Machine learning has long been a topic of interest for humanity, but only in recent years have we had broad access to great computing power to enable to the average person to dive in. [Yosh] recently decided to put an AI to work learning how to race in Trackmania.

After early experiments with supervised learning, [Yosh] decided to implement a genetic algorithm to produce an AI to drive in the game. The AI takes distance from the track walls as an input, and has steering and accelerator values as an output. Starting with 100 AIs in generation 1, [Yosh] iterated by choosing the AIs that covered the longest distance in 13 seconds. Once the AIs started to get the hang of the first few corners, he changed the training to instead prioritize the lowest time taken to traverse each of the checkpoints along the track.

The AI improved over time, and over 100 generations, got down to a 23.48s time on the test track, versus 19.63s for [Trabadia], a talented human. We’d love to see how much better the AI could do with more training. [Yosh] is trying more experiments, like providing extra feedback in the AI fitness function to keep it from hitting the walls. It’s not the first time we’ve seen a genetic algorithm used to train a racing AI, either. Video after the break.

source https://hackaday.com/2021/01/17/ai-learns-to-drive-trackmania/

Terry Pratchett once wrote, “In ancient times cats were worshipped as gods; they have not forgotten this”. [Jonathan]’s cat has clearly not forgotten, and makes it loudly known whenever her favorite chair needs to be moved to stay in the spot of sunlight. He was looking for a fun hack anyway, so he decided to give in to her majesty’s demands, and automated the task.

[Jonathan] first considered adding motorizing the chair itself, but decided to keep it simple and just drag the chair across the room with a spool attached to a motor. The rope spool was attached to a small geared DC motor, mounted on a salad bowl base, and connected to an ESP8266 via a motor driver. The ‘8266 is running  NodeMCU with a web server that accepts simple motor commands through a RESTful API. This setup can’t reset the chair to it’s starting position at the end of the day, but this is a small price to pay for simplicity. The motor was a bit underpowered, but it only needed to move the chair in small distances at a time, so [Jonathan] removed the chair’s back to reduce the weight, and upped the motor voltage.

Determining when and how far to move the chair is the second part of the challenge. [Jonathan] considered a simple lookup table for the time of day, but the motor’s movement wasn’t consistent enough. The final solution was a set of three BH1750 digital ambient light sensors to give feedback. A pair of sensors on the chair determines its position relative to the sunny spot, by comparing light levels to a reference sensor mounted in the window. These light sensors are also attached to NodeMCUs, and send movement commands to the winding unit as necessary.

Unfortunately, it appears the hack was in vain, due to the end-user being frightened off by the motor noise. However, it still helped [Jonathan] to scratch his hacking itch, and fortunately for us, he documented the adventure and shared all the code.

This is the first cat chair we covered, but we have seen a few hacks to feed them or keep them away.

source https://hackaday.com/2021/01/17/serving-the-feline-masters-a-chair-to-follow-the-sunny-spot/

It’s a little known secret that when the Hackaday writers gather in their secret underground bunker to work on our plans for world domination, we often take breaks to play our version of the corporate “Buzzword Bingo”, where paradigms are leveraged and meetings circle back to loop in offline stakeholders, or something like that. Our version, however, is “Comment Line Bingo”, and right in the middle of the card is the seemingly most common comment of all: “You should have used a 555,” or variations thereof.

So it was with vicious glee that we came across the Trollduino V1.0 by the deliciously named [Mild Lee Interested]. It’s the hardware answer to the common complaint, which we’ll grant is often justified. The beautiful part of this is that Trollduino occupies the same footprint as an Arduino Uno and is even pin-compatible with the microcontroller board, or at least sort of. The familiar line of components and connectors sprout from the left edge of the board, and headers for shields line the top and bottom edges too. “Sketches” are implemented in hardware, with jumpers and resistors and capacitors of various values plugged in to achieve all the marvelous configurations the indispensable timer chip can be used for. And extra points for the deliberately provocative use of Comic Sans in the silkscreen.

Hats off to [Lee] for a thoroughly satisfying troll, and a nice look at what the 555 chip can really do. If you want a more serious look at the 555, check out this 555 modeled on a breadboard, or dive into the story of the chip’s development.

source https://hackaday.com/2021/01/17/arduino-wannabe-should-have-used-a-555-oh-wait-it-does/

Shutter shades were cool once upon a time, but if you really want to stand out, it’s hard to go past aggressively bright LEDs right in the middle of your face. A great way to achieve that is by building a pair of RGB glasses, as [Arnov Sharma] did.

The design intelligently makes use of PCBs to form the entire structure of the glasses. One PCB makes up the left arm of the glasses, carrying an ESP12F microcontroller and the requisite support circuitry. It’s fitted to the front PCB through a slot, and soldered in place. The V+, GND, and DATA connections for the WS2812B LEDs also serve as the mechanical connection. The right arm of the glasses is held on in the same way, being the same as the left arm PCB but simply left unpopulated. A little glue is also used to stiffen up the connection.

It’s a tidy build, and one that can be easily controlled from a smartphone as the ESP12F runs a basic webserver which allows the color of the glasses to be changed. It’s not the first time we’ve seen a flashy pair of LED shades either! Video after the break.

source https://hackaday.com/2021/01/16/rgb-glasses-built-from-pcbs/

There are many schools of thought when it comes to keeping vinyl records clean. It’s a ritual that’s nearly as important as the one that comes after it — queuing up the record and lowering the needle. We’ve seen people use everything from Windex and microfiber towels to ultrasonic cleaning machines that cost hundreds or even thousands. In the midst of building a beefier ultrasonic record cleaner and waiting for parts, [Baserolokus] looked around at all the LEGO around the house and decided to build a plastic prototype in the interim.

The idea behind ultrasonic cleaning is simple — high-frequency sound waves pumped through distilled water produce tons of tiny bubbles. These bubbles gently knock all the dirt and grime out of the grooves without using any brushes, rags, or harsh cleaners. [Baserolokus] built two pieces that hang on the edge of a washtub. On one side, a Technic motor spins the record at just under one RPM, it spins against a 3D printer wheel embedded in the other side. Check it out in action after the break.

Cleaning your vinyl is a great first step, but you might be ruining your records with a sub-par turntable. Take a deep dive with [Jenny List]’s thorough primer on the subject.

source https://hackaday.com/2021/01/16/diy-lego-record-cleaner-is-revolutionary/

Fused Deposition Modelling (FDM) 3D printers which squirt out molten plastic layer by layer are by far the most popular type in general use. Most machines extrude plastic through a nozzle above print bed, and struggle to produce parts with overhangs without using support material. However, a German team of researchers have recently come up with a solution.

In a prototype built by researchers at the Zurich University of Applied Sciences (ZHAW), a standard Cartesian printer has a third rotary axis added, upon which the nozzle can rotate. Additionally, the nozzle is angled at 45 degrees to the print bed, rather than the usual perpendicular setup. This allows layers of a print to be built up in such a way that support material is not needed for the vast majority of typical overhangs. This is particularly useful for hollow parts, where removing support material can be particularly difficult.

The team believes that such technology could be implemented on existing printers by way of a simple upgrade kit, and we can imagine a few experimenters will be champing at the bit to try it out. If you do, be sure to drop us a line. Alternatively, consider using a marker to make removing supports easier. Video after the break.

[Thanks to Andreas for the tip!]

source https://hackaday.com/2021/01/16/3d-printing-without-support-material-thanks-to-an-additional-axis/

The basics of digital logic are pretty easy to master, and figuring out how the ones and zeroes flow through various kinds of gates is often an interesting exercise. Taking things down a level and breaking the component AND, OR, and NOR gates down to their underlying analog circuits adds some complexity, but the flow of electrons is still pretty understandable. Substitute all that for photons, though, and you’ll enter a strange world indeed.

At least that’s our take on [Jeroen Vleggaar]’s latest project, which is making logic gates from purely optical components. As he himself admits in the video below, this isn’t exactly unexplored territory, but his method, which uses constructive and destructive interference, seems not to have been used before. The basic “circuit” consists of a generator, a pair of diffraction patterns etched into a quartz plate, and an evaluator, which is basically a pinhole in another plate positioned to coincide with the common focal point of the generator patterns. An OR gate is formed when the two generators are hit with in-phase monochromatic light. Making the two inputs out of phase by 180° results in an XOR gate, as destructive interference between the two inputs prevents any light from making it out of the evaluator.

All seven basic gates are possible using variations of this technique, but of course, actually building such things poses challenges. The microscopic manufacturing techniques, including photolithography and etching patterns on thin metal coatings, are pretty non-trivial. And while [Jeroen] was able to construct a few basic gates and test them, there’s still a long way to go. He says that he needs to add both input beam controls — LCD shutters, perhaps — as well as output detection.

The potential applications of this are pretty exciting, and we’ll be staying tuned for details. For deeper background on the dual nature of light and how interference patterns form, check out [Jeroen]’s explanation of the double-slit experiment.

source https://hackaday.com/2021/01/16/interference-patterns-harnessed-for-optical-logic-gates/

USB-C Power Delivery 3.0 (PD3.0) introduces a new Programmable Power Supply (PPS) mode, which allows a device to negotiate any supply of 3.3-21 V in 20 mV steps, and up to 5 A of current in 50 mA steps. To make use of this new standard, [Ryan Ma] create the PD Micro, an Arduino-compatible development board, and a self-contained software library to allow easy integration of PD3.0 and the older PD2.0 into projects.

The dev board is built around an ATMega32U4 microcontroller and FUSB302 USB-C PHY. The four-layer PCB is densely packed on both sides to fit in the Arduino Pro Micro Form factor. The board can deliver up to 100W (20 V at 5 A) from an appropriate power source and shows visual feedback on the PD status through a set of LEDs.

The primary goal of the project is actually in the software. [Ryan] found that existing software libraries for PD take up a lot of memory, and are difficult to integrate into small projects. Working from the PD specifications and PD PHY chip data sheet, he created a lighter weight and self-contained software library which consumes less than 8 K of flash and 1 K of RAM. This is less than half the Flash and RAM available on the ATmega32U4.

[Ryan] is running a Crowd Supply campaign (video after the break) to get some of these powerful boards out in the wild, and has released all the source code and schematics on GitHub. The PCB design files will be released during the last week of the campaign, around 25 January 2021.

USB-C and power delivery are not simple standards, but the ability to add a high-speed data interface and a programmable power supply into almost any project has real potential.

source https://hackaday.com/2021/01/16/usb-c-programmable-power-supply-for-any-project/

Considering one of the biggest draws of the original Etch a Sketch was how simple it was, it’s always interesting to see the incredible lengths folks will go to recreate that low-tech experience with modern hardware. A perfect example is this giant wall mounted rendition of the iconic art toy created by [Ben Bernstein]. With a Raspberry Pi and some custom electronics onboard, it can even do its own drawing while you sit back and watch.

At a high level, what we’re seeing here is a standard Samsung LCD TV with a 3D printed Etch a Sketch shell mounted on top of it. That alone would be a pretty neat project, and had [Ben] just thrown some videos of designs getting sketched out onto the display, he could have achieved a similar end result with a lot less work. But where’s the fun in that?

To make his jumbo Etch a Sketch functional, [Ben] spent more than a year developing the hardware and software necessary to read the user input from the two large 3D printed knobs mounted under the TV. The knobs are connected to stepper motors with custom PCBs mounted to their backs that hold a A4988 driver chip as well as a AS5600 absolute magnetic rotary encoder. This solution allows the Raspberry Pi to not only read the rotation of the knobs when a person is using the Etch a Sketch interactively, but spin them realistically when the software takes over and starts doing an autonomous drawing.

Several Python scripts pull all the various pieces of hardware together and produce the final user interface. The software [Ben] wrote can take an image and generate paths that the Etch a Sketch can use to realistically draw it. The points that the line is to pass through, as well as variables that control knob rotation and pointer speed, are saved into a JSON file so they can easily be loaded later. Towards the end of the Imgur gallery [Ben] has created for this project, you can see the software working its way through a few example sketches.

We’ve seen several projects that motorize an Etch a Sketch to draw complex images, but this may be the first example we’ve seen where everything was done in software. This digital version doesn’t need to follow the traditional “rules”, but we appreciate that [Ben] stuck to them anyway. Incidentally this isn’t the first Etch a Sketch TV conversion to grace these pages, though to be fair, the other project took a radically different approach.

View post on imgur.com

source https://hackaday.com/2021/01/16/tv-turned-automatic-etch-a-sketch-with-raspberry-pi/

Considering one of the biggest draws of the original Etch a Sketch was how simple it was, it’s always interesting to see the incredible lengths folks will go to recreate that low-tech experience with modern hardware. A perfect example is this giant wall mounted rendition of the iconic art toy created by [Ben Bernstein]. With a Raspberry Pi and some custom electronics onboard, it can even do its own drawing while you sit back and watch.

At a high level, what we’re seeing here is a standard Samsung LCD TV with a 3D printed Etch a Sketch shell mounted on top of it. That alone would be a pretty neat project, and had [Ben] just thrown some videos of designs getting sketched out onto the display, he could have achieved a similar end result with a lot less work. But where’s the fun in that?

To make his jumbo Etch a Sketch functional, [Ben] spent more than a year developing the hardware and software necessary to read the user input from the two large 3D printed knobs mounted under the TV. The knobs are connected to stepper motors with custom PCBs mounted to their backs that hold a A4988 driver chip as well as a AS5600 absolute magnetic rotary encoder. This solution allows the Raspberry Pi to not only read the rotation of the knobs when a person is using the Etch a Sketch interactively, but spin them realistically when the software takes over and starts doing an autonomous drawing.

Several Python scripts pull all the various pieces of hardware together and produce the final user interface. The software [Ben] wrote can take an image and generate paths that the Etch a Sketch can use to realistically draw it. The points that the line is to pass through, as well as variables that control knob rotation and pointer speed, are saved into a JSON file so they can easily be loaded later. Towards the end of the Imgur gallery [Ben] has created for this project, you can see the software working its way through a few example sketches.

We’ve seen several projects that motorize an Etch a Sketch to draw complex images, but this may be the first example we’ve seen where everything was done in software. This digital version doesn’t need to follow the traditional “rules”, but we appreciate that [Ben] stuck to them anyway. Incidentally this isn’t the first Etch a Sketch TV conversion to grace these pages, though to be fair, the other project took a radically different approach.

View post on imgur.com

source https://hackaday.com/2021/01/16/tv-turned-automatic-etch-a-sketch-with-raspberry-pi/

Software defined radio lets RF hardware take on a broad spectrum of tasks, all based on how that hardware is utilized in code. The bladeRF 2.0 micro xA9 is one such device, packing a fat FPGA with plenty of room for signal processing chains on board. As a demonstration of its abilities, [Robert Ghilduta] set about writing a software-defined WiFi implementation for the platform.

The work is known as bladeRF-wiphy, as it implements the PHY, or physical layer of the WiFi connection, in the 7-layer OSI networking model. Modulation and demodulation of the WiFi signal is all handled onboard the Cyclone V FPGA, with the decoded 802.11 WiFI packets handed over to the Linux mac80211 module which handles the MAC level, or medium access control. Thanks to the capability baked into mac80211, the system can act as either an access point or an individual station depending on the task at hand.

[Robert] does a great job of explaining the why and the how of implementing WiFi modulation on an FPGA, as well as some basics of modem development in both software and hardware. It’s dense stuff, so for those new to the field of software defined radio, consider taking some classes to get yourself up to speed!

source https://hackaday.com/2021/01/16/doing-wifi-with-software-defined-radio/

Back in the 20th century, Lego Technic was a popular toy designed to teach kids about mechanical technology, and be a lot of fun to boot. Motors and pneumatics were available, but by and large you had to move your creations and make the noises yourself. That’s not the case these days, as the [Brick Experiment Channel] demonstrates with this impressive Lego tank.

The drivetrain is straightforward, using standard Lego tank treads with each side given its own motor for easy skid steering. However, the real party piece is the slingshot cannon, which launches Lego soccer balls at 60 km/h. Utilizing several motors, it’s complete with elevation adjustment for accurate ranging, and a 6 round magazine so you can (slowly) prosecute your targets with rounds downrange.

What really makes this build great is the control system, with the tank being controlled by a PS4 controller via Sbrick, a device that lets Lego motors be controlled via Bluetooth. We’d love to build a couple of Lego vehicles and have them blast away at each other. We’ve seen the technology used before for a secret heist robot. Video after the break.

source https://hackaday.com/2021/01/15/lego-tank-fires-soccer-ball-cannon/

If there’s anything about amateur radio that has more witchcraft in it than the design and implementation of antennas, we don’t know what it would be. On the face of it, hanging out a chunk of wire doesn’t seem like it should be complicated, but when you dive into the details, building effective antennas and matching them to the job at hand can be pretty complex.

That doesn’t mean antenna topics have to remain a total mystery, of course, especially once someone takes the time to explain things properly. [Charlie Morris (ZL2CTM)] recently did this with a simple antenna tuner, a device used to match impedances between a transmitter and an antenna. As he explains in the first video below, his tuner design is really just a Wheatstone bridge where the antenna forms half of one leg. A toroidal transformer with multiple taps and a variable capacitor forms an LC circuit that matches the high impedance antenna, in this case a multi-band end-fed halfwave, with the nominal 50-ohm load expected by the transceiver. A small meter and a diode detector indicate when the bridge is balanced, which means the transceiver is seeing the proper load.

The second video below shows the final implementation of the tuner; as a fan of QRP, or low-power operation, [Charlie] favors simple, lightweight homebrew gear that can be easily taken into the field, and this certainly fits the bill. A final video shows the tuner in use in the field, with a NanoVNA proving what it can do. As usual, [Charlie] protests that he not an expert and that he’s just documenting what he did, but he always does such a good job of presenting the calculations involved in component selection that any ham should be able to replicate his builds.

source https://hackaday.com/2021/01/15/manual-antenna-tuner-shows-how-homebrewing-is-done/

Rocket League is a video game famous for being wildly popular despite being virtually unplayable without several hours practice. It involves hyper fast cars playing soccer, complete with the ability to flip, jump, and rocket boost into the ball. [mrak_ripple] decided he wanted some of that action in a real RC car, and set to work.

While rocket boosts were out of scope for this build, [mrak_ripple] was pretty confident he could build a jumping, flipping RC car modelled after the Rocket League Octane vehicle. Initial experiments involved a custom 3D printed spring mechanism, but the results were underwhelming. Instead, in the true hacker spirit, a jumping mechanism was taken from an existing toy, and installed in the car instead. This was combined with a mechanism built out of a brushless motor with a flywheel added to generate a flipping moment in mid-air.

The final result is impressive, with the car flipping relatively cleanly once refined and lightened from its original design. We’d love to see a two-axis build that can front- and back-flip as well. It’s a step up in complexity from the last build we saw from [mrak_ripple], the amusing mashed potato trebuchet. Video after the break.

source https://hackaday.com/2021/01/15/youre-going-to-flip-out-over-this-rocket-league-rc-car/

Rocket League is a video game famous for being wildly popular despite being virtually unplayable without several hours practice. It involves hyper fast cars playing soccer, complete with the ability to flip, jump, and rocket boost into the ball. [mrak_ripple] decided he wanted some of that action in a real RC car, and set to work.

While rocket boosts were out of scope for this build, [mrak_ripple] was pretty confident he could build a jumping, flipping RC car modelled after the Rocket League Octane vehicle. Initial experiments involved a custom 3D printed spring mechanism, but the results were underwhelming. Instead, in the true hacker spirit, a jumping mechanism was taken from an existing toy, and installed in the car instead. This was combined with a mechanism built out of a brushless motor with a flywheel added to generate a flipping moment in mid-air.

The final result is impressive, with the car flipping relatively cleanly once refined and lightened from its original design. We’d love to see a two-axis build that can front- and back-flip as well. It’s a step up in complexity from the last build we saw from [mrak_ripple], the amusing mashed potato trebuchet. Video after the break.

source https://hackaday.com/2021/01/15/youre-going-to-flip-out-over-this-rocket-league-rc-car/

Rocket League is a video game famous for being wildly popular despite being virtually unplayable without several hours practice. It involves hyper fast cars playing soccer, complete with the ability to flip, jump, and rocket boost into the ball. [mrak_ripple] decided he wanted some of that action in a real RC car, and set to work.

While rocket boosts were out of scope for this build, [mrak_ripple] was pretty confident he could build a jumping, flipping RC car modelled after the Rocket League Octane vehicle. Initial experiments involved a custom 3D printed spring mechanism, but the results were underwhelming. Instead, in the true hacker spirit, a jumping mechanism was taken from an existing toy, and installed in the car instead. This was combined with a mechanism built out of a brushless motor with a flywheel added to generate a flipping moment in mid-air.

The final result is impressive, with the car flipping relatively cleanly once refined and lightened from its original design. We’d love to see a two-axis build that can front- and back-flip as well. It’s a step up in complexity from the last build we saw from [mrak_ripple], the amusing mashed potato trebuchet. Video after the break.

source https://hackaday.com/2021/01/15/youre-going-to-flip-out-over-this-rocket-league-rc-car/

By now, the wearing of a facemask to protect ourselves from pandemic infection is for many of us a daily fact of life. Perhaps that means a cheap disposable mask, but there’s no reason that has to be the case. It’s easy to make more durable masks that can be washed and re-used time and time again, and our Hackaday colleague [Kristina Panos] has shared her pattern and workflow to help you do it.

Her pattern isn’t a complex cut-out but a simple rectangle, and the trick of sewing them together and flipping them inside out makes for a very tidy result. With three pleats pressed in and the elastic sewn up the result is a mask that’s neat, attractive, effective, and cheap, which is a win in our book.

It’s worth repeating her important point that these are not for use in medical environments, instead they’re the standard street-wear aerosol catchers we’re all used to.This isn’t the first time we’ve looked at masks here at Hackaday, or indeed though [Kristana]’s are by far the tidier neither is it first time one of us has made a mask. We looked at them in depth last year in our surviving the pandemic as a hacker series.

source https://hackaday.com/2021/01/15/a-surefire-way-to-make-masks/

Old Civil Defense survey meters like the V-715 are interesting conversation starters, but of very little practical use today. These devices were intended to be a sort of litmus test that survivors of a nuclear blast could use to determine when it was safe to venture out of their radiation shelter: if the needle on the meter moves, even when it’s on the most sensitive setting, you should probably go back inside. Since [Hamilton Karl] would (hopefully) never need such an indicator, he decided to have a little fun with this Cold War holdover and turn it into a Disco Containment Unit.

Technical details are a little sparse on this one, but we can infer most of it just from the pictures. In place of the original meter [Hamilton] has mounted a tiny mirrored ball inside of a protective cage, which is spun by a geared motor that’s occupying the space that used to be taken up by the ion chamber.

A handful of Adafruit NeoPixel RGB LEDs, an Arduino Nano, and a few switches to control it all round out the functional aspects of the build, and a new disco-themed trefoil replaces the original Civil Defense logo on the side. The project page mentions there’s a piezo buzzer onboard that performs a stirring rendition of “Stayin’ Alive” by the Bee Gees, but alas there’s no video that shows it in action.

Thanks to the rugged construction and built-in handle of these old survey meters, [Hamilton] can now take the party with him wherever he goes. Not that he can really go anywhere with this whole global pandemic hanging over our heads, but at least he’ll be ready when things start trending towards normal. In a way the device’s functionality has now been reversed from how it originally worked, since the meter going wild will now be an indicator that its safe to come out.

While the V-715 isn’t of much use outside of a post-apocalyptic hellscape, the V-700 is actually a proper Geiger counter that’s still useful for surveying or research. An important distinction to remember if you ever get a chance to snap one of them up at a swap meet or flea market. Whenever we can start having those again, anyway.

source https://hackaday.com/2021/01/15/civil-defense-disco-ball-rocks-ground-zero/

You may have been one of the many of us who received an email from Ubiquiti this week, recommending a password change. The email stated that there was an unauthorized access of Ubiquiti systems, and while there wasn’t evidence of user data being accessed, there was also not enough evidence to say emphatically that user data was not accessed. Ubiquiti has mentioned that the database that may have been accessed contains a user’s name, email address, hashed password, and optionally the mailing address and phone number.

Depending on how the Ubiquiti authentication system is designed, that hashed password may be enough to log in to someone’s account. In any case, updating your password would invalidate the potentially compromised hash. This event underscores a complaint voiced by Ubiquiti users: Ubiquiti has been making it harder to avoid administrating hardware with local only accounts.

Nissan Source Code

Nissan was hosting a large git repository using Atlassian’s Bitbucket. That install was still using default credentials for the admin account, and someone finally noticed. The researcher who first discovered the issue has remained anonymous, and the primary source for the linked article was caught up in the recent outbreak of Twitter censorship, with an account suspension.

The repository contained code from Nissan mobile apps, marketing information, and code for internal-only services. The 18.4 GB data dump is still available on the darker corners of the internet, via torrent files.

Zyxel Scans Seen by ISC

Remember the Zyxel problem we talked about last week? Well this didn’t take long. The Internet Storm Center (ISC) is reporting that it is already seeing SSH login attempts using those hard-coded credentials.

It’s worth taking a minute to call out the ISC and similar efforts for their invaluable work. The ISC primarily serves as a clearinghouse for data from Intrusion Detection Systems and firewalls around the internet. When new patterns emerge, volunteers watching the data can quickly identify new attacks as they emerge. In some cases, this quick response can give administrators around the world time to patch the vulnerability being targeted before they are compromised.

Dovecot Hibernation

Dovecot has released version 2.3.13, and there is a fix for a notable vulnerability, CVE-2020-24386. IMAP supports an IDLE command, putting the connection to the server in a holding pattern, ready to push real-time mail notifications to the client. The vulnerability allows a client to put its connection in this state, and then send a malicious request to the server. This request can allow for limited filesystem access, most notably the downloading of messages from other accounts. It’s possible to mitigate the flaw through disabling IMAP hibernation, but the recommendation is to simply update to the latest release.

Telegram Triangulation

Telegram is one of the go-to solutions for sending secure messages. Just over a year ago, Telegram introduced “People Near Me”, a feature for finding nearby users who have opted in to the service. If you’ve opted in, you might consider going and turning that feature off. Telegram gives a very precise and accurate distance to anyone else who is within seven miles. That distance updates in real time, which is great for meetups. What might not be immediately obvious is that it’s rather trivial to spoof a device’s location to anywhere in the world. Within a few minutes, it’s possible to precisely locate anyone in the world who has Telegram’s location service turned on. Other services have prevented this problem by giving less precise location data. So far, Telegram has responded that this is not a bug, and it doesn’t plan to make any changes.

How Solarwinds Got Hacked

More details on the Solarwinds backdoor is slowly coming to light. The more information is revealed, the more interesting the story becomes. This week, we got Crowdstrike’s write-up of the malware running on Solarwinds machines. This malware, dubbed Sunspot, isn’t the Orion backdoor itself, it is a custom-written piece of malware that modifies source code surreptitiously at compile time. This brings to mind the old Trusting Trust attack.

Sunspot was written to very carefully hide from detection, and to only take action when it detects code compiling. It checks once a second for  MsBuild.exe, and whether it was building Orion. If it is, it modifies one source code file, waits for compilation to complete, and then undoes the malicious change. A developer would be hard pressed to discover the modification, because it only exists during compilation, while the developer is out getting coffee anyway. We were somewhat skeptical when Solarwinds first called this a “sophisticated and novel” hack, but the evidence seems to affirm that opinion.

source https://hackaday.com/2021/01/15/this-week-in-security-ubiquiti-nissan-zyxel-and-dovecot/

A recent research paper shows a way to create multicolor 3D prints using a single extruder if you are too lazy to babysit the machine and switch filament. The concept: print your own “programmable” filament that has the right colors in the right place. This is the same idea as manually splicing filament but presumably is more efficient since the process works with one color at a time and doesn’t repeat. In other words, to print the 64 squares of a chessboard you’d swap filament at least 64 times on each layer. Using programmable filament, you’d load one spool, print half of the filament, load another spool, print the other half, and then finally load the newly created filament and print the chessboard. Notice that the first two operations aren’t printing the chessboard. They are printing the spool of filament you feed through on the third pass.

There are machines made to do this, of course, although they generally just splice lengths of filament together for you automatically. Using one filament solves the problems of keeping multiple heads in alignment as well as the added cost and complexity. However, you now have different problems such as the transition between materials and knowing exactly how much material will be at each point in the print.

We wondered how much filament you can print spooled up on a reasonable size print bed. According to the paper:

The maximum length of a printed filament is highly dependent on the build area given the type of 3D printer. With a 300×300 mm bed size, a printed filament in a spiral shape can be approximately 20-m long.

Printing the working filament is time-consuming, too. The paper reports that a boat that took two hours to print required five hours to print the programmable filament. Still better than swapping a bunch of plastic by hand, we suppose.

Oddly enough, the inspiration for this paper came from an Instructable. If you plan to try this, you might want to watch the videos there about printing hexagonal filaments with different colors.

If the idea of printing filament seems strange, we’ve seen it before. Of course, you don’t need any equipment at all if you are patient and willing to lie about it.

source https://hackaday.com/2021/01/15/programmable-filament-for-multicolor-printing/

With the aim of reducing virus transmission due to gatherings during the pandemic, the Dutch government have banned fireworks. The people of the Netherlands like their noisy things so we’re told that the ban has been widely flouted, but [Build Comics] are a law-abiding group of workshop tool heroes. For their lockdown noise, they created an entirely-legal pulsejet. The interesting part is that it was made entirely using fairly basic tools on a minimalist budget, with TIG and MIG eschewed in favour of a mundane stick welder.

The form of the pulse jet will probably be familiar as it has been taken from other published designs. A long tube is bent back upon itself with a combustion chamber placed in one of its arms such that the jet forms a resonant chamber that produces continuous pulses of exhaust gas. This one is made from stainless steel tube, and the exhaustive documentation should be worth a look for anyone tempted to make their own. Welding thin sheet with a stick welder requires quite a bit of skill, and in a few places they manage to burn a hole or two. One requires a patch, but the time-honoured technique of running a bead around the edge manages to successfully close another.

Their first attempt to fire it up using a leaf blower with a 3D-printed adapter fails, but following the construction of a more resilient part and a more efficient gas injector the engine starts. It’s then taken out on a farm for some serious noise without too many angry neighbours, as you can see in the video below the break.

The hero tools of Build Comics have appeared here before, most recently with an analogue meter clock.

source https://hackaday.com/2021/01/15/a-136-euro-pulse-jet-for-some-no-firework-lockdown-fun/

While most PCBs stick to tried-and-true methods of passing electrons through their layers of carefully-etched copper, modern construction methods allow for a large degree of customization of most aspects of these boards. From solder mask to number of layers, and even the shape of the board itself, everything is open for artistic license and experimentation now. [Luca] shows off some of these features with his PCB which acts as a live map of Italy.

The PCB is cut out in the shape of the famous boot, with an LED strategically placed in each of 20 regions in the country. This turns the PCB into a map with the RGB LEDs having the ability to be programmed to show any data that one might want. It’s powered by a Wemos D1 Mini (based on an ESP8266) which makes programming it straightforward. [Luca] has some sample programs which fetch live data from various sources, with it currently gathering daily COVID infection rates reported for each of the 20 regions.

The ability to turn a seemingly boring way to easily attach electronic parts together into a work of art without needing too much specialized equipment is a fantastic development in PCBs. We’ve seen them turned into full-color art installations with all the mask colors available, too, so the possibilities for interesting-looking (as well as interesting-behaving) circuits are really opening up.

source https://hackaday.com/2021/01/14/dynamic-map-of-italy-on-a-pcb/

Between hot things, sharp things, and spinny things, there’s more than enough danger in the average hacker’s shop to maim and mutilate anyone who fails to respect their power. But somehow lasers don’t seem to earn the same healthy fear, which is strange considering permanent blindness can await those who make a mistake lasting mere fractions of a second.

To avoid that painful fate, high-power laser fan [Brainiac75] undertook building a beam dump, which is a safe place to aim a laser beam in an experimental setup. His version has but a few simple parts: a section of extruded aluminum tubing, a couple of plastic end caps, and a conical metal plumb bob. The plumb bob gets mounted to one of the end caps so that its tip points directly at a hole drilled in the center of the other end cap. The inside and the outside of the tube and the plumb bob are painted with high-temperature matte black paint before everything is buttoned up.

In use, laser light entering the hole in the beam dump is reflected off the surface of the plumb bob and absorbed by the aluminum walls. [Brainiac75] tested this with lasers of various powers and wavelengths, and the beam dump did a great job of safely catching the beam. His experiments are now much cleaner with all that scattered laser light contained, and the work area is much safer. Goggles still required, of course.

Hats off to [Brainiac75] for an instructive video and a build that’s cheap and easy enough that nobody using lasers has any excuse for not having a beam dump. Such a thing would be a great addition to the safety tips in [Joshua Vasquez]’s guide to designing a safe laser cutter.

source https://hackaday.com/2021/01/14/beam-dump-makes-sure-your-laser-path-is-safely-terminated/

Soil moisture sensors are cheap and easy to interface with, to the point that combining one with an Arduino and blinking an LED when your potted plant is feeling a bit parched is a common beginners project. But what about on the long term? Outside of a simple proof of concept, what would it take to actually read the data from these sensors over the course of weeks or months?

That’s precisely the question [derflob] recently had to answer. The goal was to build a device that could poll multiple soil sensors and push the data wirelessly into Home Assistant. But since it would be outside on the balcony, it needed to run exclusively on battery power. Luckily his chosen platform, the ESP32, has some phenomenal power saving features. You just need to know how to use them.

In his extremely detailed blog post, [derflob] goes over some of the tricks he used to eek out as much runtime as possible. The big one, and arguably the key to the whole project, is his use of the ESP32’s Ultra-Low Power coprocessor (ULP). This powerful feature allows the microcontroller to do some useful work, albeit with a reduced instruction set, without waking up the power-hungry primary processor.

Since the ULP can read from the ADC, he was able to craft some assembly code that only wakes the CPU if the readings from one of the six connected sensors has changed more than a set threshold value since the last update. This keeps the chip asleep the majority of the time, and so far, [derflob] says his board has been humming along for six months on a single LiFePO4 18650 cell. If you’re not so keen on diving into assembly, we’ve seen a similar technique accomplished with a simple circuit that only wakes the microcontroller when there’s work to be done.

So where does the project go from here? Now that the PCB is largely finalized, [derflob] is turning his attention to waterproofing both it and the sensors so they can survive outside in the elements. Some kind of epoxy encapsulation is probably in order, though that will be a project for another day.

source https://hackaday.com/2021/01/14/esp32-soil-monitors-tap-into-ultra-low-power-mode/

Soil moisture sensors are cheap and easy to interface with, to the point that combining one with an Arduino and blinking an LED when your potted plant is feeling a bit parched is a common beginners project. But what about on the long term? Outside of a simple proof of concept, what would it take to actually read the data from these sensors over the course of weeks or months?

That’s precisely the question [derflob] recently had to answer. The goal was to build a device that could poll multiple soil sensors and push the data wirelessly into Home Assistant. But since it would be outside on the balcony, it needed to run exclusively on battery power. Luckily his chosen platform, the ESP32, has some phenomenal power saving features. You just need to know how to use them.

In his extremely detailed blog post, [derflob] goes over some of the tricks he used to eek out as much runtime as possible. The big one, and arguably the key to the whole project, is his use of the ESP32’s Ultra-Low Power coprocessor (ULP). This powerful feature allows the microcontroller to do some useful work, albeit with a reduced instruction set, without waking up the power-hungry primary processor.

Since the ULP can read from the ADC, he was able to craft some assembly code that only wakes the CPU if the readings from one of the six connected sensors has changed more than a set threshold value since the last update. This keeps the chip asleep the majority of the time, and so far, [derflob] says his board has been humming along for six months on a single LiFePO4 18650 cell. If you’re not so keen on diving into assembly, we’ve seen a similar technique accomplished with a simple circuit that only wakes the microcontroller when there’s work to be done.

So where does the project go from here? Now that the PCB is largely finalized, [derflob] is turning his attention to waterproofing both it and the sensors so they can survive outside in the elements. Some kind of epoxy encapsulation is probably in order, though that will be a project for another day.

source https://hackaday.com/2021/01/14/esp32-soil-monitors-tap-into-ultra-low-power-mode/