Trinocular Lens Makes Digital Wigglegrams Easier To Take

Everyone likes a good animated GIF, except for some Hackaday commenters who apparently prefer to live a joyless existence. And we can’t think of a better way to celebrate moving pictures than with a 3D printed trinocular camera that makes digital Wigglegrams a snap to create.

What’s a Wigglegram, you say? We’ve seen them before, but the basic idea is to take three separate photographs through three different lenses at the same time, so that the parallax error from each lens results in three slightly different perspectives. Stringing the three frames together as a GIF later results in an interesting illusion of depth and motion. According to [scealux], the inspiration for building this camera came from photographer [Kirby Gladstein]’s work, which we have to admit is pretty cool.

While [Kirby] uses a special lenticular film camera for her images, [scealux] decided to start his build with a Sony a6300 mirrorless digital camera. A 3D printed lens body with a focusing mechanism holds three small lenses which were harvested from disposable 35 mm film cameras — are those still a thing? Each lens sits in front of a set of baffles to control the light and ensure each of the three images falls on a distinct part of the camera’s image sensor.

The resulting trio of images shows significant vignetting, but that only adds to the charm of the finished GIF, which is created in Photoshop. That’s a manual and somewhat tedious process, but [scealux] says he has some macros to speed things up. Grainy though they may be, we like these Wigglegrams; we don’t even hate the vertical format. What we’d really like to see, though, is to see everything done in-camera. We’ve seen a GIF camera before, and while automating the post-processing would be a challenge, it seems feasible.

Continue reading “Trinocular Lens Makes Digital Wigglegrams Easier To Take”

Stop Silicone Cure Inhibition, No Fancy Or Expensive Products Required

Casting parts in silicone is great, and 3D printing in resin is fantastic for making clean shapes, so it’s natural for an enterprising hacker to want to put the two together: 3D print the mold, pour in the silicone, receive parts! But silicone’s curing process can be inhibited by impurities. What’s cure inhibition? It’s a gross mess as shown in the image above, that’s what it is. Sadly, SLA-printed resin molds are notorious for causing exactly that. What’s a hacker to do?

Firstly: there are tin-cure and platinum-cure silicones, and for the most part tin-cure silicone works just fine in resin-printed molds. Platinum-cure silicones have better properties, but are much more susceptible to cure inhibition. Most workarounds rely on adding some kind of barrier coating to molds, but [Jan Mrázek] has a cheap and scalable method of avoiding this issue that we haven’t seen before. Continue reading “Stop Silicone Cure Inhibition, No Fancy Or Expensive Products Required”

Hackaday Links Column Banner

Hackaday Links: October 30, 2022

Sad news for kids and adults alike as Lego announces the end of the Mindstorms line. The much-wish-listed line of robotics construction toys will be discontinued by the end of this year, nearly a quarter-century after its 1998 introduction, while support for the mobile apps will continue for another couple of years. It’s probably fair to say that Mindstorms launched an entire generation of engineering careers, as it provided a way to quickly prototype ideas that would have been difficult to realize without the snap-fit parts and easily programmed controllers. For our money, that ability to rapidly move from idea to working model was perhaps the strongest argument for using Mindstorms, since it prevented that loss of momentum that so often kills projects. That was before the maker movement, though, and now that servos and microcontrollers are only an Amazon order away and custom plastic structural elements can pop off a 3D printer in a couple of hours, we can see how Mindstorms might no longer be profitable. So maybe it’s a good day to drag out the Mindstorms, or even just that big box of Lego parts, and just sit on the carpet and make something.

Continue reading “Hackaday Links: October 30, 2022”

The SSD described, a green board with a ZIP connector, a controller chip and two out of four NAND chips populated. There's traces of flux on the chip, as it hasn't been washed after soldering yet.

ZIF HDDs Dying Out? Here’s An Open-Source 1.8″ SSD

A lot of old technology runs on parts no longer produced – HDDs happen to be one such part, with IDE drives specifically being long out of vogue, and going extinct to natural causes. There’s substitutes, but quite a few of them are either wonky or require expensive storage medium. Now, [dosdude1] has turned his attention to 1.8 ZIF IDE SSDs – FFC-connected hard drives that are particularly rare and therefore expensive to replace, found in laptops like the Macbook Air 1,1 2008 model. Unsatisfied with substitutes, he’s designed an entire SSD from the ground up around an IDE SSD controller and NAND chips. Then, he made the design open-source and filmed an assembly video so that we can build our own. Take a look, we’ve put it below the break!

For an open-source design, there’s a respectable amount of work shared with us. He’s reverse-engineered some IDE SSDs based on the SM2236 controller to design the schematic, and put the full KiCad files on GitHub. In the video, he shows us how to assemble this SSD using only a hot air station and a soldering iron, talks about NAND matching and programming software intricacies, and shows the SSD working in the aforementioned Macbook Air. Certainly, assembly would have been faster and easier with a stencil, but the tools used work great for what’s a self-assembly tutorial!

Continue reading “ZIF HDDs Dying Out? Here’s An Open-Source 1.8″ SSD”

Garmin HUD Got Discontinued, But Not Trashed

The Garmin HUD+ was a small Bluetooth device intended for the dashboard of a car, meant to be used as a GPS heads-up display for data from Garmin smartphone apps. It used a bright VFD (vacuum fluorescent display) which was viewed through a clear reflector, and displayed GPS information and directions. It was discontinued in 2015, but [Doz] was fond of his and used it happily until a phone upgrade meant it no longer worked. Was it destined for a landfill? Not if he had anything to say about it!

The first thing [Doz] tried was using an alternate Android app, but since it also didn’t work, it was time to sit back and reflect on the scope of the issue. In [Doz]’s case, he really only wanted some basic meaningful data displayed, and decided he could do away with the phone altogether if he had the right hardware. Continue reading “Garmin HUD Got Discontinued, But Not Trashed”

Playing With The Power Of Full G-Code Control

Slicing software needs to maintain a balance between ease-of-use and control, while handling handle any STL file you throw at it. If you eliminate the need to convert an existing 3D model, and create G-Code directly, you gain a lot of design freedom, at the cost of increased design effort. By taking advantage of this freedom and making it more accessible, [Andrew Gleadall] and [Dirk Leas] created the FullControl Design Library.

Each model is a mathematically generated extrusion path with a host of adjustable design parameters and print settings. This allows you to print things like a single-layer non-planar part, or 90° overhangs without any support (video after the break). The website was built using the python version of the original Excel-based FullControl Designer (unreleased at the time of writing), and threejs for the 3D visualization.

Go browse the library, play with some parameters and see what strikes your fancy. For ideas, help and updates, keep an eye on the FullControl Subreddit.

Continue reading “Playing With The Power Of Full G-Code Control”

Recreating The “Stuck Throttle” Problem On A Toyota

A few years ago, Toyota was in the news for a major safety issue with a number of their passenger vehicles. Seemingly at random, certain cars were accelerating without concern for driver input, causing many crashes and at least 37 confirmed deaths. They issued recalls both for the floor mats which were reported to have slid forward to jam the accelerator pedal, but this didn’t explain all of these crashes. There was another recall for stuck throttles, which [Colin O’Flynn] demonstrates a possible cause for on his test bench.

While most passenger vehicles older than about 15-20 years controlled the throttle with a cable connected directly from the throttle body to the accelerator pedal, most manufacturers have switched to a fly-by-wire system which takes sensor input from the accelerator pedal and sends that position information to the vehicle’s computer which in turn adjusts the throttle position. This might be slightly cheaper to manufacture, but introduces a much larger number of failure modes to a critical system. Continue reading “Recreating The “Stuck Throttle” Problem On A Toyota”