The Vintage Computer Festival last weekend featured racks and racks of old minicomputers, enough terminals for an entire lab, and enough ancient storage devices to save a YouTube video. These storage devices – hard disks, tape readers, and 8″ disk drives – were only connected to vintage hardware, with one exception: a DEC RL02 drive connected to a modern laptop via USB.
The DEC RL02 drive is the closest you’re going to get to a modern mechanical hard drive with these old machines. It’s a huge rack unit with removable platters that can hold 10 Megabytes of storage. [Chris] found one of these old drives and because he wanted to get into FPGA development, decided to create a USB adapter for this huge, old drive.
The hardware isn’t too terribly complex, with a microcontroller and an FPGA that exposes the contents of the drive over USB mass storage. For anyone trying to bootstrap a PDP-11 or -8 system, [Chris] could download disk images from the Internet, write them to the disk, and load up the contents of the drive from the minicomputer. Now, he’s using it with SimH to have a physical drive for an emulated system, but the controller really doesn’t care about what format the disk pack is in. If [Chris] formatted a disk pack with a FAT file system, he would have the world’s largest and heaviest USB thumb drive in the world.
Since the Raspberry Pi 2 was released, everyone building RetroPi emulators has been graced with four USB ports. For those of us doing useful stuff with the Pi, those ports are a little anemic: you can’t plug in a webcam and a WiFi module at the same time without suffering CPU brownouts. The maximum current all USB peripherals can draw from the USB port is 600mA. By changing a value in the /boot/config.txt file, this current limit can be increased to 1.2A for all four ports.
Because the USB current limit is set in software, there must be a few bits of hardware that do the actual work. Tucked away below the right hand of the GPIO header is the hardware that does exactly that. It’s an AP2253 current-limited power switch (PDF), and the current is adjustable by tying a resistor to pin 5 on the chip.
Pin 5 on the AP2253 is connected to two resistors. One resistor goes directly to a ground plane, while the other is switched through a FET. The gate of this FET goes to another resistor, and when a GPIO pin is high, these resistors are wired in parallel. This means the resistance is halved when the GPIO pin is high, doubling the current limiting circuit in the AP2253.
This setup provides a relatively easy mod to increase the current limiting of the USB ports so they can provide 4x500mA, meeting the USB spec. The AP2253 power switch’s current limiting can be set by a single resistor, anywhere from 10kΩ to 232kΩ. By removing R50 and R4, and replacing R50 with a 10kΩ resistor, the current limiting of the AP2253 switch will be set to its maximum, 2.1A. Divide that by four, and you have 500mA per port, just like every other computer on the planet.
There is a reason the Raspberry Pi foundation set the current limiting of the USB ports so low. The Pi was originally intended to run off of a micro USB phone charger. There aren’t many phone chargers out there that will supply more than 1A, and the CPU and related peripherals will take half of that. If you’re going to change the /boot/config.txt file, you’re going to need a beefy power supply. Increasing the current limiting of the USB ports to 2A will require an even bigger, beefier supply.
The USB Implementers’ Forum doesn’t make things easy for anyone building a product with a USB port. To sell anything with USB and have it work like USB should, you need to buy a USB Vendor ID, a $5000 license that grants you exclusive use of 65,536 USB Product IDs. Very few companies will ever release 65,000 products, and there are a lot of unused PIDs sitting around out there.
Now, someone has finally done the sensible thing and put an unused USB VID to work. pid.codes obtained the rights to a single VID – 0x1209 – and now they’re parceling off all the PIDs that remain to open source hardware projects.
This is not a project supported by the USB Implementers’ Forum, and is more of a legal game of chicken on the part of pid.codes. The only thing the USB-IF could do to stop this is revoke the original VID; useless, because they can’t reassign it to anyone else. The original owners of the VID, InterBiometrics, licensed their VID before transferring or sublicensing VIDs and PIDs was prohibited by the USB-IF.
You can get a PID by forking the pid.codes repo, claiming a PID, and sending a pull request. Once that’s accepted, that PID is yours forever.
The availability of Smart RGB LED’s, either as individual units, as strips or even as panels, have made blinky light projects with all kinds of color control and transition effects easy to implement using even the simplest of controllers. Libraries that allow control of these smart LEDs (or Smart Pixels as they are sometimes called) make software development relatively easy.
[overflo] at the Metalab hackerspace in Vienna, Austria recently completed development of usblinky – a hacker friendly blinky USB stick. It can control up to 150 WS2812B smart LED’s when powered via an external power supply, or up to 20 LED’s when powered via a computer USB port. The micro-controller is an ATTiny85 running the Micronucleus bootloader which implements software USB using vUSB. The hardware is based on the DigiSpark platform. The usblinky software sources are available on their Github repo. The section on pitfalls and lessons learned makes for interesting reading.
Metalab plans to run workshops around this little device to get kids into programming, as it is easy enough and gives quick visual feedback to get you started. To round off the whole project, [overflo] used OpenSCAD to design a customizable, 3D printable “parametric orb” which can house the LED strip and make a nice enclosure or psychedelic night light. Check out the mesmerizing video of the usblinky Orb after the break.
[Dark Purple] recently heard a story about how someone stole a flash drive from a passenger on the subway. The thief plugged the flash drive into his computer and discovered that instead of containing any valuable data, it completely fried his computer. The fake flash drive apparently contained circuitry designed to break whatever computer it was plugged into. Since the concept sounded pretty amazing, [Dark Purple] set out to make his own computer-frying USB drive.
While any electrical port on a computer is a great entry point for potentially hazardous signals, USB is pretty well protected. If you short power and ground together, the port simply shuts off. Pass through a few kV of static electricity and TVS diodes safely shunt the power. Feed in an RF signal and the inline filtering beads dissipate most of the energy.
To get around or break through these protections, [Dark Purple]’s design uses an inverting DC-DC converter. The converter takes power from the USB port to charge a capacitor bank up to -110VDC. After the caps are charged, the converter shuts down and a transistor shunts the capacitor voltage to the data pins of the port. Once the caps are discharged, the supply fires back up and the cycle repeats until the computer is fried (typically as long as bus voltage is present). The combination of high voltage and high current is enough to defeat the small TVS diodes on the bus lines and successfully fry some sensitive components—and often the CPU. USB is typically integrated with the CPU in most modern laptops, which makes this attack very effective.
[Artificial Intelligence] has made a desk lamp out of parts he had kicking around in his parts bin. Most recognizable are the 4 CDs that make up the base and the shade. To start this project, [Artificial Intelligence] sketched out a circular pattern on one of the CDs and marked 7 locations where the LEDs will be. Holes were drilled at those marked locations, the LEDs inserted and hot glued into place. Each LED has its own current limiting resistor soldered in a series configuration.
[Artificial Intelligence] mentions the resistor value was determined by a nice LED resistor calculator he found online, ledcalc.com. Then each LED/resistor combo was wired together in a parallel configuration and covered up by another CD to clean up the look and protect the wiring.
The base, like the top, is also made from 2 CDs, but this time there are 5 AA batteries underneath the CDs. These batteries don’t power the lamp, they are only used as a counterweight to prevent the lamp from tipping over. A USB cord runs to the lamp base, goes through an on/off switch and then up a pair of large-gauge solid core wire before connecting to the LED’s in the top of the lamp. The thick solid core wire acts as the only support for the lamp shade and LEDs. Since it is still just wire, the lamp can be bent to shine light in the most convenient position, as any good desk lamp would be capable of.
There’s many complex systems for automatically pointing a telescope at an object in the sky, but most of them are too expensive for the amateur astronomer. [Kevin]’s Arduino ST4 interface lets you connect your PC to a reasonably priced motorized telescope mount, without ripping it apart.
The ST4 port is a very basic interface. There’s one pin per direction that the mount can move, and a common pin. This port can be added to just about any motorized mount with some modification to the controller. To connect to an Arduino, a TLP521-4 quad optoisolator is used. This keeps the Arduino and PC fully isolated from the motor circuits. but lets the Arduino take control of the mount.
With the hardware in place, [Kevin] cranked out some software which is available on Google Code. A simple Arduino sketch provides the USB interface, and a custom driver allows the ASCOM Platform to control the mount. Since many astronomy software tools support ASCOM, this allows the mount to be controlled by existing software.
With the interface in place, the mount can be used to find objects (GOTO) and automatically follow them with high accuracy (autoguiding). You can watch the telescope move on its own after the break.