Harmonic Drive Uses Compliant Mechanism To Slim Down

April 7, 2021 by Dan Maloney 15 Comments

[Levi Janssen] has a secret: he doesn’t like harmonic drives. But rather than abandon the torque-amplifying transmission completely, he decided to see about improving them using 3D-printed compliant mechanisms.

For the uninitiated, harmonic drives, also known as strain-wave gears, are a compact, high-torque gearbox that has become popular with “robotic dog” makers and other roboticists. The idea is to have a rigid, internally-toothed outer ring nested around an externally-toothed, flexible cup. A wave generator rotates within the inside cup, stretching it so that it meshes with the outer ring. The two gears differ by only a couple of teeth, meaning that very high gear ratios can be achieved, which makes them great for the joints of robot legs.

[Levi]’s problem with the harmonic drive is that due to the depth of the flexible spline cup, compactness is not among its virtues. His idea is to couple the flex spline to the output of the drive through a flat spring, one that allows flexion as the wave generator rotates but transmits torque efficiently. The entire prototype is 3D-printed, except for the wave generator bearings and stepper motor, and put to the test.

As the video below shows after the excellent introduction to harmonic drives, the concept works, but it’s not without its limitations. Even lightly loaded, the drive made some unpleasant crunching sounds as the PLA springs gave out. We could easily see that being replaced with, say, a steel spring, either machined or cut on a water-jet machine. That might solve the most obvious problem and make [Levi]’s dream of a compact harmonic drive a reality. Of course, we have seen pretty compact strain-wave gears before.

Continue reading “Harmonic Drive Uses Compliant Mechanism To Slim Down” →

This Custom Dynamometer Is A Stirling Example Of Homebrewing

April 6, 2021 by Dan Maloney 25 Comments

[Leo Fernekes] has fallen down the Stirling engine rabbit hole. We mustn’t judge — things like this happen in the best of families, after all. And when they do happen to someone like [Leo], things can get interesting mighty quickly.

His current video, linked below, actually has precious little to do with his newfound Stirling engine habit per se. But when you build a Stirling engine, and you’re of a quantitative bent, having some way to measure its power output would be handy. That’s a job for a dynamometer, which [Leo] sets out to build in grand fashion. Dynos need to measure the torque and rotational speed of an engine while varying the load on it, and this one does it with style.

[Leo]’s torque transducer is completely DIY, consisting of hand-wound coils on the ends of a long lever arm that’s attached to the output shaft of the engine under test by a magnetic coupling. The coils are free to move within a strong magnetic field, with a PID loop controlling the current in the coils. Feedback on the arm’s position is provided by an optical sensor, also DIY, making the current necessary to keep the arm stationary proportional to the input torque. The video goes into great detail and has a lot of design and build tips.

We just love the whole vibe of this build. There may have been simpler or quicker ways to go about it, but [Leo] got this done with what he had on hand for a fraction of what buying in off-the-shelf parts would have cost. And the whole thing was a great learning experience, both for him and for us. It sort of reminds us of a dyno that [Jeremy Fielding] built a while back, albeit on a much different scale.

Continue reading “This Custom Dynamometer Is A Stirling Example Of Homebrewing” →

Flight Of The Pterothopter: A Jurassic-Inspired Ornithopter

April 6, 2021 by Zach Zeman 14 Comments

Ornithopters look silly. They look like something that shouldn’t work. An airplane with no propeller and wings that go flappy-flappy? No way that thing is going to fly. There are, however, a multitude of hobbyists, researchers, and birds who would heartily disagree with that sentiment, because ornithopters do fly. And they are almost mesmerizing to watch when they do it, which is just one reason we love [Hobi Cerdas]’s build of the Pterothopter, a rubber band-powered ornithopter modeled after a pterodactyl.

All joking aside, the science and research behind ornithopters and, relatedly, how living organisms fly is fascinating in itself — which is why [Lewin Day] wrote that article about how bees manage to become airborne. We can lose hours reading about this stuff and watching videos of prototypes. While most models we can currently build are not as efficient as their propeller-powered counterparts, the potential of evolutionarily-perfected flying mechanisms is endlessly intriguing. That alone is enough to fuel builds like this for years to come.

As you can see in the video below, [Hobi Cerdas] went through his own research and development process as he got his Pterothopter to soar. The model proved too nose-heavy in its maiden flight, but that’s nothing a little raising of the tail section and a quick field decapitation couldn’t resolve. After a more successful second flight, he swapped in a thinner rubber band and modified the wing’s leading edge for more thrust. This allowed the tiny balsa dinosaur to really take off, flying long enough to have some very close encounters with buildings and trees.

For those of you already itching to build your own Pterothopter, the plans come from the Summer 2017 issue of Flapping Wings, the official newsletter of the Ornithopter Society (an organization we’re so happy to learn about today). You can also find more in-depth ornithopter build logs to help you get started. And, honestly, there’s no reason to limit yourself to uncontrolled flight; we’ve come across some very impressive RC ornithopters in the past.

Continue reading “Flight Of The Pterothopter: A Jurassic-Inspired Ornithopter” →

Decoding S/PDIF With A Microcontroller Brings A Few Headaches

April 4, 2021 by Lewin Day 14 Comments

The average punter shunts audio around with analog 3.5 mm cables, RCA jacks, or Bluetooth on a regular basis. A useful standard that hasn’t really bothered most of us is S/PDIF, standing for Sony/Phillips Digital Interface. It’s a useful way to pump digital audio around over copper cables or optic fiber. [Andrew Jeddeloh] got curious about the standard after contemplating some long cable runs in his home, and decided to try decoding it.

The target for [Andrew]’s development efforts was the STM32L476 Discovery, which had no SPDIF decoding hardware on board. Instead, [Andrew] tinkered with the peripherals he had to see what would work. In the end, a cavalcade of internal timers were daisy chained to allow the microcontroller to recover a clock from the self-clocked S/PDIF signal. This was then used to generate a clock to sync up the onboard SPI hardware to actually read in the 16-bit PCM data from the S/PDIF signal.

[Andrew]’s original broader plan was to pipe the S/PDIF data to the onboard I2S DAC, though he struggled manipulating the remaining resources on the STM chip to do so successfully. Anyone wishing to have a crack can take a look at [Andrew]’s code over on GitHub. If completed, the STM32L476 would become a useful analog endpoint for S/PDIF streams, allowing you to pump tunes digitally over long distances without signal degradation. If you know the key to getting it done, sound off in the comments! Alternatively, if you need to get up and running more quickly, the Teensy platform has you covered!

Finger Bend Is A Textile Flex Sensor You Can Sew At Home

April 2, 2021 by Lewin Day 2 Comments

So often, we use control devices for electronics that involve our fingers directly grasping, touching, or moving another object or surface. It’s less common for us to use interfaces that detect the motion of our bodies directly. Flex sensors are one way to do that, and it’s exactly what [WillpowerStudios] aims to do with Finger Bend.

The construction of the sensor is simple, using piezoresistive fabric which changes its resistance when deformed. By sewing this into a sheath that can be placed on the finger, and wiring it up with conductive threads, it can be used to detect the flexion of the wearer’s digits by sampling the resistance with an analog to digital converter on any garden variety microcontroller. Expanding the technique to a full hand is as simple as creating a Finger Bend per digit and wiring up each one to its own ADC channel. If you want to get really fancy, you could even scan through them at speed with a multiplexer.

It’s similar to the technology used in Nintendo’s infamous Power Glove, and while it’s never caught on in the mainstream, it may have applications yet. Video after the break

Continue reading “Finger Bend Is A Textile Flex Sensor You Can Sew At Home” →

Adjustable, Low-Impact Keeb Is About As Comfortable As It Gets

April 2, 2021 by Kristina Panos 19 Comments

What’s the coolest-looking way to ease the repetitive stress of typing without quitting altogether? Move nothing but your fingers, and move them as little as possible without any stretching or reaching. We’ve been fans of the weirdly wonderful DataHand keyboard since we first laid eyes on one, but [Ben Gruver] has actually been using these out-of-production keyboards for years as a daily driver. And what do we do when we love something scarce? Make our own, improved version like [Ben] has done, with the lalboard.

[Ben] has been using the lalboard for about two years now and has a laundry list of improvements for version two, a project we are proud to host over on IO. Many of the improvements are designed to make this massive undertaking a bit easier to print and put together. Version one uses copper tape traces, but [Ben] is working on a fab-able PCB that will use something other than a pair of Teensy 2.0s, and perhaps QMK firmware.

Something that won’t be changing is the fantastic optical key switch design that uses an IR LED and phototransistor to capture key presses, and tiny square magnets to return the key to the home position and deliver what we’re quite sure is a satisfying clack.

The absolute coolest part of this keyboard is that it’s so adjustable. Every key cluster can be adjusted in 6 directions, which includes the ability to dial in different heights for each finger if that’s what works best. Once that’s all figured out, then it’s time to print some perfect permanent standoffs. Want to make one of these sci-fi clackers for yourself? [Ben] has the BOM, some printing instructions and tips, and a guide to making the copper tape PCBs over on GitHub. Check it out in action after the break as [Ben] rewrites Kafka’s Metamorphosis at 120 WPM.

Interested in learning more about the original DataHand keyboard? Here’s our take.

Continue reading “Adjustable, Low-Impact Keeb Is About As Comfortable As It Gets” →

Flashpen Is A High Fidelity Pen Input Device

April 1, 2021 by Lewin Day 10 Comments

Pen input has never really taken off in the computing mainstream, though it’s had somewhat of a renaissance in the last decade or so. Various smartphones and tablets are shipping with the technology, and some diehard users swear by it as the best way to take notes on the go. Recently, researchers at the Sensing, Interaction and Perception Lab at ETH Zurich have been working on Flashpen, a high-fidelity pen interface for a wide range of applications.

The fundamental technology behind the pen is simple, with the device using an optical flow sensor harvested from a high-end gaming mouse. This is a device that uses an image sensor to detect the motion of the sensor itself across a surface. Working at an update rate of 8 KHz, it eclipses other devices in the market from manufacturers such as Wacom that typically operate at rates closer to 200Hz. The optical sensor is mounted to a plastic joint that allows the user to hold the pen at a natural angle while keeping the sensor parallel to the writing surface. There’s also a reflective sensor on the pen tip which allows cameras to track its position in space, for use in combination with VR technology.

The team show off the device being used in several ways, primarily in VR tasks, but also in simple handwriting and coloring work. It’s a project that could readily be replicated by any eager experimenter by gutting a gaming mouse and getting down to work; our writers will expect six of your submissions by June 1st to the tipsline. Those eager to learn more can check out the project paper, and may also find the team’s TapID technology interesting. Video after the break. Continue reading “Flashpen Is A High Fidelity Pen Input Device” →