Watch Sony Engineers Tear Down Sony’s VR Hardware

Teardowns are great because they let us peek not only at a product’s components, but also gain insight into the design decisions and implementations of hardware. For teardowns, we’re used to waiting until enthusiasts and enterprising hackers create them, so it came as a bit of a surprise to see Sony themselves share detailed teardowns of the new PlayStation VR2 hardware. (If you prefer the direct video links, Engineer [Takamasa Araki] shows off the headset, and [Takeshi Igarashi] does the same for the controllers.)

The “adaptive trigger” module responsible for the unique feedback.

One particularly intriguing detail is the custom tool [Araki] uses to hold the headset at various stages of the disassembly, which is visible in the picture above. It looks 3D-printed and carefully designed, and while we’re not sure what it’s made from, it does have a strong resemblance to certain high-temperature SLA resins. Those cure into hard, glassy, off-yellow translucent prints like what we see here.

As for the controller, we get a good look at a deeply interesting assembly Sony calls their “adaptive trigger”. What’s so clever about it? Not only can it cause the user to feel a variable amount of resistance when pulling the trigger, it can even actively push back against one’s finger, and the way it works is simple and effective. It is pretty much the same as what is in the PS5 controller, so to find out all about how it works, check out our PS5 controller teardown coverage.

The headset and controller teardown videos are embedded just below. Did anything in them catch your interest? Know of any other companies doing their own teardowns? Let us know in the comments!

Continue reading “Watch Sony Engineers Tear Down Sony’s VR Hardware”

Tiny PCB Banishes Soldering Fumes, Automatically

A fan to remove fumes is a handy thing to have when soldering, even better is a fan furnished with a filter. Better still is a fan that activates only when the iron is in use, turning off when the iron is in its stand. Now that’s handy!

[Petteri Aimonen] made exactly such a device when he noticed his JBC BT-2BWA soldering station could detect when the iron is removed from its stand, and indicate its operating mode via status LEDs. Broadly speaking, when the iron is removed from its cradle the green “in use” LED is on. By turning the fan on whenever that LED is lit (and turning it off when it becomes unlit), fume extraction gets a little more elegant and efficient.

Instead of tapping directly into the soldering station’s hardware to detect the LED’s state, [Petteri] went for a completely noninvasive solution that made good use of a few spare parts and a small bit of copper-clad board. The PCB is nothing more than piece of copper-clad board with lands scratched out with a hobby knife.

This tiny board sits atop the soldering station, parking a photodiode directly above the “in use” LED. The circuit is a simple comparator whose output controls fan power via a MOSFET, and a top-facing LED provides as a duplicate “in use” indicator, since the original is hidden under the tiny board.

Even for one-off designs like this, creating a PCB layout in an EDA program like KiCad is still worth doing because one can use it to scratch out lands on a copper-clad board, a technique with similarities to Manhattan-style circuit construction.

How To Roll Your Own Custom Object Detection Neural Network

Real-time object detection, which uses neural networks and deep learning to rapidly identify and tag objects of interest in a video feed, is a handy feature with great hacker potential. Happily, it’s also possible to make customized CNNs (convolutional neural networks) tailored for one’s own needs, and that process just got easier thanks to some new documentation for the Vizy “AI camera” by Charmed Labs.

Raspberry Pi-based Vizy camera

Charmed Labs has been making hacker-friendly machine vision devices for a long time, and the Vizy camera impressed us mightily when we checked it out last year. Out of the box, Vizy has a perfectly functional object detector application that runs locally on the device, and can detect and tag many common everyday objects in real time. But what if that default application doesn’t quite meet one’s project needs? Good news, because it’s possible to create a custom-trained CNN, and that process got a lot more accessible thanks to step-by-step examples of training a model to recognize hands doing rock-paper-scissors.

Person and cat with machine-generated tags identifying them
Default object detection works well, but sometimes one needs custom results.

The basic process is this: Start with a variety of images that show the item of interest. Then identify and label the item of interest in each photo. These photos (a “training set”) are then sent to Google Colab, which will be used to generate a neural network. The resulting CNN model can then be downloaded and used, to see how well it performs.

Of course things rarely work perfectly the first time around, so at this point it’s pretty common for some refinement to be needed to increase accuracy. Luckily there are a number of tools to help do this without creating a new model from scratch, so it’s just a matter of tweaking until things perform acceptably.

Google Colab is free and the resulting CNNs are implemented in the TensorFlow Lite framework, meaning it’s possible to use them elsewhere. So if custom object detection has been holding up a project idea of yours, this might be what gets you over that hump.

Stadia Controller’s Two Extra Buttons Get Seen With WebHID

The Google Stadia game streaming service relied on a proprietary controller. It was a pretty neat piece of hardware that unfortunately looked destined for landfills when Google announced that Stadia would discontinue. Thankfully it’s possible to use them as normal gamepads, and related to that, [Thomas Steiner] has a developer blog post about how to talk to the Stadia controller via WebHID. Continue reading “Stadia Controller’s Two Extra Buttons Get Seen With WebHID”

RCA’s Clear Plastic TV Wowed Crowds In 1939

In the United States in 1939, television sets still had a long way to go before they pretty much sold themselves. Efforts to do just that are what led to RCA’s Lucite Phantom Telereceiver, which aimed to show people a new way to receive broadcast media.

Created for the 1939 World’s Fair, the TRK-12 Lucite Phantom Telereceiver introduced people to the concept of television. Production models were housed in contemporary wood cabinets, but the clear acrylic (itself also a relatively new thing) units allowed curious potential customers to gaze within, and see what was inside these devices.

One interesting feature is the vertically-mounted cathode ray tube, which reflects off a mirror in the top cover of the cabinet for viewing. This meant that much of the bulk of the TRK-12 could be vertical instead of horizontal. Important, because the TRK-12 was just over a meter tall and weighed 91 kilograms (or just over 200 lbs.)

Clearly a luxury item, the TRK-12 sold for $600 which was an eye-watering sum for the time. But it was a glimpse of the future, and as usual, the future is made available a few ticks early to those who can afford the cost.

Want to see one in person? You might be in luck, because an original resides at the MZTV Museum of Television in Toronto, Canada.

Domino Ring Machine Tips Tiles In A Never-ending Wave

Like to see dominoes fall? [JK Brickworks] has got what you need, in the form of a never-ending ring of falling and resetting tiles. LEGO pieces are the star in this assembly, which uses a circular track and moving ramp to reset tiles after they have fallen. Timed just right, it’s like watching a kinetic sculpture harmoniously generating a soliton wave as tiles fall only to be endlessly reset in time to fall again.

A Mindstorms IR sensor monitors a tile’s state for timing.

It’s true that these chunky tiles aren’t actually dominoes — not only are they made from LEGO pieces and hinged to their bases, they have a small peg to assist with the reset mechanism. [JK Brickworks] acknowledges that this does stretch the definition of “dominos”, but if you’re willing to look past that, it’s sure fun to see the whole assembly in action.

The central hub in particular is a thing of beauty. For speed control, an IR sensor monitors a single domino’s up/down state and a LEGO Mindstorms EV3 with two large motors takes care of automation.

The video does a great job of showing the whole design process, especially the refinements and tweaks, that demonstrate the truly fun part of prototyping. [JK Brickworks] suggests turning on subtitles for some added details and technical commentary, but if you’re in a hurry skip directly to 4:55 to see it in action.

Want to see more automated domino action? This domino-laying robot sets them up for you to knock down at your leisure, and this entirely different robot lays out big (and we do mean BIG) domino art displays.

Continue reading “Domino Ring Machine Tips Tiles In A Never-ending Wave”

With ChatGPT, Game NPCs Get A Lot More Interesting

Not only is AI-driven natural language processing a thing now, but you can even select from a number of different offerings, each optimized for different tasks. It took very little time for [Bloc] to mod a computer game to allow the player to converse naturally with non-player characters (NPCs) by hooking it into ChatGPT, a large language model AI optimized for conversational communication.

If you can look past the painfully-long loading times, even buying grain (7:36) gains a new layer of interactivity.

[Bloc] modified the game Mount & Blade II: Bannerlord to reject traditional dialogue trees and instead accept free-form text inputs, using ChatGPT on the back end to create more natural dialogue interactions with NPCs. This is a refinement of an earlier mod [Bloc] made and shared, so what you see in the video below is quite a bit more than a proof of concept. The NPCs communicate as though they are aware of surrounding events and conditions in the game world, are generally less forthcoming when talking to strangers, and the new system can interact with game mechanics and elements such as money, quests, and hirelings.

Starting around 1:08 into the video, [Bloc] talks to a peasant about some bandits harassing the community, and from there demonstrates hiring some locals and haggling over prices before heading out to deal with the bandits.

The downside is that ChatGPT is currently amazingly popular. As a result, [Bloc]’s mod is stuck using an overloaded service which means some painfully-long load times between each exchange. But if you can look past that, it’s a pretty fascinating demonstration of what’s possible by gluing two systems together with a mod and some clever coding.

Take a few minutes to check out the video, embedded below. And if you’re more of a tabletop gamer? Let us remind you that it might be fun to try replacing your DM with ChatGPT.

Continue reading “With ChatGPT, Game NPCs Get A Lot More Interesting”