800 X 600 VGA With The STM32F4

January 4, 2015 by Brian Benchoff 12 Comments

Generating VGA is a perennial favorite on the Hackaday tips line, and it’s not hard to see why. Low-res video games, of course, but sending all those pixels out to a screen is actually a pretty challenging feat of coding. The best most project have attained is the original VGA standard, 640×480. Now that we have fast ARMs sitting around, we can bump that up to 800×600, like [Karl] did with an STM32F4 Discovery board.

The problem with generating VGA on a microcontroller is the pixel frequency – the speed at which pixels are shoved out of the microcontroller and onto the screen. For an 800×600 display, that’s 36 MHz; faster than what the 8-bit micros can do, but a piece of cake for the STM32F4 [Karl] is using.

[Karl] started his build by looking at the VGA project Artekit put together. It too uses an STM32, but a 36-pin F103 part. Still, it was fast enough to generate a line-doubled 800×600 display. [Karl] took this code and ported it over to the F4 part on the Discovery board that has enough space for a full 800×600 frame buffer.

With all that RAM on board the F4 part, [Karl] was able to expand the frame buffer and create a relatively high-resolution display with DMA and about a dozen lines of code. It looks great, and now we just need a proper application for high-resolution VGA displays. Retrocomputing? A high-resolution terminal emulator? Who knows, but it’s a great use for the STM32.

If circles and some text aren’t your thing, Artekit also has Space Invaders running on the 36-pin STM32.

Sound Reactive Drums Of Trailing Light

January 4, 2015 by Sarah Petkus 8 Comments

If you’re going to be the drummer in a band for a Back to the Future themed New Years Eve party, you really need to add something to your gig that captures that kitschy futuristic ambiance as seen by the 80s. Rainbow LEDs will do the trick.

For his drum set’s reactive trailing light display, [Alec Smecher] was inspired by a similar project he’d seen in the past where Neopixels were added to a regular drum kit and activated with several individual microphones. Since the microphones ultimately heard all of the thundering noise from every drum and cymbal at once, there was a lot of bleed over in the response of the LEDs. To remedy this, [Alec] used piezo pickups which listen to discrete surface vibrations rather than sound in order to clean up the effect produced by the lights. Each of the five LED strips lining the stands of his cymbal and inside of his drums were programmed to react with a burst of light equal in brightness to the intensity of the vibration sensed by the piezo.

To insure everything kept together amidst all the constant motion and shaking during performance, [Alec] soldered his connections directly onto his Trinket’s pins as well as the fragile pickup of the piezo. The pickup of the sensors were taped directly against the skin of his drums and along the inside of each cymbal to maximize responsiveness. After ringing in the new year appropriately as the ‘band from the future’, [Alec] reports that his colorful addition worked fantastic the whole night.

Those interested in building their own can find a nice schematic on [Alec’s] blog as well as the code he used on github. Difficulty level taken into account, this is a great first project for a musician who has yet to dabble in electronics… and seeing that it’s a brand new year, there’s no better time to have a go at something new.

Continue reading “Sound Reactive Drums Of Trailing Light” →

Digital To Analog To Digital To Analog To Digital Conversion

January 3, 2015 by Brian Benchoff 9 Comments

[Andy] had the idea of turning a mixing desk into a MIDI controller. At first glance, this idea seems extremely practical – mixers are a great way to get a lot of dials and faders in a cheap, compact, and robust enclosure. Exactly how you turn a mixer into a MIDI device is what’s important. This build might not be the most efficient, but it does have the best name ever: digital to analog to digital to analog to digital conversion.

The process starts by generating a sine wave on an Arduino with some direct digital synthesis. A 480 Hz square wave is generated on an ATTiny85. Both of these signals are then fed into a 74LS08 AND gate. According to the schematic [Andy] posted, these signals are going into two different gates, with the other input of the gate pulled high. The output of the gate is then sent through a pair of resistors and combined to the ‘audio out’ signal. [Andy] says this is ‘spine-crawling’ for people who do this professionally. If anyone knows what this part of the circuit actually does, please leave a note in the comments.

The signal from the AND gates is then fed into the mixer and sent out to the analog input of another Arduino. This Arduino converts the audio coming out of the mixer to frequencies using a Fast Hartley Transform. With a binary representation of what’s happening inside the mixer, [Andy] has something that can be converted into MIDI.

[Andy] put up a demo of this circuit working. He’s connected the MIDI out to Abelton and can modify MIDI parameters using an audio mixer. Video of that below if you’re still trying to wrap your head around this one.

Continue reading “Digital To Analog To Digital To Analog To Digital Conversion” →

The Smallest Portable Pi

January 3, 2015 by Brian Benchoff 17 Comments

What do you get when you take an extremely small Raspberry Pi clone and stuff it inside a Game Boy Advance SP? We don’t know what to call it, but it’s probably one of the best portable gaming machines ever made, able to run emulators ranging from the Apple II to playing Quake III natively on a tiny flip-top display.

This isn’t the first time we’ve seen [frostedfires]’ work on a tiny system stuffed into a Game Boy. The initial post on this build over on the bacman forums just covered the basics – getting an Odroid W up and running, and putting Quake III on the tiny display. Now that the build is complete, we can get a look at what it takes to turn a Raspberry Pi clone into one of the smallest portable projects we’ve ever seen.

Using a Raspi clone as the only component in a tiny portable emulation station isn’t possible, so [frostefires] added a few other bits of electronics to make everything work. There’s a joystick from a PSP in there to work as the mouse, a few extra buttons in addition to the stock Game Boy ones, A USB hub, WiFi adapter, speaker and amplifier, a battery and the related charging electronics, and a Teensy 3.1 to handle all the input.

It’s a very impressive build that can run emulators ranging from the Apple II to later generation Nintendo consoles and handhelds (including the Game Boy Advance), but since the HDMI connector is availble on the outside of the case, [frostedfires] can also use this as a tiny, portable media center. Check out the video below to see this Game Boy in action, playing Mario Kart and 1080p video.

Continue reading “The Smallest Portable Pi” →

Boeing 777 From Manilla Folders, A 6+ Year Effort

January 3, 2015 by Matt Freund 67 Comments

The closer you look the more you will be in awe of this shockingly intricate 777 replica. The fully-articulating landing gear alone has over 2,000 parts and 200 hours of assembly, not even including the penny-sized tires with individually-cut lug nuts. All carved from manilla office folders by hand.

A high school ~~art~~ architecture class in 2008 inspired this build by teaching a few papercrafting techniques. When [Luca] got a hold of a precision Air India 777-300ER schematic, he started building this 5 foot long 1:60 scale model. His project has received a fair amount of media attention over the years, including some false reports that he was so focused on the build that he dropped out of college (he did, for 2 years, but for other reasons). 6.5 years in the making, [Luca] is rounding the homestretch.

The design is manually drawn in Illustrator from the schematics, then is printed directly onto the manilla folders. Wielding an X-acto knife like a watch-maker, [Luca] cuts all the segments out and places them with whispers of glue. Pistons. Axles. Clamps. Tie rods. Brackets. Even pneumatic hoses – fractions of a toothpick thin – are run to their proper locations. A mesh behind the engine was latticed manually from of hundreds of strands. If that was not enough, it all moves and works exactly as it does on the real thing.

Continue reading “Boeing 777 From Manilla Folders, A 6+ Year Effort” →

Automatic 3D Scanning On The Cheap

January 3, 2015 by Brian Benchoff 24 Comments

After hearing about a few 3D object scanners, [Will] thought one of these tools could find a place in his workshop. The price of these scanners made him reconsider simply buying one, so he just made one out of parts that were sitting around. This was the first version of his 3D scanner. It worked, but there were a few shortcomings. [Will] had to rotate the object manually. That’s a cheap way of doing it, but the method is tedious.

Now [Will] is back for round two. He’s made some improvements, and this time a few bits of electronics automate the process, allowing [Will] to hit a button, walk away, and come back to a scanned object.

Even though [Will] has improved his setup immensely, the theory of how to scan an object remains the same. He’s projecting a straight vertical line on an object, taking a few snapshots with a webcam, and reconstructing the object with computer vision algorithms and Meshlab. The new additions include a BeagleBone Black, a stepper motor and an EasyDriver from Sparkfun, and a turntable.

[Will] wrote two scripts for this project. The first does the mechanical heavy lifting – turning the stepper motor and taking a picture, while the second converts the output from the webcam to a point cloud. From there, the point cloud is sent over to Meshlab, and an object appears on [Will]’s hard drive.

There’s about $80 in hardware invested in this setup, and considering the inspiration for this project was the $800 Makerbot Digitizer, we’re going to call [Will]’s experiments in 3D scanning a success.

Peculiar Radial Mill From Car Parts

January 3, 2015 by Matt Freund 62 Comments

Whether 3D printer, lasercutter, or mill, most CNC machines use human-friendly, square-angle Cartesian geometry. This intriguing concept mill instead uses radial axes where motion is derived from scrap Chevy flywheels. It may look and feel weird at first, but it works – sort of.

Cartesian axes are intuitive. If you want to go to the right, increase X. If you want to go to away from you, increase Y. If you want to lift, increase Z. On a manual mill this is easy for making rectangles and blocks, or, with creative clamping, straight lines of any sort. But if you want to carve a circle? As we all learned on an Etch-A-Sketch, you increase your swearing and then throw it in the corner.

[Jason] knew that with a CNC machine all geometry problems are reduced to math done by software. With two offset discs, any position is possible by rotating both the correct way. It may look odd that both plates drunkenly meander about just to draw a straight line but the computer is ambivalent. Software can be complicated without penalty and is free once written – more on that later. If a machine is physically simple then it can be built and repaired easily and cheaply. This design does away with almost all the familiar – and [Jason] argues complicated – components of normal hobby CNC machines. No slides, rails, carriages or belts here. His design uses only about a dozen parts.

Because automotive flywheels are made from cast iron the machine is rigid and naturally dampening. Sticking with the junkyard theme he pulled bearings from an F-450 truck, good for a few thousand pounds. Some steppers and a Raspberry Pi and he was done – well, sort of.

[Jason] let us know that his project has sat for long enough that he has become passionate about other things and decided to move on. He documented his progress and submitted the tip in hope to inspire someone else to continue the design further. Any type of CNC is possible, not just a mill. 3D printer perhaps?

Two big caveats: it needs a Z-axis (linear, probably standard) and there appears to be deeper-seated-than-expected G-code demands to chit-chat about rectangles and only rectangles. Nothing insurmountable, just nothing he has solved yet himself.

[Jason] said not to expect any further updates from him but he would love to see what the next person could do with it.

See the video after the break of the mill drawing our skull and wrenches logo, (soft of, without a Z-axis to lift).

Continue reading “Peculiar Radial Mill From Car Parts” →