Trochotron Industries

One product very similar to the Teensy USB development board are the open-source (hardware and software) avropendous boards, which now seem to be available in a DIP form factor ala the Teensy board, at a similarly low price.  The AVRopendous-DIP ($22) uses the same AT90USB162 as the Teensy board, and allows selecting 5v or 3v3 operation via a jumper.  There seem to be some more recent additions to the AVRopendous-DIP family, in particular the ATmega32U4-based AVRopendous2-DIP ($29), the AT90USB646-based AVRopendous3D-DIP ($32), and the AT90USB647-based AVRopendous3H-DIP (not yet for sale?)  The 3D and 3H variants provide a more powerful microcontroller, and the 3H variant adds the ability to function in USB host or OTG modes.  Not to be dismissed however, the AVRopendous2-DIP has more than double the memory (both flash and RAM) of the AT90USB162-based products and can supply the most current (450ma) at 3.3v.

Most of the things I’m interested in interfacing these days is 3v3, which tend to require extra components to safely interface to a 5v device like the Teensy board.  I looked into what it would take to modify the Teensy operate at 3v3, it should just barely be doable with some very delicate soldering.  (I had hoped there was some bit of the Teensy design that would make this trivial, like a 0-ohm resistor I could move to make it work at 3.3v.  Alas, no such luck…)  The AVRopendous design does make this trivial,  with it’s jumper-selectable voltage.  It’s not a perfect substitution since the AVRopendous lacks the compact bootloader that the Teensy ships with (although made up for by additional flash memory on the AVRopendous2-DIP), and the device has a slower clock (8mhz) that is required to work at both voltages.  However, the $25 ATmega32U4-based AVRopendous2-DIPnh (with no headers) option is a deal I couldn’t pass up… Expect to see some pix and a review here when it arrives.

We live in an era of liberal return policies.  After purchasing a defective item from a store, we’ve become accustomed to returning the item to the store in exchange for a working item.  Ever wonder where those items go after they’re returned?  It’s nice to imagine that they somehow make their way back to the factory where they are repaired and resold (presumably as “refurbished” items.)  However the reality is that for most inexpensive items, the items are usually “destroyed” by the retailer in exchange for credit from the manufacturer, where destroyed means they are in some way rendered unsellable and then sent out in the trash.  Quite often the item can be easily repaired, but for mass-market items it’s usually not cost effective for the manufacturer to do so.  One of my personal resolutions has been to buck this trend by fixing items rather than returning them, and to spend the effort to fix items when it would be more convenient to just toss it and buy another one.  Generally I learn something about how things work in the process, and feel good about keeping something out of the landfill.  (Not to mention reducing my carbon footprint by a small amount.)

From this idea I tried to imagine a different world, one where our money doesn’t flow overseas to third-world countries where underpaid workers work tireless hours assembling low-quality goods.  Instead items are designed to be readily serviced, and defective items are generally repaired, more out of principle than whether it makes economic sense to do so.  (Probably with the help of legislation overriding the economic incentive for manufacturers to make their items increasingly disposable.)  An item that lasts a lifetime is true wealth, whereas cheap, disposable items are a constant burden, forever needing to be replaced. Our service economy seems to be working overtime to invent new services that we truly do not need.  ($5 latte anyone?)  If only we could find a way to channel some of that money into cost-effective repair services, we might actually make a difference.

I enjoy working with leather.  As a material, leather has a nice old-world feel to it, being nicely organic yet luxurious and strong.  Often I’ve gotten some inexpensive gadget in a cheaply molded plastic case and have wanted to replace it with something more sturdy.  Unfortunately in many of these gadgets the case serves a crucial function insulating the internal PCB from the outside world.  For this application, leather is not a very good replacement, since it conducts electricity about as well as human skin does, especially when humid or damp. While considering possible solutions, I hunted for some kind of paint-on coating that would serve as an insulating layer, ideally being as durable and flexible as the leather itself.  I ended up finding a product variously known as “liquid electrical tape” or “brush-on electrical tape” sold in small cans having an integral brush.    It is available in various colors, ranging from clear to black.  I was surprised to find this is stocked by my local hardware store, albeit only in black.

A first test of the liquid electrical tape on a piece of scrap leather (veg-tanned tooling hide) shows some promising results.  I painted on three coats, waiting about 5 minutes for it to dry between coats.  After drying, the liquid tape becomes a somewhat rubbery substance that happily flexes and stretches with the leather.  The resulting coating seems to function as a good insulator, even when the leather is completely soaked.  (Not too suprising for a product sold as a replacement for electrical tape.)

Next up is to find a likely candidate for an enclosure makeover…

Don’t get me wrong, I think screensavers are a nifty invention.  However, they can be a major hassle.  Every time I sit down in front of my computer, I’m faced with the arduous task of wiggling the mouse in order to get rid of the screensaver.  This got me thinking… If only there were some way to automate this task.   Now thanks to the Teensy USB board and a spare Parallax PIR motion sensor I had floating around, I have a solution to this problem.  The Parallax sensor is low cost ($10) and pretty much ready-to-use out of the box.  Keith Neufeld has an excellent, detailed write-up of the motion sensor on his blog, along with some revealing pics of all the naughty bits. The sensor is very low power (<100uA) and can operate from a 3.3V to 5V power supply.  It has a single logic output pin which transitions to high when motion is detected, so it could not be any easier to interface.  I connected its power to the Teensy board and wired the output to a GPIO pin on the Teensy board and made a quick hack of the Teensy usb mouse example so that it polls the PIR sensor and wiggles the mouse cursor when  it detects motion after no motion is detected for 10 minutes.  Now whenever I sit down at the machine, the board automatically wakes up the screensaver, saving me from the laborious toil of moving the mouse and all of its associated hazards.

Here are some pictures of the Teensy, now with pins (soldered in pointing upwards) and  interfaced to the PIR motion sensor, also in a custom enclosure.

Lately I’ve been exploring low-cost USB solutions for interfacing microcontrollers to systems without legacy serial or parallel ports.  A obvious and popular solution is a USB to TTL-level RS232 converter.  These are available as an on-board device (as seen in the official Arduino boards) or as a cable or add-on board (popular with the smaller ‘duino clones.)  The best of which seem to be the FTDI TTL232 series of cables which can be had for around $20.   The Modern Device Company recently released an even lower cost USB BUB converter board that looks to be a nice solution.

Beyond simply emulating a serial port are approaches that provide a native USB interface to the microcontroller, allowing a full range of device types to be presented to the host system.  One inspiring example for hobbyists is the USBtiny approach which emulates a USB low-speed device using an Atmel ATtiny microprocessor that lacks native hardware support for USB.  (The Adafruit USBtinyISP AVR programmer uses this approach to great effect.)  This design can be achieved for very low cost with a small number of components; however, the quirk that makes it possible limits it to low-speed operation.  Yet it is still sufficient for many applications.

A step up from there are some low-cost microcontrollers with native USB support.  Some time ago I picked up an Atmel AT90USB1287-based AT90USBKEY demo board, which at $30 seemed like an excellent low-cost hobbyist platform, particularly with its 16MB of onboard flash and support for USB OTG-style host mode.  However the registration-required, EULA encumbered nature of their examples pretty much prevented any sort of open-source community surrounding the board.  One person mentioned on a forum that they had taken the time to convert the example sources to compile under avr-gcc, but they felt they could could not share the patches legally.  So the board sat around collecting dust until earlier today, when I decided to place an order for the $19 AT90USB162-based  Teensy USB development board from PJRC, based on their claims of an open-source USB library and some budding support for a “Teensyduino“  integration with the Arduino IDE.  The find of the day turned out to be that open-source USB library, namely Dean Camera’s LUFA (Lightweight USB Framework for AVRs.)  Interestingly, Dean’s default target just happens to be the aforementioned AT90USBKEY.  A short download and a (rarely occurring) successful first-time ”make all” later and I had a nice set of working examples for the board.  Much kudos to Dean, now I have something else I can hack on while I await the arrival of the Teensy board.  (Not that I ever seem to have a shortage of things to hack on…)