[TheHyperFix] had a problem. He’d spied a brilliant camera slider, but didn’t want to lay out big money to acquire it. The natural solution? Build one! Only, life is seldom so straightforward.

The plan was straightforward – take an old broken 3D printer, and repurpose its parts to make a camera slider instead. The build started with a aluminium extrusion, some V-slot wheels, and a 3D printed platform to hold the camera. Moving the platform was done via a belt drive, using the stepper motors and some software to tell the original printer controller what to do.

Unfortunately, the early experiments failed when the controller blew up under load. An Arduino was subbed in with a CNC shield, which got things back on track, and [TheHyperFix] had a somewhat functional slider with relatively jerky movement. A tough iterative design process ensued to work out problems with bearings and the Arduino’s pulse limit, among others.

As it stands, the slider is semi-functional, but it’s not quite well behaved enough to use for professional shooting. Still, for a first attempt at electronics prototyping, we think [TheHyperFix] did a pretty solid job. It might not be all there yet, but it’s well on the way, and a great deal was learned in the process.

If you’re trying to build a camera slider in a hurry, you might like to try recreating one of the builds we’ve featured before. Video after the break.

youtube.com/embed/AiuP5UIWeO4?…

hackaday.com/2026/05/04/buildi…

Once you peel back the hype and mysticism, large language models (LLMs) are a fascinating application of statistical models, effectively what you get when you dial a basic auto-complete model up to eleven. In order to analyze a mind-boggling amount of text and produce meaningful auto-completion results quite a bit of math is involved, with a recent three-part article series by [Giles] going through the basics of inference, being the prediction step using a trained model.

The text is encoded in the LLM’s vector space as token IDs, each token being a text fragment that has some probability of following another ID, such as when cats may be found on desks, as in the above photo by [Giles]. With inference multiple of such IDs are retrieved in a vector from which in successive steps a sentence can be pieced together. These so-called logits are detailed in the first article in the series, with the second article focusing on vocabulary space and embedding, as well as the matrix operations used for inference.

Finally, the third article puts all of this together and looks at transformers, which is a crucial part of GPT (generative pretrained transformer) LLM architecture. Of note is the attention mechanism, which takes GPTs beyond merely being glorified auto-complete systems by adding pattern matching. Here we can see how the statistical model of the LLM is used to generate a rather plausible output, which is where the human has to ask themselves in how far they feel that it is correct.

Of course, there goes a lot more into making LLMs and GPTs performant, such as key-value caches that massively speed up inference.

hackaday.com/2026/05/04/the-ma…

The remains of a gimbal camera after its drone was shot down. (Credit: Le labo de Michel, YouTube)
The Iranian Shahed-136’s basic design has seen many changes and additions since Russia began using them, with some featuring interesting payloads such as cameras in a gimbal, making these drones useful for tasks like surveillance. Recently [Michel] got his hands one one such camera that was recovered from a shot-down drone in Ukraine, providing the opportunity for an in-depth look at what hardware is in these cameras.

The teardown thus covers the gimbal mechanism itself as well as the electronics and camera. First up is an Artix-7 FPGA-based board, followed by the range finder assembly. Unsurprisingly the camera feed handling is performed by an Hi3519 SoC, as this appears to be the off-the-shelf option you find all over on AliExpress and similar sites. There’s also an Artix-7 FPGA-based board here, which presumably performs some machine vision tasks or similar.

Continuing the ‘bought off AliExpress’ vibe, the power supply board (pictured above) is quite literally just that. A relay board follows the same pattern, with apparently the entire contents of the camera consisting of off-the-shelf development boards and modules that are readily found for sale online.

For the camera there is a thermal camera presumably for night operations, as most of these drone swarms are launched towards Ukraine at night. Looking at the gimbal assembly it similarly feels like it was sourced off AliExpress, featuring mostly Western components, sometimes with the typical lasered-off component markings and such.

This makes one wonder how much has changed here since nearly two years ago we saw an air data computer from a similar drone that could have been sourced off AliExpress, while the Russian missile teardowns show significantly more custom hardware, presumably because those are harder to source off AliExpress.

youtube.com/embed/6CXfxhH8n00?…

hackaday.com/2026/05/04/teardo…

When life gives you lemons, you make lemonade. What if life gives you a pile of old e-book readers? Well, when [spiritplumber] got box of old Nook Simple Touch devices, he decided to design solar-powered cases to help boost the old batteries. It makes perfect sense to us: sunlight readable screen, sunlight chargeable battery.

It looks like he’s got a pair of panels built into the 3D printed case. He recommends using any TP4056-based charger, and tying into the battery test points, not the 5 V supply. It won’t hurt anything if you do, apparently, but the device will think it’s plugged in an refuse to turn off the WiFi. That’s no big deal when you’ve got a continental power grid on the other end of the cable, but charging from a small panel on the back of the case doesn’t always give you enough juice to waste on unneeded radio activity. Especially indoors — these panels are apparently big enough to trickle-charge the device under artificial light, which is a nice, if doubtless slow feature.

The design is open source, and includes SketchUp design files as well as the exported .STL, so if you’ve got a hankering to edit this to fit a different e-book reader, you can. He also provides a handy-dandy guide to root this model of Nook, and if you’re on Hackaday we probably don’t need to explain why you might want to.

We’ve seen the Nook Simple Touch go some interesting places — like into the clouds as a glider computer — but solar power is a new hack for this device, at least on this site. We don’t know if [spiritplumber] has a green thumb, but he’s evidently got some environmental bones in his body: his last featured project was about improving quadcopter efficiency with a wing and a prayer.

hackaday.com/2026/05/04/sunlig…

The ionosphere is of great importance to shortwave radio transmissions, since it allows radio waves to be refracted and reflected over the horizon, and it’s therefore unfortunate that the height and thickness of the ionosphere depends on the time of day or night, weather, season, and the solar cycle. To get a better idea of current transmission conditions, [mircemk] built this shortwave propagation monitor.

The monitor provides a basic measure of ionosphere conditions by measuring the strength of received shortwave signals: if the conditions for transmission are good, it should receive a relatively high level of existing signals, and a weak signal if conditions are bad. It has an external antenna connected to a signal strength indicator circuit based on the CA3089, which amplifies signals in the 1-40 MHz range and outputs a smoothed voltage indicating the RF energy in this range. The output signal can be read by any voltmeter, in this case an Arduino Nano with an OLED display. Assuming the same antenna is always used, the signal should noticeably fluctuate between night and day as the solar wind affects the ionosphere.

Of course, the distance at which you’ll be receiving a signal means nothing unless you have a receiver, which can range from the antique to the modern.

youtube.com/embed/ly-2Ns3jC0M?…

hackaday.com/2026/05/04/a-shor…

Making stuff cool and keeping it that way has been a pretty essential part of human civilization for thousands of years, with only in the past few hundred years man-made methods having become available that remove the reliance on the whims of nature and lugging around massive blocks of ice. The most important cooling method is undoubtedly that of vapor-compression refrigeration, but this is hardly the only method to transfer thermal energy from one location to another.

For example, we recently covered an elastocaloric cooling project by a group of scientists that uses strips of NiTi metal. By flexing these they induce a cooling effect which when put in a number of stages serves to transfer a significant amount of thermal energy between both sides, much like a vapor-compression system but without the gases and compressor. Meanwhile the Seebeck effect is relatively well-known from Peltier thermocouple devices, and features heavily in portable refrigerators and kin where these solid-state devices can also transfer thermal energy.

Of course, along with how they function the major question with all of these cooling technologies is how efficient they are, as this determines when you’d want to even consider them for a specific application.

The Science Of Cold

Although as animals we have an intuitive understanding of what concepts like ‘cold’ and ‘hot’ are in the sense of comfort levels, on a fundamental level the related concept of temperature is about the kinetic energy of the particles in a system. Essentially, the more kinetic energy exists in the system, the higher the temperature of said system is, regardless of whether it’s a liquid, gas, solid or plasma. Hence a temperature of zero Kelvin is the complete absence of any such kinetic energy in the system, also known as the Third Law of thermodynamics:

As the temperature of a system approaches absolute zero, all processes cease and the entropy of the system approaches a minimum value.

When we talk about moving thermal energy from one location to another – as in refrigeration – this thus means transferring said energy from one system to another in some fashion, something which is covered by the First Law of thermodynamics:

In a process without transfer of matter, the change in internal energy, , of a thermodynamic system is equal to the energy gained as heat, , less the thermodynamic work, , done by the system on its surroundings.

In the case of a hot water bottle or ice bag we are actively changing the energy balance of a system by transferring matter. This makes such transfers rather lossy, which is not a quality that is generally desirable in a refrigeration system. Thus we prefer a closed system in which the matter is ideally never lost, and thus all the energy transfer occurs via reversible processes.

Vapor-Compression

Single-stage vapor-compression refrigeration system components. (Credit: mbeychok, Wikimedia)
In vapor-compression refrigeration a liquid – the refrigerant – is circulated through the system, alternately changing state into a gas by absorbing thermal energy from the environment, before shedding this energy again while condensing back into a liquid.

A key component in this system is the compressor, which takes in the saturated vapor. This means that said vapor contains enough energy to effect the liquid-gaseous transition, but is still pretty close to the condensing point.

By compressing this vapor into a smaller volume its temperature increases since roughly the same amount of kinetic energy exists within the system. This superheated vapor then passes through the condenser, like the radiator found at the back of the average kitchen refrigerator. Here the superheated vapor condenses back into a liquid, with the higher temperature and pressure helping to make the condensing process more efficient. This is also why said refrigerator radiator can feel so warm to the touch.

The role of the expansion valve is effectively the opposite of the compressor: as the name suggests this is where the liquid refrigerant at high pressure suddenly transitions back to a low pressure, causing adiabatic flash evaporation of part of the liquid into a vapor. This reduces the temperature of the refrigerant, making it colder than e.g. the inside of the refrigerator and drawing in kinetic energy from the air inside said refrigerator before the vapor makes its way to the compressor again.

Elastocaloric Cooling

The elastocaloric effect. (Credit: Fatemeh Kordizadeh, Wikimedia)
With elastocaloric cooling (ECC) there is no liquid refrigerant or a pressure differential. Instead they rely on the elastocaloric effect, which is thermomechanical in nature.

Similar to how the refrigerant with vapor-compression refrigeration can absorb energy as it transitions from liquid to vapor and vice versa, with the elastocaloric effect it is the material itself that absorbs thermal energy from its environment when it’s mechanically loaded.

The aforementioned NiTi alloy is also known as a shape-memory alloy (SMA), which are generally known to be heat sensitive, finding use in applications like thermal fuses and sensors.

While the application of heat or cold can cause the deformation, this also works the other way around when mechanical force is applied. This is readily demonstrated with a strip of NiTi SMA and a thermal camera, as in this video by Helge Wurst:

hackaday.com/wp-content/upload…

As the strip is bent, the area experiencing the deformation becomes rather warm to the touch, with subsequent relaxation causing the same area to become cold to the touch.

Using such strips and mechanical actuators capable of applying 900 MPa of pressure, Guoan Zhou et al. were able to achieve freezing temperatures. They did this by combining multiple of such elastocaloric stages with CaCl₂ as heat-exchange fluid. This is not a mainstream cooling method so far, but it should be quite reliable and low-maintenance.

Magnetocaloric Cooling

Comparison between magnetocaloric effect and vapor-compression cooling. (Source: Wikimedia)
The magnetocaloric effect (MCE) was first observed in 1881 by German physicist Emil Warburg, with the early 20th century seeing significant progress towards using it for cooling applications. This particular effect as the name suggests consists of exposing a material to a magnetic field, with this material then drawing in thermal energy. Upon removal of the magnetic field the material sheds this gained energy as well as some additional energy, thus cooling down relative to its environment.

Similar to the elastocaloric effect, this relies on an adiabatic process: without the transfer of any matter or entropy. This makes it a fully reversible process that can be repeated by successive applications of said magnetic field.

The biggest disadvantage with this effect for cooling purposes is that it’s only a very strong effect (giant MCE, or GMCE) in a limited number of alloys discovered so far. The first significant here was a rare-earth gadolinium-based alloy, Gd₅(Si₂Ge₂), that showed GMCE at 270 K. This relatively low temperature and the use of rare-earths made this a tough sell.

More recently discovered alloys like Ni₂Mn-X, where X is a variety of additives, display the GMCE near room temperature and even saw GE demonstrate an Ni-Mn-based magnetic refrigerator in 2014. So far commercialization of GMCE-based refrigeration is still rather limited but there is a push to make it work for generally less efficient vapor-compression-based home refrigerators.

Electrocaloric Cooling

Although easy to confuse with the magnetocaloric effect, the electrocaloric effect (ECE) pertains to the application of an electric field in dielectric materials. The effect is roughly the same, with the dipoles in the material either assuming an ordered or disordered state, depending on whether the field is respectively applied or turned off.

So far ECE-based cooling hasn’t seen commercialization yet either, though the past years there have been a range of breakthroughs, with for example Xin Chen et al. demonstrating ECE polymer films in 2023 that was subsequently used to create a thin-film refrigerator prototype with. This was claimed to achieve a Coefficient of Performance (COP) of a rather astounding 24, which compared to traditional heat pumps would make it a rather interesting solution if it can be commercialized.

Thermoelectric Cooling

Diagram of a thermoelectric cooler. (Credit: Ken Brezier, Wikimedia)
The thermoelectric effect and the associated Peltier cooling devices are probably the most well-known and most heavily commercialized on this list along with vapor-compression. Within the thermoelectric effect, the Peltier effect concerns thermocouples and their associated temperature differences, thus lending its name to what are alternatively called ‘Peltier coolers’ as well as ‘thermoelectric coolers’, or TECs.

Rather than a refrigerant or rearranging of dipoles here the transfer of kinetic energy is performed using charge carriers within the TEC. On average charge carriers move to the ‘cool’ side, allowing them to transfer heat away from the other side.

As is well-known, this Peltier effect is rather limited when used as a heat pump, with very low efficiency and strict limitations on temperature differences. This is why their use in dehumidifiers and portable refrigerators is at best questionable.

The main reason why TECs are so popular can be said to be due to vapor-compression refrigeration being so bulky and neither elastocaloric, nor MCE, nor ECE solid-state coolers being quite ready for prime-time yet at the low-low price level that TECs can achieve due to being dead-simple semiconductor devices.

Pulse Tube Cooling

Stirling-type pulse tube refrigerator. (Credit: Mbeljaars, Wikimedia)
Another interesting, partially solid-state cooling method is the pulse tube refrigerator (PTR), which has seen limited use in commercial and other applications. Its main advantage is that it can be used as a cryocooler, making it ideal for space telescopes where sensors have to remain super-cold.

At its core it’s reminiscent of vapor-compression refrigerating, in that it uses a gas and a compressor, yet there’s no circulating loop of refrigerant. Inside the tube a piston alternately compresses the gas – often helium – which forces it through the regenerator. As the compression raises the temperature of the gas, this heat is then passed onto the material of the regenerator. On its way back through the regenerator this heat is then returned to the gas, explaining the name of this component.

The hot and cold sides of the regenerator are hereby used for cooling, though other PTR configurations are possible, such as the coaxial design. The relatively straightforward mechanical design and low temperatures achievable are why hobbyists are tinkering with PTRs in order to do things like making their own liquid nitrogen.

Chill Choices

Ultimately the question of what the right cooling method is for your particular task depends on a range of factors, including the required efficiency, available space and whether or not that big research grant budget just became available.

In terms of commercially available options that aren’t outrageously expensive, your options are somewhat limited, especially if you do not have a lot of space available. It’s possible that in a number of years these alternate technologies will be commercialized and wipe the floor with TECs in particular, but unless you’re currently heavily into tinkering with strips of NiTi SMA to build your own cooler, the primary options would seem to be either vapor-compression or TECs.

That said, considering that only a hundred years ago we were only just beginning to transition from iceboxes to vapor-compression refrigeration, it’s already pretty neat that we have some rather chill options to use today, and absolutely cool ones to look forward to.

Featured image: “Frosted Flakes“, National Park Service photo by [Neal Herbert]. Thumbnail image: “Frost” by [XoMEoX].

hackaday.com/2026/05/04/strang…

Solarpunk is sometimes thought of as the “good ending” to cyberpunk– there’s technology, but it’s community-focused instead of in the hands of evil conglomerates, and– if the name doesn’t give it away– renewably powered. [Victor Frost] found that image of the future inspiring enough to create this ESP32-hosted community hub. Yes, it looks like a lantern, but it’s actually a very-local webserver.
It looks like a lantern, but it’s got a server inside. Plus two 18650 cells to charge from a solar panel that’s presumably off-camera.
Local webserver sounds like an oxymoron, but this device does serve a page over HTTP… just, not on the world-wide web. Instead the only way to access it is via its own Wireless Network– he’s using the ‘captive portal’ that forces you to log into public wifi to direct people to a community message board.

It’s unmoderated, and unfiltered– users can post what they like, but given that they have to be within a few meters of the device, it’s not exactly anonymous. It’s a lot like the community center corkboard brought into the 21st centruy, which is very in keeping with the solarpunk ethos.

For ease of updates, he’s subdivided the ESP32’s flash into three partitions: one for the data, and two for the software, using LittleFS. This allows live updates and keeping a known-good backup for the quickest possible turnaround and/or rollback. One interesting thing is that his who UI– the actual web site, HTML, CSS, and JS– is all crammed into a single string in PROGMEM rather than files on the little file system. It’s an interesting choice, and makes for quick updates, firmware and UI in one go. Not everyone will like it, but it works for [Victor]. The code is, of course, on GitHub under the GPL— there’s a lot of overlap between the open source and solarpunk ethos, after all.

It’s a bit of a pity that he missed our Green Powered Challenge, as this project would have fit right in to the PV category, considering it runs on a 6W panel. For all the cyberpunk and solar power you see on this website, you’d think the “solarpunk” tag would be more popular, but no– all we have is this stained-glass robot.

Thanks to [Victor] for the tip! If you missed our contest, too, no worries– we take projects of all colours, green or otherwise, all the time. Just drop us a tip.

hackaday.com/2026/05/04/esp32-…

Leaded fuel is considered one of the greatest environmental failures in modern human history. Adding tetraethyl lead to gasoline reduced knock in internal combustion engines, which was widely considered a good thing. It was only later that the deleterious health effects came into view, by which point there was a massive fleet of lead-dependent automobiles and an industry reluctant to change. Still, the tide turned, and over the last 50 years, unleaded fuel has become the norm for automotive use across the world.

And yet, there remains a hold out—a world where engines still burn leaded fuels and spray their noxious fumes across the countryside. In the aviation sector, leaded fuel remains a normal part of everyday operations to this day amidst concerted efforts to eliminate it for good.

“Low” Lead

Leaded gasoline is a thing of the past in the automotive world, but remains a standard fuel for piston-engined aircraft to this day. Credit: Ahunt, public domain
Piston-engined aircraft do not typically run on the same fuels as automobiles. Instead, they burn aviation gasoline, or Avgas, which comes in specific grades and is designed to suit the needs of aircraft engines, by being less volatile and more suitable for high-performance applications.

The most common grade is 100LL (low lead), which is used widely across North America and Western Europe. Despite the moniker, the fuel contains 0.56 grams/litre of tetraethyl lead (TEL), somewhat higher than many leaded automotive fuels used in the 20th century. As with ground-based applications, the additive is used to provide a measure of valvetrain protection by offering cooling and preventing microwelds between contacting parts. It also provides an easy increase to the fuel’s effective octane rating. The latter is particularly useful in aviation contexts where engines run under high load conditions for extended periods of time, and where performance is critical.

Other grades of aviation fuel are also in regular use in various parts of the world, many of which still contain significant levels of TEL as well. It’s worth noting that turbine-based aviation engines are not relevant to this issue, as they burn kerosene-based fuels which are lead-free.
100LL fuel is dyed blue for easy identification on the flight line. It’s one of the most widely used fuels in piston-engined aircraft. Credit: Ahunt, public domain
The basic makeup of aviation gasoline was largely decided by the mid-1940s, a period in which fuels were heavily developed to suit the needs of then-cutting-edge piston military aircraft. At the time, knock resistance was key to enabling supercharged aircraft engines to achieve higher power levels, a point of key military interest during World War II. Tetraethyl lead was an easy way to achieve this, and this requirement also led to development of technologies like water-methanol injection.

Unfortunately, burning leaded fuel effectively sprayed significant amounts of lead into the environment. This lead to elevated blood lead levels in the population, causing premature deaths, neurological damage, and negatively impacting development in children. This is perhaps somewhat galling given that the inventor of TEL, Thomas Midgley Jr., himself suffered significant health effects from the compound. Many workers would also die during early efforts to produce industrial amounts of TEL in the 1920s. It’s one of many examples from the 20th century of industrial will prevailing in spite of obvious severe health risks from a dangerous but otherwise useful chemical.

Despite early knowledge of the dangers, it took some time for the negative impacts of TEL to become readily apparent on a wide scale. Japan lead the charge with a leaded fuel ban for automotive use in 1986, with other developed countries following suit in years to come. It would take decades for the last domino to fall, with Algeria finally outlawing the fuel in 2021.
As per the MSDS, 100LL fuel is not good for humans or the environment. Credit: Shell MSDS
However, the aviation world has not been so quick to abandon lead. Much of the reasoning behind this comes down to practicality. Aviation piston engines simply require high octane fuel and TEL has proven one of the easiest ways to achieve a high rating. 100LL, for example, has a MON rating of 100, which is quite high compared to even premium gasoline used in automotive applications.

Engines designed to run on TEL often rely on the additive to prevent excessive valve wear, too, so running these engines on non-leaded fuels can significantly increase wear. This would be an expensive inconvenience in an automotive application, but when the engine is what’s keeping you in the sky, it’s less desirable to risk a failure by running a cleaner fuel.

In 2019, the FAA estimated that there were 167,000 aircraft in the United States that relied on 100LL avgas, and 230,000 worldwide. The agency had asked in 2014 for industry proposals to make a transition towards unleaded fuels for internal combustion applications.

However, testing revealed issues with proposed alternatives, and was eventually halted in 2018. The FAA has since provided a draft plan in 2026 that lays out the timeline to phase out leaded aviation fuel for good. The intent is to end the use of 100LL fuel in the United States by 2030, excepting Alaska, which will phase out the fuel in 2032. The intention is to take an incremental approach, giving the industry time to develop and certify unleaded replacement fuels—with G100UL, 100R, and UL100E all candidates for FAA approval.

Real-world use of these fuels will then be monitored for compatibility and safety and to determine if further support or changes are required to manage the transition away from 100LL. For now, the timelines are still subject to change, particularly in Alaska, where piston-engined aircraft are particularly vital for transport and logistics are harder to manage. However, it marks a very real commitment to ending the use of leaded aviation for good, at least in the United States. If the FAA does manage to pull off this feat, it should be readily achievable for other countries around the world.

Ultimately, leaded aviation fuels aren’t causing the same level of damage to humanity and the environment as leaded automotive fuels, purely by virtue of their more limited use. Still, it’s never ideal to be spraying lead into the environment, and the health risks are always going to be elevated for those near general aviation airports or under flightpaths of piston-engined aircraft. It’s positive that there is a real commitment to end the use of these fuels, but much work remains to be done to end the reign of tetraethyl lead for good.

Featured Image: “Tetraethyl Lead” by [David Brodbeck]

hackaday.com/2026/05/04/why-le…