sam] has measured the sound at up to 125 dB. Not exactly the best when it comes to ear safety!

If you work on MRI machines regularly, you already have the tooling to do your job. However, it’s neat to see that such a specialized tool can be easily and reliably 3D printed… with the slight drawback that you need a $60,000 SLS printer to do it. SLS isn’t readily available at the DIY level just yet, but it is slowly getting there. We’re waiting with bated breath.

hackaday.com/2026/02/28/3d-pri…

Sodium Chloride has a melting point of 801 C (1,474 F), putting it comfortably between commonly-cast materials like aluminum and bronze. Which led to [Robinson Foundry] asking the question: can you cast salt like a metal? The answer, surprisingly, was yes!

[Robinson] tries casting the salt with two different methods: like it was glass, and like it was metal. In the glass-like casting, he packs a ceramic mold with salt and tosses it into an electric kiln, there to melt and very slowly cool. In metal-like casting, he just tosses salt into a crucible and melts it in the same beer-can kiln we saw when we featured his lost-pla casting a while back. The molten salt is poured very carefully into sand casting molds. If you’re familiar with the technique, you can skip to about 5:20 when he does the reveal.

As it turns out, the sand casting works out much better. While the glass-style casting in the electric kiln grew much larger crystals and so is more translucent, it’s also stuck completely inside the porous ceramic. Perhaps the ceramic would need glazed to pull off that technique?

On the other hand, the sand reacts with the salt in some way– molten salt isn’t exactly a noble gas, after all–to create a lovely gunmetal finish to the parts. They almost look like metal, though the brittleness gives away the game when he opens the mold to show a dagger in several pieces. For the decorative busts and megalodon teeth in the test, though, it is a great success.

Now, we’re not going to say this video came about because of high metal prices, or comment on what sort of trade policies might be driving up the price of metals like aluminum in the USA, but we do think this a great hack. While salt-based castings are obviously going to have very different physical properties than metal, for decorative work, it creates a lovely finish out of a material that’s cheap as dirt. Hopefully he comes back to the glass-style casting; we would not want to trust that black coating around food, and a salt crystal salt shaker sounds too good to pass up.

The only times we’ve seen molten salt around here is in nuclear reactors, and in homemade batteries, though that first one obviously wasn’t table salt.

youtube.com/embed/HaY-gvgS6JY?…

hackaday.com/2026/02/28/nacl-c…

[Joshua Clay] recently unveiled his newest RC Nerf Dart Robot and talks through his design choices, pointing out that in his aim to have it launch darts fast and hard he may have somewhat overshot the mark. He found out first hand during testing that it shoots hard enough to leave welts through a sweatshirt and probably should be downgraded a bit. Thankfully, one of the features of his new unit is a highly modular design that makes iterating easier than ever.
A modular, glue-free assembly that leaves wiring accessible helps make design iterations faster and easier.
This model is an evolution of his first Nerfbot, and the new one is a smaller, tighter design that trades a wheeled base for a tracked one, among other changes.

The tank platform is one example of [Joshua] using affordable, off-the-shelf solutions where it makes sense to do so. For example, the inexpensive tank-track platform means he can focus on the rest of the bot without having to design or make his own tank treads. Similarly, to control the bot he opts for a PlayStation 4 controller, paired to the bot over Bluetooth. It’s high quality, inexpensive, commonly available, and easily interfaced with the RP2040 that runs the show.

[Joshua] aims for a modular, LEGO-inspired mechanical assembly that makes maintenance, wiring, and iteration as easy as possible. We especially like how the battery, wiring, and things like gears for the pan-and-tilt mechanism of the Nerf launcher are easily accessible.

The dart launcher uses two flywheels to grip and propel each dart fed from a high-capacity magazine, and you can watch it move and shoot around the 9:44 mark in the video, embedded below. It’s plenty loud, but the camera is barely able to register darts leaving the barrel.

If you like the looks of [Joshua]’s newest Nerfbot, keep an eye out because he’s got more to share about it and is considering other features like a camera. In the meantime, there are a few more photos on his website.

youtube.com/embed/k_1R4x8YI2A?…

hackaday.com/2026/02/27/evolve…