| Rick’s first purpose built Sumo robot was constructed during the summer of 2001. The primary purpose of that first Sumo was to test various theories on wheel camber and toe in/out. Also, this new Sumo was to be as small as possible to limit detection by other Robots. |
| The stiff chassis was constructed out of extruded aluminum parts glued together with aircraft quality structural epoxy. Four Tower Hobbies TS35MG servos provided the power. Three IR sensors embedded into the aluminum front bumper provided opponent detection and a Parallax BS2 provided the brains. The BS2 issued “Pulse Out” commands to two TS35MG electronics packages on the processor board. The output from the TS35MG electronics packages drove two FET H bridges via optical isolators. The TS35MG servos without electronics were just a very nicely packaged motor and gear train. Initial testing resulted in four smoked motors in the TS35MG servos. They were replaced with four Hitec HS925MG servos with their electronics removed. Seven Nickel-Metal Hydride batteries provided the power. About four pounds of lead was added so that the 3 Kilogram weight was achieved. Also, a handle on the top made catching the runaway robot easier. The speed of this robot was 17 inches per second. Since this robot was substantially smaller than the size limit, the name of “Runt” was adopted.
|
| Since the IR sensors include IR detectors in the front bumper, the addition of some software in the BS2 resulted in a Sony TV remote control start. The ten numbers on the TV remote control resulted in ten different start routines. This remote control start was only used once and prohibited at all other events that I have attended. |
| Tests were undertaken on the Exert-O’Meter and silicone coated tires. Obviously, clean tires are mandatory. Beyond that, toe changes did not result in a measurable advantage. It was discovered that having the tires square to the surface and having the weight distributed evenly side-to-side was very important. Careful adjustments with paper shims under the servo mounts was time consuming and paid big dividends. The 3 Kilogram Runt with silicone tires could easily push over ten pounds. |
| The February, 2002 Northeast Indiana Robot Games resulted in Runt winning the 3 Kilogram Sumo class. As a result of this competition, three changes were required. First, adjustable range sensors were mandatory. Runt’s sensors had 18 inches of range at home, but only six inches of range at the sun-lit competition. Second, another method of selecting the different start routines was mandated. Third, Runt was too slow to attempt any alternative attack strategy. |
| The adjustable range problem was resolved with an LM317LZ adjustable voltage regulator controlling the current through the IR LEDs. The voltage regulator gave a range adjustment of four inches to about three feet. The different start routines were resolved with two pushbuttons on the BS2. The right pushbutton reset the BS2 and commanded the start routine to the left. The left pushbutton incremented the different start patterns and commanded the start routine to the right. A Red/Green LED was added to indicate the start routine that was presently loaded and to show sensor status. Also, the BS2 was replaced by a BS2SX in order to improve sensor performance.
|
| Since the weight on the tires had such a major impact on push, some method of balancing Runt was needed. The solution was a rocking rear axle. The two rear drive servos were solidly mounted to an aluminum plate and the plate was mounted to the chassis via two grommets on the chassis center line. This allowed the “rear axle” to pivot slightly on the chassis center line. This change produced much more stable readings on the Exert-O’Meter. |
| At the August, 2002 Northeast Indiana Robot Games Runt easily won the three pound class. The only problem was a 500 gram Sumo! During one of the bouts we locked bumpers and spun into a draw. Surely some minor reprogramming could prevent this from happening again. |
| After the Robot Games, the pattern of the three IR sensors was drawn. The center sensor had a range of 16 inches. The two side sensors (about 20 degrees to either side) had a range of 14 to 15 inches. Between these three peaks, there were two major valleys where the range dropped to less than nine inches. Clearly, more that three sensors were needed to give an even detection pattern. |
| Runt was a few ounces too heavy for the 1 Kilogram class, so a major effort was undertaken to get it down to that weight. I figured that a few holes in the extruded aluminum chassis would same the weight. On the third hole, the drill bit caught on the soft aluminum, spun the chassis against the drill press and broke every epoxy joint. Runt was dead and I stopped playing with extruded aluminum.
No photos of Runt survived. |