Over the last few years, LED candles have become increasingly common; and for good reason. From a distance a decent LED candle is a pretty convincing facsimile for the real thing, providing a low flickering glow without that annoying risk of burning your house down. But there’s something to be said for the experience of a real candle; such as that puff of fragrant smoke you get when you blow one out.
Which is why [Keith] set out on an epic three year quest to build the most realistic LED candle possible, with a specific focus on the features that commercial offerings lack. So not only does it use real wax as a diffuser for the LEDs, but you’re able to “light” it with an actual match. It even ejects a realistic bit of smoke when its microphone detects you’ve blown into it. Ironically, its ability to generate smoke means it doesn’t completely remove the possibility of it setting your house on fire if left unattended, but we suppose that’s the price you pay for authenticity.
As you might have gathered by now, [Keith] is pretty serious about this stuff, and has gone to great lengths to document his candle’s long development process. If you’d care to build a similar candle, his written documentation as well as the video after the break will certainly get you on the right track. He’s even broken the design down into “milestones” of increasing complexity, so for example if you don’t care about the smoking aspect of the candle you can just skip that part of the build.
So what did [Keith] put into his ultimate LED candle? In the most basic form, the electronics consist of a Arduino Pro Mini and a chunk of RGB WS2812B strip holding six LEDs. Add in an IR sensor if you want the candle to be able to detect the presence of a match, and a microphone if you want to be able to blow into the candle to turn it off. Things only get tricky if you want to go full smoke, and let’s be honest, you want to go full smoke.
To safely produce a puff of fragrant smoke, [Keith] is using a coil of 28 gauge wire wrapped around the wick of a “Tiki Torch”, and a beefy enough power supply and MOSFET to get it nice and hot. The wick is injected with his own blend of vegetable glycerin and aromatic oil, and when the coil is fired up it produces an impressive amount of light gray smoke that carries the scent of whatever oil you add. Even if you’re not currently on the hunt for the ultimate electronic candle, it’s a neat little implementation that could be used come Halloween.
You might be surprised to learn that LED candles are a rather popular project within the hacking community. From the exceptionally simple to the exceedingly complex, we’ve seen an impressive array of electronic candles over the years. Perfect for setting the mood when listening to the smooth sounds of the latest Hackaday podcast episode. Continue reading “This Super Realistic LED Candle Is Smoking Hot”
Last week , I covered the ridiculously low barriers to entry to amateur radio, both in terms of financial outlay and the process of studying for and passing the FCC examination. You’ve had seven days, so I assume that you’ve taken the plunge and are a freshly minted amateur radio operator. The next big question may be: Now what?
We briefly mentioned the image that ham radio is a rich old person’s hobby, and that reputation is somewhat deserved. For ham gear, there really is no upper limit on what you can spend. Glossy brochures and slick web pages hawk transceiver bristling with knobs and switches and loaded with the latest features, all of which will probably be obsolete within a few years when the Next Big Thing comes along and manufacturers respond with new, must-have models – looking at you, ICOM IC-7300. It’s no different than any other technology market, and enough people fall for that marketing to make it a going concern.
But thankfully, while there is no apparent ceiling on what you can spend on ham gear, there certainly is a floor, and it can be very, very low. Our $50 budget can go quite a long way to getting a new Technician on the air, if you’re willing to make some compromises and can forego the latest and greatest for a while.
We see a lot of weird and esoteric stuff here at Hackaday, but even by our standards, Bell Lab’s Plan 9 operating system is an oddball. Named after the science fiction film Plan 9 from Outer Space, it was designed to extend the UNIX “everything is a file” mentality to the network. It envisioned a future where utilizing the resources of another computer would be as easy as copying a file. But as desktop computers got more powerful the idea seemed less appealing, and ultimately traditional operating systems won out. Of course, that doesn’t mean you still can’t play around with it.
Logically to make use of a distributed operating system you really need something to distribute it on, but as [Andrew Back] shows, today that’s not nearly the challenge it would have been back then. Using the Raspberry Pi, he builds a four-node Plan 9 cluster that’s not only an excellent way to explore this experimental operating system, but looks cool sitting on your desk. Even if you’re not interested in drinking the Bell Lab’s Kool-Aid circa 1992, his slick desktop cluster design would work just as well for getting your feet wet with modern-day distributed software stacks.
The enclosure for the cluster is built from laser cut acrylic panels which are then folded into shape with a hot wire bending machine. That might seem like a tall order for the home hacker, but we’ve covered DIY acrylic benders in the past, and the process is surprisingly simple. Granted you’ll still need to get access to a beefy laser cutter, but that’s not too hard anymore if you’ve got a hackerspace nearby.
[Andrew] uses short extension cables and female panel mount connectors to keep everything tidy, and with the addition of some internal LED lighting the final product really does look like a desktop computer from a far more fashionable future. Combined with the minimalist keyboard, the whole setup wouldn’t look out of place on the set of a science fiction movie. Perhaps that’s fitting, giving Bell Lab’s futuristic goals for Plan 9.
Its been the better part of a decade since we first brought you word that Plan 9 was available for the Raspberry Pi, and yet in all that time we’ve never really seen it put to use. Hopefully builds like this will inspire others to play around with this fascinating piece of computing history.
[Thanks to Dave for the tip.]
Reinforcement learning has been a hot-button area of research into artificial intelligence. This is a method where software agents make decisions and refine these over time based on analyzing resulting outcomes. [Kamil Rocki] had been exploring this field, but needed some more powerful tools. As it turned out, a cluster of emulated Game Boys running at a billion FPS was just the ticket.
The trick to efficient development of reinforcement learning systems is to be able to run things quickly. If it takes an AI one thousand attempts to clear level 1 of Super Mario Bros., you’d better hope you’re not running that in real time. [Kamil] started by coding a Game Boy emulator in C. By then implementing it in Verilog, [Kamil] was able to create a cluster of emulated Game Boys that enabled games to be run at breakneck speed, greatly speeding the training and development process.
[Kamil] goes into detail about how the work came to revolve around the Game Boy platform. After initial work with the Atari 2600, which is somewhat of a defacto standard in RL circles, [Kamil] began to explore further. It was desired to have an environment with a well-documented CPU, a simple display to cut down on the preprocessing required, and a wide selection of games.
The goal of the project is to allow [Kamil] to explore the transfer of knowledge from one game to another in RL systems. The aim is to determine whether for an AI, skills at Metroid can help in Prince of Persia, for example. This is arguably true for human players, but it remains to be seen if this can be carried over for RL systems.
It’s rather advanced work, on both a hardware emulation level and in terms of AI research. Similar work has been done, training a computer to play Super Mario through monitoring score and world values. We can’t wait to see where this research leads in years to come.