Thursday, February 23, 2006


Sparkfun Electronics is a great site I've been poking around at for a long time. Just wish I could afford some of it.

Current wishlist that others might be interested in:
-Single transducer sonar unit with decent signal filtering onboard. Looks really nice.
-Various GPS units up the wazoo with lots of antenna options.
-Fully functional cell phone modules. Talk about tempting to build your own!
-Speaking of phones, they have a contest going on currently to figure out the hardware for their surplus, undocumented cameras from the Samsung E700.
-(I already have these) Affordable hot air rework and soldering stations. Quite useful
-accelerometers and gyros, ranging from single units to full IMU's. They are actually required to NOT sell the chips loose, but this is a decent way to get into the projects.
-Magnetic sensors for compass navigation.

I'm most interested in a small GPS module, a 3 axis mag sensor, and the sonar unit. That's about $170 worth of parts, so I'm afraid I have to hold off for a LONG while.

Saturday, February 18, 2006

Impact Mechanics

One of my odder interests are impact and protection. Slashdot recently ran an article with someone advocating using basically a memory foam for hardening armor. Now, I don't have a sample of memory foam right now, but I've seen Silly Putty and high density water-corn starch concoctions flow and move at one time and act like steel at a different time. I think that professional systems for impact absorption (special sparring suits) use a Neoprene outer shell, EVA underlay and a high density memory foam core, and probably similar layering going out the other side.

Now, WHY am I suddenly interested in this?

I've always got my mind on using stuff I build or want to build. I often wander back to Airsoft or Paintball as test grounds and play fields. Paint ball especially can HURT, so people wear armor. Now, most is fairly hard, but what if a lighter weight, fully pliable material was possible? Because it's local hardening, it should give more warning that you've been hit while still keeping the impact down. I've had a line of balls fired down my midsection before, it knocked the breath out of me. I'd have loved to have an inch or two of soft armor between me and them. Airsoft, due to the lower delivered energy, doesn't need to worry about this unless 25mm canister rounds start flying.

Thursday, February 16, 2006

CAN: Universal bus?

Well, I'm still learning gEDA. Finally got it loaded on a PC that I actually use. I got in my head a slightly modified version of my RC CAN interface. Instead of using a CAN enabled RC controller, how about a CAN enabled smart digital motor? I could then control the PID and rate control, increase accuracy, and have more information and control. So, this board should be designed to interface with the mechanics of the standard RC servo and replace it. So, I'm considering the following capabilities (no particular order):
  • H bridge based motor control
  • Temperature measurement
  • Input voltage measurement
  • Module amperage measurement
  • Programmable potentiometer for maximum position measurement
The above would allow for shutdown of the unit before motor burnout, power efficiency calculations and power usage control. Each board would be designed and built and made to support as large a range of servos as possible, hopefully allowing wide ranging retrofit. Target manufacturer is currently Hitec, although if a low cost, good quality servo is found it will be considered.

Offboard support circuitry:
  • High current buck/boost controller: provides servos and modules with approximately 5V of regulated voltage, several amps of power (I hope). Keeps the electronics and servos electrically isolated from the main motor voltage spikes.

Now, onboard circuitry, preliminary part selections:
  • PIC18F2580 (Microchip, QFN): integrated CAN, smallest board space for a CAN chip
  • MCP1252-33x50 (Microchip, MSOP): Power regulator, will provide 5V 120mA for 2 to 5.5V input, may be replaced if a good alternative for 2.5V to 10V is found and also will phase out the buck/boost converter from some designs. Isolates electronics from servo voltage spikes.
  • MCP2551 (Microchip, SOIC): CAN voltage converter (Requires 5V)
  • MCP6S21 (Microchip, MSOP): PGA w/ SPI controls for measurement of the amperage via high side shunt, may need an external SPI delta sigma ADC
  • MCP42100 (Microchip, TSSOP): Dual Digital Potentiometer w/ SPI controls. Used to scale line voltage to monitor raw 5V/high voltage line level, tune the high and low voltages of the position capture circuit to maximize accuracy.
  • MCP9800 (Microchip, SOT): I2C temperature sensor, due to lower rate, may be software driven
Note: space constraints will limit what can be implemented. The minimum design will include the PIC, voltage regulator, and CAN chip.

Now, another objective is using the Kodak 9630 camera chip and building a camera around it. It needs a 10MHz clock and 3.3V of power (sounds perfect for the MCP1252 above!). I'd configure it with the outputs buffered to work with whatever the logic is of the host processor. What's so special?

  • 8 bit B&W
  • 126x98 pixels (approximately 12K of RAM required)
  • Single shot and video modes
  • 10MHz SPI (~90FPS) or 10MHz 8 bit parallel (~580FPS), plus H sync and V sync pins
  • 16 pin interface expected (8 data, 2 I2C, ground, voltage, Vsync, Hsync, clock out, reset). Reset is also tied to the Power Good input of the camera.
  • Possible variant: Replace the clock module with a crystal and PIC18, allow for slight variance in clock, some onboard processing, power control of an auto-focus/zoom lens onboard.