First Look: Macchina M2

In the past few years, we’ve seen a growth in car hacking. Newer tools are being released, which makes it faster and cheaper to get into automotive tinkering. Today we’re taking a first look at the M2, a new device from the folks at Macchina.

The Macchina M1 was the first release of a hacker friendly automotive device from the company. This was an Arduino compatible board, which kept the Arduino form factor but added interface hardware for the protocols most commonly found in cars. This allowed for anyone familiar with Arduino to start tinkering with cars in a familiar fashion. The form factor was convenient for adding standard shields, but was a bit large for using as a device connected to the industry standard OBD-II connector under the dash.

The Macchina M2 is a redesign that crams the M1’s feature set into a smaller form factor, modularizes the design, and adds some new features. With their Kickstarter launching today, they sent us a developer kit to review. Here’s our first look at the device.

Continue reading “First Look: Macchina M2”

A Full Speed, Portable Apple //e

A while back, [Jorj] caught wind of a Hackaday post from December. It was a handheld Apple IIe, emulated on an ATMega1284p. An impressive feat, no doubt, but it’s all wrong. This ATapple only has 12k of RAM and only runs at 70% of the correct speed. The ATapple is impressive, but [Jorj] knew he could do better. He set out to create the ultimate portable Apple IIe. By all accounts, he succeeded.

This project and its inspiration have a few things in common. They’re both assembled on perfboard, using tiny tact switches for the keyboard. The display is a standard TFT display easily sourced from eBay, Amazon, or Aliexpress. There’s a speaker for terribad Apple II audio on both, and gigantic 5 1/4″ floppies have been shrunk down to the size of an SD card. That’s where the similarities end.

[Jorj] knew he needed horsepower for this build, so he turned to the most powerful microcontroller development board he had on his workbench: the Teensy 3.6. This is a 180 MHz ARM Cortex M4 running a full-speed Apple IIe emulator. Writing a simple 6502 emulator is straightforward, but Apple IIe emulation also requires an MMU. the complete emulator is available in [Jorj]’s repo, and passes all the tests for 6502 functionality.

The project runs all Apple II software with ease, but we’re really struck by how simple the entire circuit is. Aside from the Teensy, there really isn’t much to this build. It’s an off-the-shelf display, a dead simple keyboard matrix, and a little bit of miscellaneous circuitry. It’s simple enough to be built on a piece of perfboard, and we hope simple enough for someone to clone the circuit and share the PCBs.

Pancake-ROM: Eat-only Memory?

You can store arbitrary data encoded in binary as a pattern of zeros and ones. What you do to get those zeros and ones is up to you. If you’re in a particularly strange mood, you could even store them as strips of chocolate on Swedish pancakes.

Oddly enough, the possibility of the pancake as digital storage medium was what originally prompted [Michael Kohn] to undertake his similar 2013 project where he encoded his name on a paper wheel. Perhaps wisely, he prototyped on a simpler medium. With that perfected, four years later, it was time to step up to Modified Swedish Pancake Technology (MSPT).

pancake_rom_bottomHighlights of the build include trying to optimize the brightness difference between chocolate and pancake. Reducing the amount of sugar in the recipe helps increase contrast by reducing caramelization, naturally. And cotton balls placed under the spinning cardboard platform can help stabilize the spinning breakfast / storage product.

Even so, [Michael] reports that it took multiple tries to get the sixteen bytes (bites?) of success in the video below. The data is stenciled onto the pancake and to our eye is quite distinct. Improvement seems to be more of an issue with better edge detection for the reflectance sensor.

Continue reading “Pancake-ROM: Eat-only Memory?”

A 6502 Retrocomputer In A Very Tidy Package

One of the designers whose work we see constantly in the world of retrocomputing is [Grant Searle], whose work on minimal chip count microcomputers has spawned a host of implementations across several processor families.

Often a retrocomputer is by necessity quite large, as an inevitable consequence of having integrated circuits in the period-correct dual-in-line packages with 0.1″ spaced pins. Back in the day there were few micros whose PCBs were smaller than a Eurocard (100 mm x 160 mm, 4″ x 6.3″), and many boasted PCBs much larger.

[Mark Feldman] though has taken a [Grant Searle] 6502 design and fitted it into a much smaller footprint through ingenious use of two stacked Perf+ prototyping boards. This is a stripboard product that features horizontal traces on one side and vertical on the other, which lends itself to compactness. Continue reading “A 6502 Retrocomputer In A Very Tidy Package”

Shape Programmable Matter Is More Magnetic Magic

How could you build an artificial tadpole? Or simulate the motion of a cilium? Those would be hard to do with mechanical means — even micromechanical because of their fluid motion. Researchers have been studying shape-programmable matter: materials that can change shape based on something like heat or magnetic field. However, most research in this area has relied on human intuition and trial and error to get the programmed shape correct. They also are frequently not very fast to change shape.

[Metin Sitti] and researchers at several institutions have found a way to make rapidly changing silicone rubber parts (PDF link) that can change shape due to a magnetic field. The method is reproducible and doesn’t seem out of reach for a hackerspace or well-equipped garage lab.

Continue reading “Shape Programmable Matter Is More Magnetic Magic”

The Elements Converge For ±.002 In Tolerance

What can be accomplished with just a torch and compressed air? We can think of many things, but bringing a 17-foot-long marine shaft into ±.002 in tolerance was not on our list.

Heat straightening (PDF) utilizes an oxy-acetylene flame that is used to quickly heat a small section of a workpiece. As the metal cools, it contracts more than it expanded when heated, resulting in a changed volume. With skill, any distortions on a shaft can theoretically be straightened out with enough time (and oxy-acetylene). Heat straightening is commonly applied to steel but works on nickel, copper, brass and aluminum additionally.

[Keith Fenner’s] standard process for trueing stock is sensitive enough that even sunlight can introduce irregularities, but at the same time is robust enough to carry out in your driveway. However, even though the only specialty tools you need are a torch, compressed air and work supports, watching [Keith] work makes it clear that heat straightening is as much an art as it is a science. Check out his artistry in the video below the break. Continue reading “The Elements Converge For ±.002 In Tolerance”

Neural Nets And Game Boy Cameras

Released in 1998, the Game Boy camera was perhaps the first digital camera many young hackers got their hands on. Around the time Sony Mavica cameras were shoving VGA resolution pictures onto floppy drives, the Game Boy camera was snapping 256×224 resolution pictures and displaying them on a 190×144 resolution display. The picture quality was terrible, but [Roland Meertens] recently had an idea. Why not use neural networks to turn these Game Boy Camera pictures into photorealistic images?

Neural networks, deep learning, machine learning, or whatever other buzzwords we’re using require training data. In this case, the training data would be a picture from a Game Boy Camera and a full-color, high-resolution image of the same scene. This dataset obviously does not exist so [Roland] took a few close up head shots of celebrities and reduced the color to four shades of gray.

[Roland]'s face captured with the Game Boy Camera (left), and turned into a photorealistic image (right)
[Roland]’s face captured with the Game Boy Camera (left), and turned into a photorealistic image (right)
For the deep machine artificial neural learning part of this experiment, [Roland] turned to a few papers on converting photographs to sketches and back again, real-time style transfer. After some work, this neural network turned the test data back into images reasonably similar to the original images. This is what you would expect from a trained neural network, but [Roland] also sent a few pics from the Game Boy Camera through this deep machine artificial learning minsky. These images turned out surprisingly well – a bit washed out, but nearly lomographic in character.

We’ve seen a lot of hacks with the Game Boy Camera over the years. Everything from dumping the raw images with a microcontroller to turning the sensor into a camcorder has been done. Although [Roland]’s technique will only work on faces, it is an excellent example of what neural networks can do.