Prototype See document: ffc8.jpg See document: ffc21.jpg 1. Hardware 1.1 Robot Base As mentioned in the Analysis section, the base is 6" x 6" Lynxmotion Carpetrover mobile platform powered by a pair of Hitec HS-300B servos modified for continuous rotation. One point missed by all the servo modifiacation procedures is that you must transpose the motor leads inside the Left servo for it to be programmable in the same way as the right motor. This is because they are mounted transversely to each other. Stan Kusmider designed a second 6" x 6" platform and custom 5 11/16" threaded stand-offs to support it above the base, as a "second story addition" to our robot. This was needed space for mounting our homebrew flame-extinguishing fan at approximately 7 3/4" above the ground (candlestick height) and ample room for the Parallax Basic Stamp II carrier board extra batteries, flame sensors, and an electronic compass. The Lynxmotion "First-Step" Basic Stamp I controller, the Sensor Board, the four Object detection IR Receivers, along with wheel encoders, batteries, and servos took up all available space on the lower deck. Whenever any mechanical mounting method was needed, Stan took care of it. 1.2 Sensors 1.2.1 Flame Detector The photo devices from the Electronic Goldmine arrived mid-March, about 1 month from the Trinity Fire-Fighting Robot Contest date. With little time to spare, they were experimentally tested in the Ward College lab, and within a few days, it was determined one set of devices performed better than all the rest in our IR-noisy environment (energy-saving electronic ballast flourescent lighting). These were unmarked Honeywell IR phototransistors in a tiny black rectangular package. The Electronic Goldmine part# for these is G9412. Even with a reasonably high value collector resistor of 27K, the overhead lighting did not even begin to turn on the device. Yet a small flame at a foot or more distance easily saturated the collector. The only problem was the omni-directional pickup pattern: it would pick up a flame in almost any direction, making it impossible to determine what direction it was coming from. Some experimentation with flashlight type silvered reflectors was tried. This idea was discarded when it became apparent the increased optical gain was too much, and had higher gain at visible 500-750 nM. light wavelengths than at the desired 900-1000 nM. Infra-Red region. While scouring the junkboxes, a better solution for a narrow aperture housing was seen: Black plastic Waldom WN-10 Wire Connectors for twist splicing AWG-10 house wiring. After removing the springs from several of these, 2 small holes were drilled to allow the phototransistor leads to exit, and the devices were glued about halfway into these "housings". Three were mounted on the underside of the robot's upper deck at approximately 7.5" height above ground using tie-wraps and adhesive backed clamps, facing to the front, right, and left sides at a slight downtilt angle. Rather than run the phototransistor outputs directly into the Basic Stamp II, it was decided to add comparator circuitry with a small amount of hysteresis. A threshold reference of 1/2 Vcc would be employed, which should reduce false triggering. See document: robot2.gif 1.2.2 Wheel Optical Encoders The circular paper disks colored with 8 sets of alternating flat black and flat white color bands were printed on a laser printer and glued to each tires inside surface. 1.2.3 Sensor Board The board containing the IR Object detector circuitry will now be called the "Sensor Board". It was removed from the robot and additional circuitry for controlling the flame sensors was added. This included "ambient noise" reduction by processing each signal through a comparator stage. A spare 4-1 multiplexor was used to "steer" any flame detector output to a single pin on the Basic Stamp II (P1). Also, this circuit was wired such that the selected flame sensor would always be OPPOSITE the selected object IR sensor. This is so our own Object detector IR LEDS would not generate interference to a flame sensor on the same side of the robot. A 5V coil reed relay in a 14 pin DIP was added to allow the Basic Stamp II to control turning the Fan on or off via I/O pin P11. Here is the complete schematic of the robot's Sensor Board: See document: robot.gif If you'd like to print the schematic, here it is in 2 parts: See document: robot1.gif See document: robot2.gif Here is the complete Circuit Board layout: See document: robot4.gif Here is the Bill of Materials: See document: robot.bom Here is the Robot BS2 Wirelist: See document: robotwir.txt 1.2.4 Electronic Compass An Dinsmore 1490 electronic compass sensor was wired on a small circuit board. It has 4 active low outputs related to compass points N E S W with some overlap so an additional 4 directions NE SE SW NW may be intpolated. The 4 open collector outputs were pulled up to +5v by 22K pullup resistors and routed to the Basic Stamp II I/O pins P4,P5,P6,P7. See document: robot3.gif View the next section "Testing" to see how this all worked out.
Back to Robot Main Page