 DC gearmotors are the next step for robot drives. They can be relatively expensive ($150 plus) if buying new parts, but that's what surplus houses ($20 to $30) are for, right? A sort of "Holy Grail" for surplus motor searches is the DC gearmotor with encoder. Additional things to check for include voltage, current, torque and rpm. You need to know how fast you want your robot to move and about how much it may weigh. Two feet per second is a fairly fast speed for most small robots. Many travel slower, only a few faster. If you specify a speed and a wheel size, you can figure out what the specs for a motor may be. For example, you have some 3" diameter wheels, you have a 12 VDC 8 AH battery, and you want your robot to travel 2 feet per second. 2 feet per second would mean 24" per second, and 3" wheels travel 9.42" per rotation. 24/9.42 is 2.5 revolutions per second or 153 rpm. You would need to find a 12 volt motor that runs about 150 rpm under some load. Load is approximated from how much your robot may weighs and the expected environemt it would move in. Perhaps your robot will run in a contest arena, with smooth floors. And with batteries, frame, motors and wheels, the robot weighs in at 2 lbs. Very minimal torque is required for smooth floors, which is another reason why the RC servo notors above are used in many robots. But perhaps there is a slight incline to climb, perhaps 15 degrees. To overcome that, you would calculate how gravity pulls the robot back through finding the backwards force that is part of the gravitational force. For 15 degrees, only 2 lbs times the sine of 15 is the backwards force or about a half a pound. This is applied at the edge of the 3 inch wheel, so the torque needed is about 1.55 pound inches or 24.85 ounce inches or 0.175 newton meters. Many standard RC servos are rated around 40-50 oz-in, so there is no real problem for a 15 degree grade. But what if the surface was carpet? More drag, more power needed. And the robot has tracks instead of wheels? Even more power. So you can see it's not a bad thing to "fudge a little" on the torque side and perhaps double the expected need for torque. Especially to get over any humps or bumps your robot may encounter. One ohter thing to consider is in sumo competitions. There, a robot must overcome it's own mass, and that of another robot fighting against it. So doubling the torque in this case may be too modest. Perhaps tripling or more might be in order. One other consideration is that DC motors also increse torque as they go faster, up until their maximum speed. In general though, many do not bother calculating out the torque requirements for a robot, and simply go with a motor that seems good. Two intial prototype robots of mine didn't work though, because of weak motors. One used a type of steering gearmotor that had a slipping clutch in the geartrain. It would simply spin the motor without much power getting to the wheels. Another was using directly driven stepper motors, which also turned out weak, partial because of the nature of steppers, and partially because of insufficient current supplied. A cheap way to get gear motors is to use those already in toys, specifically RC cars or even wired control vehicles. The above slipping clutch motors were from the steering mechanism of such a car, but didn't pan out. My sumo, however, uses other types of steering servos that work quite well. It's a bit of trial and error, and with persistence, it works. In getting bigger surplus motors from an electronic supply, you can find out many of the specifications discussed above from the manufacturers web sites. Pittman, MicroMo, and Escap have fairly complete specs on any many good robot motors. There's even a wheelchair repair cneter that sells surplus motors for those really big robots. DC gearmotors can be a great help in getting your robot moving as you want.
|
RC Servos have been widely used in small hobby robots. They have the drive electronics already on them and can be controlled by coded pulses from the controller. This is a form of pulse width modulation, except there is a wide variation in the frequency for the pulses to be generated in, other than that pulses occur around every 20 milliseconds or so. Pulses can occur as often as every 5 ms (200 Hz) or sometimes as long as 30 ms (33 Hz) apart. The pulses themselves control how the RC servos turn. Generally a pulse width of 1.5 ms centers the servo, with a 2 ms pulse moving the servo to a clockwise position, and a 1 ms pulse moving it counterclockwise to an opposing postition. Full travel of a servo is typically 180 degrees, requiring about 0.7 seconds to get from one side to the other. Special servos may have greater speed, or even be geared externally for greater travel. Use for continuous rotation in drive wheels requires that the servo be modified. The servo modifications might need a stop on a gear removed and a disconnection of the internal potentiometer from the gear train. In some cases, people have replaced the pot with 2 equivalent resistors in a voltage divider to simulate the center position of the servo. Better methods usually simply mechanically disconnect the pot, leaving it accessable for trimming the center position. Once you have a pair of servos modified, You attach the wheels, attach them to the base, make connections, and have your controller output a 1 ms pulse on one side while a 2 ms pulse drives the other. This works well, since one servo is flipped and must be pulsed in an opposing manner the other. RC servo driven robots usually weigh less than a pound, but can weigh up to 2 or more and still travle well on a smooth surface.
