Hackaday

电动风力玩具飞机使用超级电容器实现自由飞行乐趣

2026-04-18T08:00:02+00:00

There’s something to be said for a simple wind-up, free flight model airplane. With no controls, it must be built very well to fly well, and with only the limited power of a rubber band, it needs a good, high-lift design without much superfluous drag to maximize flight time. There’s also something to be said for modernity though, and prolific hacker [Tom Stanton] puts them together with this supercapacitor plane.

If that sounds familiar, it’s because [Tom] did this before back in 2023. But for that first attempt he converted a commercial R/C toy rather than a plane optimized for low-power free flight. Just like with the best rubber-band machines, his goal for the new production is more flight time than winding time. Plus lots of views on YouTube, but that goes without saying.

Thus this machine is smaller and lighter than the previous iteration. Rather than balsa and tissue like the free-flight aircraft of our youths, [Tom] is using 3D printed plastic for the structure. But he’s got a neat hack built in: he’s printing the wings and control surfaces directly onto tissue paper, eliminating the bonding step. Of course that means his wings are printed flat, but a bit of heat and some bending and he has a single-surface airfoil. Single-surface airfoils are normal in this application, anyway: closed wings add too much weight for too little gain. If you want to try the technique, he’s got files on Printables.

Another interesting factoid [Tom] discovered is that the energy density of supercapacitors decreases sharply below 10 F. As you might imagine by the square-cubed law, bigger is better, but the sharp drop-off dictated he use a single 10 F cap for this build, along with a micro motor. Using the wind-up generator from his previous build, he’s able to get 45 seconds of flight out of just 4 seconds of cranking, a good ratio indeed.

[Tom] seems to like playing with different ways to power his toys; aside from supercapacitors, we’ve also seen him finessing aircraft air motors — including an attempt at a turbine for a model helicopter.

用新电容修复GameCube光驱故障

2026-04-18T05:00:43+00:00

Generally when a game console with an optical drive stops reading discs the first thing that people do is crank on the potentiometer that controls the power to the laser diode to ramp up its output. While this can be a necessary solution to eke out a bit more life out of a clearly dying laser diode, this can actually massively shorten the lifespan of a good diode that’s just held back by bad capacitors. This is demonstrated by [Skawo] with a fix on a GameCube that stopped reading discs.

While it’s absolutely true that laser diodes have a limited lifespan, so do the capacitors and other components in the system. Thus, after tearing down this Japanese GameCube, [Skawo] accesses the optical PCB for some delicate plier-based capacitor surgery. One can absolutely question such violence, as well as the replacement mix of MLCC ceramics and a stray THT electrolytic capacitor, but the results after reassembly are obvious.

Without having to adjust the laser diode’s potentiometer, the game console now happily reads the game disc while the laser diode breathes a sigh of relief. Although all GameCube consoles will face the inevitable demise of their optical drives – barring a replacement optical pickup solution appearing – with this capacitor replacement solution it’s at least possible to stave off that undesirable time for a bit longer.

FRED来到爱好者操作系统（以及Linux）

2026-04-18T02:00:41+00:00

Those who have worked on a hobby operating system for x86 will have interacted with its rather complex and confusing interrupt model. [Evalyn] shows us why and how to use Flexible Return and Event Delivery (FRED), a new standard by the x86 Ecosystem Advisory Group.

Of course, it would be silly to omit the fact that Linux received patches first. But that isn’t the interesting part; after all, Linux is often the first place to have support for this kind of thing. No, what’s interesting is [Evalyn]’s implementation, to our knowledge among — if not the first — non-Linux operating system to support it.

The kernel confirming that FRED has been detected and enabled

To know why we should switch to FRED, we must first understand what it replaces. The Interrupt Descriptor Table (IDT) tells the CPU what code to run when certain interrupts or faults happen. The big problem that the IDT has is inconsistency, most egregiously the fact that the stack layout depends on which interrupt happened. To solve the issues with the IDT, FRED was created.

[Evalyn] shows us the process, starting at the documentation, then finding an emulator capable of it and culminating in a demo where DOOM runs in EvalynOS with FRED enabled.

Pentium II die shot. Martijn Boer, Public domain.

从零开始制作青铜镜

2026-04-17T23:00:00+00:00

Although modern-day silvered glass mirrors have pretty much destroyed the market for bronze mirrors, these highly polished pieces of metal once were the pinnacle of mirror technology. Due to the laborious process required these mirrors saw use essentially only by the affluent. That said, how hard would it be to make a bronze mirror today with all of the modern technologies that even a hobbyist can acquire for their shed? Cue [Lundgren Bronze Studios] giving it a shot, starting by casting something flat-ish to start polishing.

Just getting that initial shape to start polishing is a chore, with hammering out the shape possibly being also a viable method. When casting metal it’s tricky to avoid having air bubbles and other defects forming, though using a sand mold seems to help a lot.

After you have the rough shape, polishing using power tools seems like cheating, but as you can see in the video even going from 50 to 8000 grit with a rotating disc left countless scratches. Amusingly, hand sanding did a much better job of removing the worst scratches, following which a polishing compound helped to bring out that literal mirror finish.

A quick glance at the Wikipedia entry for bronze mirrors shows that a tin-bronze alloy like speculum metal was used for thousands of years as it was much easier to polish to a good mirror finish. The metallurgy of what may seem like just a vanity item clearly goes deeper than just polishing up a metal surface.

等离子体烹饪（非火焰）

2026-04-17T20:00:11+00:00

Cooking food with fire is arguably the technology that propelled humans to become the dominant species on Earth. It’s pretty straightforward to achieve, just requiring a fuel source, a supply of oxygen from the air, and a way to initiate the reaction; then it self-sustains. You wouldn’t think there’s much to improve, but what about cooking with plasma? [Jay] from the plasma channel is no stranger here, and he thinks that there may be something in this idea, certainly enough to actually build something.

Now, let’s be straight with you, this isn’t a new concept, and you can buy a plasma-based cooking appliance right now. But they are all AC-powered devices. What if you want to go camping? [Jay] attempts (and succeeds) in building a portable, rechargeable 600W plasma cooking device that can actually cook food, but it was not all plain sailing.

The existing off-the-shelf ZVS driver modules available were a bit weak and unreliable, and the required flyback coils were hard to find with the right specs, so he needed to get down to work building custom parts. First off, the coils. Custom formers were resin-printed and machine-wound with 4000 turns of fine wire, and then resin-sealed into the former. [Jay] takes care to explain that it is crucial to get all the air out of the windings, or else local flashover breakdown will occur and wreck the coil in a short time. We reckon the resulting coils look amazing in their own right!

We do love a nicely wound coil. Oooh!

Next, the ZVS drivers on hand had low-quality capacitors (well, not enough capacitance anyway) and cheap driver transistors, so both were upgraded. The initial plan was to have four driver/coil pairs, each driving a single pair of electrodes, with a common ground ring connecting them all. It turns out this was a terrible idea: the drivers were not synchronised, so they were pulling on each other, causing catastrophic damage to the PCBs in a very short time. The solution was more complicated wiring, to give each coil secondary output a dedicated electrode pair, so there was no direct electrical connection between neighbouring coils and no coupling between them. A clever electrode arrangement meant that a pan would sit on top of a ring of electrodes, causing plasma discharges to jump directly to the pan, thereby concentrating localised heating there. We were wondering how this new direct connection (the pan is now a common connection!) didn’t also cause backfeeding and kill the ZVS drivers again, but it didn’t seem to happen.

Bang, smell, oops. The copper is supposed to be stuck to the PCB.

Anyway, [Jay] demonstrates what is possibly the world’s first rechargeable, portable plasma cooker capable of making breakfast. Which we think is very important in its own right, however, we would like a plasma-based solution to making toast next, perhaps a plasma knife that cooks the bread as you slice it?

If this plasma cooking lark rings a bell, yes, we did touch upon this way back in 2017. And whilst not strictly plasma cooking, you can make an amazing microwave plasma in this ridiculously upgraded appliance. Definitely do not try that one at home.

2026绿色动力挑战赛：SolMate 移动充电

2026-04-17T17:30:53+00:00

We’ve all had those times when your electrical device of choice is running low on battery. Sometimes you even plan ahead and are also carrying a battery pack, but what happens when you’ve forgotten to charge the battery pack? This is the problem that [Arnov Sharma] addressed with the SolMate, a portable solar panel that keeps a battery bank topped up.

The SolMate is built around an efficient 2 W photovoltaic panel that’s not much bigger than a cellphone. This panel can supply 5 V at 400 mA on a sunny day. The solar output is more than enough to keep the internal 2000 mAh battery topped up and ready for use. Charging the Li-ion battery is handled by an IP5306 power management chip, which pulls double duty: it safely regulates charging while boosting the battery’s 3.7 V to the 5 V expected at the USB charge ports. Speaking of charge ports, the SolMate includes both a USB-A and a USB-C port, plus a switch to enable or disable the unit.

The case is all 3D printed, with some clever design choices. Offsetting the bulk of the battery and PCB storage area to one side lets the SolMate naturally cant toward the sun. Even the clip used to attach it to a backpack is printed.

Be sure to check out the other entries into our latest challenge!

Hackaday播客第366集：在吐司机上玩DOOM，用LED灯带制作音乐，旧硬盘的新装

2026-04-17T16:30:38+00:00

It’s the evening before publication, and a pair of Hackaday writers convene to record the week’s podcast. This week Elliot Williams is joined by Jenny List, and it’s a bumper episode!

Of course, a bit of Hackaday news makes the cut, as it’s time to make an entry in the Green Powered Challenge. Then we make the first of a couple of sojourns into AI, as we talk about the Linux kernel stance on AI code. In short: if you submit AI code you’re responsible for its bugs. Meanwhile out of this world, we look forward to a time when astronauts breathe oxygen from moon dust.

There are hacks aplenty for your enjoyment, starting with far more than you ever thought it was possible to know about sound-reactive LED strips. Then we have among others a Mac on an ESP32 forming the UI for a weather monitor, Doom on a toaster, and a fascinating look at screw threads for plastic.

In the longer reads we have our colleague [Tom Nardi] finding Chinese people’s personal data on hard drives he bought in an electronics store, and an attempt to look at what an LLM can do that might be useful. Grab your headphones, and join us!

Download your own Podcast!

Episode 365 Show Notes:

News:

What’s that Sound?

Congrats to [Sean rules] for guessing the pencil sharpener!

Interesting Hacks of the Week:

Quick Hacks:

Can’t-Miss Articles:

可穿戴电路雕塑是一款智能手环

2026-04-17T15:30:59+00:00

Well, this might just be a Hackaday first. Certainly not the circuit sculpture part, nor the wearable aspect, but the glorious combination of the two. Behold [CMoz]’s Fashionably on Task: a Smart Bracelet for Forgetfulness. The name may be long, but the intent is concise: to showcase your top five must-dos for the day.

This lovely bracelet uses a tri-color e-paper display, and it’s WiFi enabled in order to receive input from the corresponding phone app. Although the cute pink ESP32-C3 is programmed in PlatformIO, the code will work with the Arduino IDE as well.

To get down to business, just power on the bracelet. If it can’t connect to the network you’ve hard-coded, it will broadcast it’s own access point. Connect with your phone to the custom web page, and Bob’s your uncle. From here, you can enter the tasks, change the colors around, mark tasks as complete, and remove tasks or reset recurring reminders.

The nifty part is that e-paper screen, since it will of course continue to display your list once powered down. Here’s the full code. Then you can deep-dive into the graph theory of circuit sculptures.

本周安全动态：Docker 认证、Windows 工具以及一次非常充实的补丁星期二

2026-04-17T14:00:33+00:00

CVE-2026-34040 lets attackers bypass some Docker authentication plugins by allowing an empty request body. Present since 2024, this bug was caused by a previous fix to the auth workflow. In the 2024 bug, the authentication system could be tricked into passing a zero-length request to the authentication handler. In the modern vulnerability, the system can be tricked into removing a too-large authentication request and passing a zero-length request to the authentication handler.

In both cases, the authentication system may not properly handle the malformed request and allow creation of docker images with access to stored credentials and secrets.

Bugs like these are increasing in visibility because AI agents running in Docker, like OpenClaw, may be tricked via prompt injection into leveraging the vulnerability.

Windows CPU Tools Compromised

videocardz.com notes that the popular Windows monitoring software Cpu-Z and HWMonitor appear to have been compromised. Reports indicate that the download site was compromised, not the actual packages, but that it was redirecting update requests to packages including malware. While the site has been repaired, unfortunately it looks like there is no warning to users that the downloads were compromised for a period of time.

Anecdotally, there has been a rash of Discord account takeovers in the past week, where long-standing accounts in multiple servers have been compromised and turned into spambots. While there is no evidence these events are linked, clearly a new credential or authentication stealing malware is in play, which involves stealing credentials from Discord.

X.Org and XWayland Updated

The X.Org and XWayland servers saw security updates this week, fixing a handful of vulnerabilities involving uninitialized memory use, use-after-free, and reading beyond the end of a buffer.

The vulnerabilities are generally classified as “moderate”, but of course, don’t leave known vulnerabilities when you can avoid it! Fixed releases should find their way into distributions soon.

OpenSSL 4.0 Released

OpenSSL released version 4.0 this week, adding support for Encrypted Client Hello / ECH / RFC9849 as well as deprecating some older SSL 2.0 behavior.

Encrypted Client Hello is a new enhancement to TLS (nee SSL) client handshake. When a client connects to a TLS server like a website, one of the first packets sent is the Client Hello which contains the TLS version, supported algorithms, and importantly, the server name the client is connecting to. Including the server name in the hello message allows modern multi-homed and cloud-based websites to function, because it indicates which web server and SSL certificate should be used to handle the request, but exposes the hostname the user is connecting to.

With ECH, the hello message is split into multiple messages, with the true hostname encrypted inside the second, inner message. The outer message allows routing the request to a server responsible for decrypting the inner communication and dispatching the request to the proper server. It is possible, for instance, for an ISP to see that a user has connected to a website on the Cloudflare infrastructure, but not which website hosted on Cloudflare.

For individual sites, the value of ECH is debatable – without a central server to dispatch to the specific hosts, the outer hostname is still readable – but for sites hosted behind load balancers, there is additional protection for users against identification of browsing habits. Although it brings extra complexity, adding new standards like ECH at least moves the needle towards better user privacy and protection by default.

Rockstar games breached (again)

Rockstar Games (of Grand Theft Auto and Red Dead Redemption fame) has been breached by a ransomware/extortion group. If this sounds familiar, in 2022 the company was breached and early GTA 6 gameplay was stolen.

This go around, the breach was actually of the data warehousing company Snowflake, via another service, Anodot. Used for cloud monitoring and analytics, Bleeping Computer reports that an Anodot breach was used to access Snowflake data, which is now used to extort Rockstar.

Rockstar says the data stolen does not impact players or the functioning of the company, and they will not be paying the ransom.

Linux Kernel Certificate OOB

Linux Kernel 7.0 releases this week, and includes a fix to out-of-bounds memory access in certificate handling. The fix is also being back-ported to stable and LTS kernel versions (Linux 6.4, 6.6 LTS, 6.12 LTS, 6.18 LTS, and 6.19) so be on the lookout for updates!

The out-of-bounds bug lies in the kernel keyring API; any user on the system can submit an invalid certificate to the kernel keyring. In this specific case the impact seems limited to a kernel crash instead of arbitrary privilege escalation.

NIST no Longer Enriching CVE

The NIST organization is no longer enriching CVE entries in the National Vulnerability Database, except for those in the Known Exploited Vulnerabilities catalog, used in federal government, or those in designated critical software. Previously, the NIST NVD provided additional information and severity rankings for reported vulnerabilities. Citing a lack of funding and an overwhelming number of reported vulnerabilities, they will no longer provide updated severity scores or details.

It’s understandable, but a net loss to the security community, and the Internet at large, when we lose analysis and commentary on risks. CVE details and risks are often self-assigned by the vendor, which can lead in some cases to a culture of “malicious compliance” where the released information is technically correct and complete, but contains little or no actual detail and assumes the least impactful interpretations. Third-party evaluation and classification by organizations like NIST offered additional context and analysis to identify the truly critical reports.

Patch Tuesday, Everybody Panic!

OK – don’t actually panic, but if you’re a Microsoft user, you already know. This month’s Patch Tuesday — the scheduled day for Microsoft updates, for anyone lucky enough not to have to observe — includes over 160 security updates. This makes it the second largest Patch Tuesday ever. It includes a fix to the publicly available Bluehammer exploit for bypassing Windows Defender, and over 60 patches for browser vulnerabilities.

Additionally, Chrome published fixes for 20 vulnerabilities, and Adobe published fixes for Reader, with evidence on both that the bugs are already being publicly exploited.

This is your monthly reminder to stay on top of security updates whenever they are available, on whatever platform you use. Unknown zero-day exploits might get all the attention, but outdated software with known, patched bugs can be the biggest vector for exploits and malware. Once a bug is known and patched, there is no reason to save the exploit for targeted attacks; the days and weeks after a bug is publicly fixed can be a wave of automated exploits, and many of the largest attacks use vulnerabilities fixed weeks or months prior.

Botconf Talks Streaming

Finally, a quick aside for anyone interested in pursing more related content, the Botconf EU conference about fighting botnets and malware is streaming the conference content; by the time this post goes live the conference is likely to be concluded, but the talk streams are accessible!

微软最终上调FAT32大小限制

2026-04-17T11:00:55+00:00

You probably don’t spend a lot of time using the FAT32 file system anymore, since it’s thoroughly been superseded many times over. Even so, Microsoft has seen fit to deliver an upgrade for FAT32 for the latest Windows 11 Insider Preview build. Finally, the stock Windows tools will let you format a FAT32 drive up to 2 TB instead of locking you to a 32 GB maximum!

The size limit was never baked into the FAT32 spec itself. With a 32-bit field for counting sectors, the file system supports up to 2 TB volumes with 512-byte sectors. However, as explained by former Microsoft developer [Dave Plummer], it just so happened that the 32 GB limit came about because of a random decision made when slapping together the Format dialogue box over 30 years ago.

The pending change was first announced in 2024, affecting the command line format tool as well. It’s actually been possible to create larger FAT32 volumes for some time, you just couldn’t easily do it with Microsoft’s standard formatting tools.

FAT32 is still a terrible file system to use in 2026, mostly because it has a hard limit on file size that tops out at 4 GB. It’ll ruin your life if you’re shooting HD or 4K video. We often don’t spend a lot of time musing over file systems in detail, but they’re right at the heart of everything we do on our computers on a daily basis. Sometimes, it bears thinking about!

游戏立方体热水浴缸动画带入现实

2026-04-17T08:00:02+00:00

Unlike the current era where most consumer electronics are black rectangles, or the early 90s where most consumer electronics were black rectangles, we got a brief glimmer of color, light, and hope in the 2000s. Cell phones had all kinds of shapes and sizes, laptops came in bright colors, and even video game consoles got in on the fun. The Nintendo GameCube not only featured its namesake shape but came in several vibrant colors, most famously a bright purple. In fact, its design was such a hit that it continues to inspire artists and console modders alike. An animator named [kidd.gorgeous] recently envisioned a GameCube as a hot tub, and [BigRig Creates] set out to make this animation a reality.

Of course, this won’t be a life-sized hot tub capable of holding a human, but [BigRig Creates] did want it to be a usable, playable Game Cube with all of the features from the animation present in the final version. Since the lid won’t be operational with a hot tub model on the top lest all of the water spill out every time a game is changed, he’s modified it with some modern tools to hold his games inside the console itself. With the case open he’s also added the LED accent lighting featured in the animation as well as added the 3D-printed hot tub to the top. The hot tub is filled with mineral oil for electronics safety, and has a small pump built in to give the appearance of a working hot tub.

The buttons around the outside are functional as well, toggling the various lighting features and hot tub operation. And of course, the console diorama is fully playable, with the staircase railing able to easily detach in order to access the leftmost controller ports. It’s a faithful adaptation of the original animation, and [BigRig Creates] has a few games on queue that are properly themed for the new hot tub addition like Wave Race 64, Super Mario Sunshine, and Pool Paradise.

组装大型遥控迷你卡车

2026-04-17T05:00:13+00:00

Kei trucks are some of the smallest commercial vehicles out on the roads today. You can also get lots of cute kei RC cars if you’re into the toy side of things. [Victor] wanted to split the difference with Truck-Kun, and built a 1:3 scale kei truck to show off at a recent anime convention.

The truck is modelled on the Suzuki Carry.

The build started with a classic hacker favorite—a bunch of old hoverboard motors. These brushless hub motors are pretty easy to drive and have plenty of torque right out of the box. A simple ladder frame was whipped up with a hoverboard wheel at each corner, with a body whipped up out of cardboard, paint, and a few 3D printed parts to hold everything together.

Steering was courtesy of a leadscrew and stepper motor from an old 3D printer. RP2040s were thrown in to talk to the motor controllers, while an ESP32-S3 was charged with communicating with an Xbox One controller to receive directional commands.

It was a neat build that stood tall over most any other kei RC car at Anime Los Angeles. Only, it ended in tears when the 100-pound machine accidentally drove at full throttle into a wall and smashed itself to pieces. Still, that’s just an excuse for [Victor] to build a better one for next year.

We love a good RC build around these parts, and we love kei trucks too.

对比12伏直流空气炸锅与常规的240伏交流空气炸锅

2026-04-17T02:00:23+00:00

That boiling water is a contentious topic of discussion is clear, but what about hot air? When you take a 12 VDC, 280 Watt-rated air fryer and pit it against a bog-standard 240 VAC, 1400 Watt unit, which one would you want to use when you’re doing some camping or other exciting off-the-grid opportunities? Unlike with boiling water the physics aren’t as clear-cut here, so [Cahn] did some testing to figure out exactly what the efficiency numbers look like

Since air fryers rely on the transfer of thermal energy from the resistive heating element into the food, any thermal energy that’s not immediately transferred is effectively wasted. This, combined with the relatively low power rating and thus much higher time demand of the low-voltage air fryer is enough to set one’s expectations pretty low.

As scientific test samples chicken nuggets were used with the test, following a preheating period for the 12 VDC unit. Both units managed to hit a safe temperature inside the nuggets after 20 minutes, thus successfully staving off food poisoning, but the browning with the 240 VAC air fryer was much better.

As for the efficiency, the 12 VDC unit required 150 Wh for 20 minutes plus the 10 minutes of preheating, with 45 minutes total at 225 Watt to get proper browning. Meanwhile the 240 VAC unit burned through 250 Wh in 20 minutes, with no pre-heating, though only 230 Wh with no inverter losses included. As a final test, the 12 VDC unit was run at 400 Watt using 14.6 VDC input, which did indeed get it up to temperature much faster.

Thus both are equivalent, just with the caveat that the low-voltage unit will take considerably more time to get the same result. This mirrors the results with boiling water, where most options mostly vary in how much time they require to get water up to a boiling temperature.

我们真的需要另一个开发板吗？

2026-04-16T23:00:54+00:00

It’s fair to say that there are a lot of development board form factors for MCUs, with [Tech Dregs] over on yonder YouTube on the verge of adding another one to the pile, but not before he was having some serious thoughts on the implications of such a decision. Does this world really need another devboard with the ubiquitous 2.54 mm (0.1″) pitch pin headers, all so that it can perhaps be used in the same traditional 2.54 mm pitch breadboards?

The thought that [Tech Dregs] is playing with is to go for something more akin to the system-on-module (SoM) approach that’s reminiscent of the Raspberry Pi compute module form factor. This means using a 1 mm pitch for the headers and castellated edges in case you want use it as an SMT part, while breaking out many more pins of the onboard ESP32 module in far less space.

Obviously, the main advantage of this approach is that much like with compute modules you can leave most of the tedious cheap stuff on a carrier board, while the expensive to manufacture components are on a self-contained module. Meanwhile with the much finer pitch on the SoM contacts it’d straddle the divide between a 2.54 mm breadboard-capable devboard and a fully custom PCB, while making any mistakes on the carrier board much cheaper to redo.

The counterpoint here is of course that something like an ESP32 module is already a module with a finer pitch, but if you need more than just what it offers, or you want to use an STM32 or RP MCU across boards it could make a lot of sense.

Having 1 mm pitch breadboards would honestly also be rather nifty, natch. That said, what are your thoughts on this matter?

FPGA驱动DIY USB示波器和信号发生器

2026-04-16T20:00:11+00:00

Oscilloscopes and to lesser extent signals generators are useful tools for analyzing, testing and diagnosing circuits but we often take for granted how they work. Luckily, [FromConceptToCircuit] is here to show us how they’re made.

[FromConceptToCircuit] starts by selecting the hardware to use: an Artix-7-based FPGA and an FT2232 USB-serial converter. RS245 in synchronous FIFO mode is selected for its high bandwidth of about 400 Mbps. Then, they show how to wire it all up to your FPGA of choice. Now it’s time for the implementation; they go over how the FT2232 interfaces with the FPGA, going through the Verilog code step-by-step to show how the FPGA makes use of the link, building up from the basic transmission logic all the way up to a simple framed protocol with CRC8-based error detection. With all that, the FPGA can now send captured samples to the PC over USB.

Now it’s PC-side time! [FromConceptToCircuit] first explains the physical pipeline through which the samples reach the PC: FPGA captures, transmits over RS245, FT2232 interfaces that with USB and finally, the software talks with the FT2232 over USB to get the data back out. The software starts by configuring the FT2232 into RS245 mode, sets buffer sizes, the whole deal. With everything set up, [FromConceptToCircuit] explains how to use the FT2232 driver’s API for non-blocking communication.

As a bonus, [FromConceptToCircuit] adds a signal generator feature to the oscilloscope using an I2C DAC chip. They start by explaining what exactly the DAC does and follow up with how it’ll be integrated into the existing system. Then it’s time to explain how to implement the I2C protocol bit-for-bit. Finally combine everything together for one final demo that shows a sine wave on the DAC’s output.

桌面数字化仪让笔记捕获变得轻而易举

2026-04-16T18:30:22+00:00

While it might seem quaint these days, we’ve met many makers and hackers who reach for a pen and a pad when learning something new or working their way through some technical problem. But even if you’re the type of person who thinks best when writing something out on paper, there’s still a good chance that you’ll eventually want to bring those notes and sketches into the digital realm. That’s where things can get a little tricky.

[Spencer Adams-Rand] recently wrote in with his clever solution for capturing written notes and pushing them into Notion, but the hardware design and digitization workflow is flexible enough that it could be adapted to your specific needs — especially since he was good enough to release all the files required to build your own version.

Whether they are hand-written notes, old photographs, or legal documents, digitization boils down to taking a high resolution digital photo of the object and running it through the appropriate software. But getting good and consistent photos is the key, especially when you’re working your way through a lot of pages. [Spencer] started out just snapping pictures with his phone, but quickly found the process was less than ideal.

His custom scanning station addresses that first part of the problem: getting consistent shots. The images are captured using a Raspberry Pi 5 with attached Camera Module 3, while the 3D printed structure of the device makes sure that the camera and integrated lighting system are always in the same position. All he needs to do is place his notepad inside the cavity, hit the button, and it produces a perfect shot of the page.

Using a dedicated digitizing station like this would already provide better results than trying to freehand it with your phone or camera, but [Spencer] took things quite a bit farther. The software side of the project puts a handy user interface on the 5 inch touch screen built into the top of the scanner, while also providing niceties like a REST API and integration with the OpenAI Vision API for optical character recognition (OCR).

Those with an aversion to AI could certainly swap this out for something open source like Tesseract, but [Spencer] notes that not only is OpenAI’s OCR better at reading his handwriting, it spits out structured markdown-like data that’s easier to parse. From there it goes into the Notion API, but again, this could be replaced with whatever you use to collect your digital thoughts.

A device like this would go a long way towards answering a question we posed to the community back in January about the best way to digitize your documents.

肯塔基州洞穴战争，以及1925年的病毒式传播

2026-04-16T17:00:15+00:00

Floyd Collins, the unfortunate star of this post. (Public Domain)

Information, it seems, flows at the speed of media. In the old days, information traveled with people on ships or horses, so if, say, a battle was won or lost, it could be months or even years before anyone back home knew what happened. While books and movable type let people store information, they still moved at the speed people moved. Before the telegraph, there were attempts to use things like semaphores to speed the flow of information, but those were generally limited to line-of-sight operations. Carrier pigeons were handy, but don’t really move much faster than people.

The telegraph helped, but people didn’t have telegraph stations in their homes. At least not ordinary people. But radio was different. It didn’t take long for every home to have a radio, and while the means of broadcasting remained in the hands of a few, the message could go everywhere virtually instantly. This meant news could go from one side of the globe to the other in seconds. It also meant rumors, fads, and what we might think of today as memes could, too.

You might think that things “going viral” is a modern problem, but, in reality, media sensations have always been with us. All that changes is the number of them and their speed.

One of the earliest viral media sensations dealt with William Floyd Collins, an unfortunate man who was exploring caves during the Kentucky Cave Wars.

Background

Mammoth Cave in Kentucky had become a major tourist attraction. The accessible entrance to the cave was located on land owned by the Croghan family. The massive cave system had been made famous in the 19th century, and with the construction of a lock and dam nearby in 1906, Mammoth Cave became accessible to ordinary tourists.

Rescuers weighing options at the entrance to Sand Cave. (Public Domain)

However, the cave wasn’t completely under the Croghan land. There were also other caves that may or may not have been connected with Mammoth Cave. This led to fierce competition. The Croghan family suppressed information about exactly what land was over the cave. Meanwhile, other cave “owners” would intercept people heading for the cave, tell them that Mammoth Cave was closed, and “helpfully” direct them to another location.

By the 1920s, George Morrison blasted new entrances to the cave on non-Croghan land. There was fierce interest in finding new entrances to the cave or nearby caves to capture tourist money.

Back to Floyd

Floyd Collins found an entrance into what would become known as the “Great Crystal Cave” in 1917 and opened it to tourists in 1918. Unfortunately, the cave was hard to access, so it didn’t make much money.

Floyd had started entering caves in 1893 at the age of six. He discovered his first cave in 1910. But Great Crystal Cave was too far off the main road. He entered into a deal with three farmers who owned land closer to the main highway. If Floyd could find a suitable cave or, even better, an entrance to Mammoth Cave, he’d partner with them and create a mutually profitable tourist attraction.

Floyd found a hole in what would become known as Sand Cave. Some of the passages he had to move through were as tight as 9 inches, which, of course, would not be suitable for tourists, but they opened, apparently, into a large grotto. He was determined to expand the entrance to make the cave commercially viable.

In January of 1925, he was working in the cave when his gas lamp started to dim. He tried to leave, but while trying to move through a small passage, he knocked over his light, leaving him in total darkness.

In the dark, he put his foot against a seemingly stable wall and caused a shift that pinned his leg with a rock weighing nearly 30 pounds. He was also buried in gravel. At this point, he was 150 feet from the hole to the surface.

The Media

The next day, people noticed Floyd was missing, but no one would dare to follow him through the narrow passages. His younger brother finally got close enough to determine what happened. He was able to give Floyd food and water as plans for a rescue developed.

After four days in the cave, several people tried to pull Floyd out using a rope and a harness, but they only wound up injuring him. Meanwhile, the media had taken interest in the case, and the publicity drew hundreds of tourists and amateur spelunkers. Campfires and, possibly, the electric light that had been placed to give Floyd some light and warmth, melted ice inside the cave, creating puddles of water around the trapped man.

Two days after the failed rescue attempt, rain and the melting ice caused the cave passage to collapse, and the rescue team determined it was too dangerous to dig it back out after making an attempt to do so. They decided to dig straight down to reach Floyd.

Digging

Unfortunately, the cave drew air in so they decided they could not use mechanical diggers without risking suffocating Floyd. That meant humans would have to dig the 55-foot shaft to reach the victim. The initial estimate that 75 volunteers could dig the shaft in 30 hours proved optimistic, as conditions worsened and the hole grew deeper.

Someone disconnected the wires from the light bulb and connected them to an audio amplifier to detect signs of life from the victim. They believed the repetitive crackling noise meant he was breathing.

The light bulb went open on February 11th, twelve days after the incident started. Five days later, they reached his body. He had died and had been dead for several days.

You can find a well-done documentary from Remix Films in the video below. For a movie inspired by the event, check out the Billy Wilder film Ace in the Hole (1951) starring Kirk Douglas.

Viral

A newspaper reporter, William Miller, was on the scene and, being a small man, was able to actually help remove gravel from Floyd before the cave-in. His interview with the man from inside the cave won a Pulitzer Prize.

Not a circus. A cave rescue.

There was a time when this would have been only a sensational local story, but by the modern year of 1925, reports “went out on the wire” by telegraph and were picked up by newspapers worldwide. The nearest telegraph station was miles away, so two ham radio operators (9BRK and 9CHG) provided a link between the site, the newspaper, and the authorities.

The first broadcast radio station, KDKA, was only five years old, but stations provided news bulletins detailing the progress. Thanks to the media, crowds were reported to number in the tens of thousands. Eventually, the National Guard arrived to help control the crowds.

Vendors popped up to sell hamburgers and memorabilia like a macabre circus. As you can see in the video below, memorabilia about the event and Floyd Collins can be worth a pretty penny to collectors.

The whole thing became one of the three largest media events between World War I and World War II. The other two were Lindbergh’s transatlantic flight (1927) and the kidnapping of Lindbergh’s baby (1932). Oddly, Lindbergh was an acquaintance of Floyd’s and also flew news photos from the scene (although, reportedly, to the wrong newspaper).

While it wasn’t quite as big an event, Canada’s 1936 Moose River Gold Mine collapse was a similar situation and also received worldwide media attention. It has the distinction of being the first 24-hour radio coverage of a breaking news story in Canada.

Today

These days, sensational news stories pop up everywhere. It seems as if they hardly get started when they are displaced by another one. But we submit that “going viral” isn’t a modern phenomenon. Only the speed at which it happens. Even an 1835 newspaper was able to spur a viral hoax.

Featured image: “Mammoth Cave Saltpeter Mine” by [Bpluke01]

EpiPen紧急杠杆式启动装置

2026-04-16T15:30:04+00:00

Imagine you and your friend are enjoying a nice sunny day, and BAM — they start to have a severe allergic reaction to who knows what. You have an EpiPen, but your friend is on the other side of a field! The solution? Obviously [Emily The Engineer] has only one option: build an entire EpiPen launcher!

Starting off the life-saving project, [Emily] prototyped with a 3D printed blank and a simple solenoid-controlled glorified potato cannon. This proved effective, as one would expect of such a project after successful tests on a human subject. However, there was one simple problem: what if you missed your initial shot?

To ensure no possible failed missions, a bolt-action magazine was retrofitted onto the device. Additionally, an air compressor placed in a mobile backpack carrier allows for repeated mobile use. Official testing was done on ballistic gel before a “war game” scenario played out involving an anaphylactic friend. As one would assume, this went perfectly, ignoring the time delay of having to wait for the compressor to build up enough pressure…

Anyways, even if you won’t be using this EpiPen launcher anytime soon, there are some actual DIY medical miracles you can look into! Something that’s a tad less insane to hack together than an EpiPen gun would be a splint. That is exactly what you can learn about here!

2026年尝试自行搭建消费级路由器

2026-04-16T14:00:07+00:00

Although we have many types of networking equipment with many unique names, at their core they can usually be reduced to just a computer with some specific peripherals. This is especially the case for something like a router, a device found in just about any home these days. Certain consumer-grade routers may contain something special like a VDSL modem, but most of them just have a WAN Ethernet jack on one end and one or more LAN-facing Ethernet ports.

All further functionality is implemented in software, including any firewall, routing and DHCP features. What this means is that any old PC with at least two Ethernet ports or equivalent can be a router as long as you install the appropriate software.

In this article we’ll be taking a look at what consumer-level options there exist here today, ideally something so simple that the average home user could set it up with a bit of coaching.

When Routers Were Terrible

Start-up screen of FreeSCO. (Credit: Lewis “Lightning” Baughman, Wikimedia)

Although these days most people are probably happy to use whatever consumer router gets tossed their way by their ISP or purchased from any of the many electronics retailers, the era of home routers starting in the 90s was a rough one. Not only did many ISPs for a long time have strict rules against even thinking of installing any kind of connection-sharing router on the precious internet connection they so gracefully provided access to, the routers that were available even during the 2000s also weren’t particularly good.

Perhaps the most egregious issue that you’d run into as a somewhat internet-savvy user was that these units came with the tiniest amount of RAM and a routing table that would fill up within seconds if you so much as thought of doing anything naughty like downloading Linux ISOs with Bittorrent. Even more than the lightest internet use risked issues such as the router freezing up and requiring a reboot.

Being the nerdy type who’d hang out on various tech forums during that time, I instead opted to toss some ISA NICs into an old 486DX2-66 PC, put FreeSCO on a 1.44 MB floppy and used that instead. Even the 16 MB or so of EDO or FP RAM in that old system easily kept up with whatever internet traffic-related abuse I threw at it along with the rest of the family.

Later I’d also use Smoothwall on a trashed Pentium 166 system that I had salvaged from e-waste as a healthy upgrade. Yet over time as consumer routers stopped being terrible, I’d end up using those instead. That said, despite recent improvements there are still reasons today to put your own router together.

These reasons can vary from a simple increase in control, as well as peace of mind in an era where commercial consumer router firmware do not appear to get quite the quality assurance they should get. There is also the specter of repressive government policies, and of course less dramatic reasons such as just being into tinkering with networking gear.

Modern Options

Although we have moved on quite a few years from the 2000s, it’d be nice if we could still theoretically dig an ancient 486 PC out of a closet and use it, just to push back the minimum requirements as far as possible. Looking at the list of router and firewall distributions over at Wikipedia we can get a bit of an idea of what’s out there today. Naturally we’d like something that’s open source, recently updated and well-supported as we’ll be exposing it to the worst that the modern Internet has to offer.

Sadly, FreeSCO isn’t even the list, and based on the project’s website it doesn’t appear to have been updated since 2014. Similarly, Smoothwall’s free version (Express) doesn’t appear to be quite supported any more, with the commercial offering being promoted instead. A promising project is something like OPNsense, which is a FreeBSD-based distribution that does basically everything networking-related and then some. It’s also rather resource-heavy, demanding 4+ GB of RAM and so on, for features that the average home user is unlikely to use.

Ultimately only two projects really jump out: the first is the Linux Embedded Appliance Framework (LEAF), which feels like a spiritual successor to FreeSCO in targeting 486 or better hardware while running read-only off floppy discs or better.

The second project is the well-known OpenWrt, which doesn’t just provide alternative firmware images for commercial routers, but also for generic x86 and ARM hardware. This should tick all the aforementioned boxes, including the ability to run off an x86-compatible potato, while theoretically also providing driver support for just about any NIC and maybe even WNICs.

Amusingly, my current internet router is a Xiaomi device that runs OpenWrt with a custom web UI, so I’m already familiar with it in that sense. I also have no complaints about its stability, so this should be smooth sailing as a first try.

A quick sidenote here: although I did say that consumer-grade routers aren’t terrible any more today, this Xiaomi router replaced a horrid TP-Link Archer C80 with its tiny 4 MB Flash, incompatibility with OpenWrt and broken IPv6 support. Caveat emptor, as they say, for consumer-grade routers can absolutely still be terrible.

It’s Not Hoarding, I Swear

Like any self-respecting tech-enthusiast I happen to have a few drawers and boxes with relevant computer hardware, including a stash of industrial Intel Atom mini-ITX boards and some PCI Gbit NICs. Although I could have tossed this into a scruffy old case with a questionably-but-probably-fine PSU, I decided to get a new budget case and PSU, just to feel more confident about flipping on the power switch:

With the powers of e-waste and non-sketchy PSUs combined… a DIY router. (Credit: Maya Posch)

The mainboard is a 2009-era Intel Desktop Board D410PT, with two sticks of 1 GB DDR2 RAM. There’s just an onboard 10/100 Mbit Ethernet adapter in addition to the PCI 1 Gbit adapter, but this is good enough to test basic principles. Plus older Ethernet chips definitely ought to be supported.

With this whole kit built together, the OpenWrt Wiki entry on x86 installations was followed, with an Ext4-based non-EFI image written to a 512 MB microSD card. This card was inserted into the system with a USB adapter, though as the Wiki page makes clear there are many more ways to install OpenWrt, including on an internal (SATA/IDE/etc.) storage device.

Here the OpenWrt image selection process is definitely a bit more complicated than that for e.g. FreeSCO, as you have to know whether the system boots using a UEFI or traditional BIOS, as well as decide between the SquashFS and Ext4 root partition option. Even after looking at the pros and cons of both I’m still torn on which one is best.

There is also the issue of supported network cards, with the typical image supporting Intel and Realtek Ethernet chipsets. Having compatible NICs installed is even more important than back in the generic NE2000 days.

Sad Trombone

After firing up the newly built rig and briefly powering it down to replace the very much drained CMOS battery, I had hoped to see something promising on the connected screen. Yet other than the PXE boot attempts from both Ethernet cards before a sad ‘no bootable devices found’ error message nothing happened.

I did go through the BIOS to make sure that all ‘boot from USB’ options were enabled and tried multiple SD-card-to-USB adapters in multiple USB ports, but it would seem that either the provided OpenWrt image isn’t bootable or the Curse of Bootable USB Devices has struck again.

While I haven’t given up yet on DIY-ing an x86 router on older hardware in 2026 and still have to give LEAF a shot, I must say that compared to the FreeSCO method of just slapping that floppy into the FDD, running through the config and being done, things are a lot more complicated today. I’d definitely not trust the average person to run through these steps to set up their own consumer-grade router.

Perhaps for the LEAF version I’ll dig up a rig with FDD header on the mainboard and run it off a 1.44 MB floppy like in the good old days. Until then feel free to sound off in the comments about what obvious mistakes I have made with OpenWrt, how I should just use OPNsense on an old 19″ rackmounted Xeon server rig, or maybe your own experiences with FreeSCO, SmoothWall, OpenWrt, and/or LEAF on lovingly restored e-waste.

Featured image: “Ethernet Router” by [gratuit].

Optocam Zero 的照片堪称完美

2026-04-16T11:00:54+00:00

Yeah, we know, there’s a camera on your phone that does this and that. But these days its become trendy to turn towards older digital cameras in place of smartphones, and we can tell you from experience, that the joys of having a dedicated photographing contraption are many.

There’s plenty of cheap digital cameras on the secondhand market, but instead of fending off the clothes bros and other reseller types at the thrift store, stay home and build [Doruk Kumkumoğlu]’s Optocam Zero. Inspired by the Kodak Charmera and the like, [Doruk] aimed for something that’s playful, enjoyable, and intuitive to use.

Optocam Zero uses an auto-focus camera module and features eight photo filters. The screen dims when inactive to preserve battery life, but it can be charged back up with USB-C, and you can use it for the duration. And unlike my young adult camera, you don’t have to take out the SD card to see the pictures, just use the custom hotspot interface to transfer them.

If you need to capture a cat yawning real fast, you might want to just use your phone instead, because takes 22 seconds for Optocam Zero to get camera-ready. But when you have a less spontaneous subject in mind, this thing looks like a great choice. Be sure to check out the excellent build guide (PDF) whether you build one or not.

We think the images from the Optocam Zero look pretty crispy. But if you want to go lo-fi, we have that, too.