Introduction      Mechanical System      Electrical System      Strategies
 

The drive system for Brannigan consists of two tank-treads powered by two independently controlled Pittman motors. Selection of wheels and tread material was the first and most fundamental process in designing the drive system. Since slipping between the wheel and the tread would be fatal, we investigated toothed tread systems. Pre-manufactured systems did not fit the specifications because they were either too light-duty or too expensive. Eventually we came upon timing belts as the treads and timing belt pulleys as the wheels. Timing belts and pulleys are generally used in industrial power transmission applications, but we decided to modify the belts and pulleys to drive our robot. We selected the XL series, a small pitch (1/5"), light duty belt system.

The pulleys are generally manufactured in plastic or steel. We opted for plastic because of cost, weight, and machining requirements. Since standard pulleys (in the XL series) have only one rib, we decided to place two pulleys together, creating one wide wheel that had two ribs. The ribs were important to prevent the belt from slipping off the side. Some machining was required on each pulley, namely removing the hub protrusion. This was done on a manual lathe.

The tread system has three wheels (each composed of two timing belt pulleys). The center wheel is the driving wheel, since it is connected to the motor. The two slave wheels are smaller in diameter and run freely. As shown in the picture above, there are two tensioners that provide a large area of contact between the belt and the driving wheel. If these tensioners were not present, slipping between the belt and the driving wheel would likely occur. Timing belt specifications, like those found in Machinery's Handbook, suggest at least six teeth in contact. The geometry of the belt path was governed by this criteria and the overall dimensions of the robot.

To maintain the dimensions of the placement of the wheels and tensioners, the housing of the tread system needed to be strong. It also needed to be protective, since the competition was head to head. Using a upside-down U-channel provided two rigid bearing surfaces via a through-hole in legs of the channels. We decided on aluminum 1 1/2" x 1/8" Wall Square U-channels for several reasons: 1. The excellent strength-to-weight ratio of aluminum coupled with the shape of channels provided complete coverage and dimensional stability with a reasonable weight. 2. Aluminum channel is extruded, unlike steel channel which is bent after being rolled. This provides: a. Sharp corners (we needed the space) and b. Better parallelism between the legs of the channels. Since the legs of the channels hold the shafts of the wheels, this was critical. 3. Machining the aluminum channel is easier since it is all one piece. Most holes in the channel were through-holes, which ensures that the axes line up. Machining two plates would require the holes to be matched up perfectly. Also, the channel could be remachined by rechucking it in the mill and touching off the original datums. 4. The two channels were the structural grounding of our robot. Because of the large moment of inertia of the shape of the channel, he forces required to deflect the channel any reasonable amount over its entire length were very high. A plate between the two channels was all that was needed to create a rigid and reliable robot.

We decided to use round head machine screws as the shafts of the wheels. This meant that the wheels had to be tapped. The slave shafts were 5/16"-18 and the driving shafts were 3/8"-16. The shafts ran on bronze sleeve bearings, which were also threaded. Generally, aluminum is not a good bearing material, but the relatively low speed of the system allowed for it. If friction in the drive system were more critical, we could have added ball bearings on the aluminum channel.

XL series belts come in two standard sizes: 1/4" and 3/8". This posed a problem, since we wanted a width of at least 3/4" to drive the robot. Another problem was that the ribs of the wheels protruded past the belt. Tackling both of these problems at once, we decided to create a belt composed of two timing belts covered with a strip of rubber. This raised the belt above the ribs of the wheels and tied the two belts together. We used industrial strength rubber cement - Pliobond, very strong stuff. The rubber selected was Adhesive-Grade Neoprene, which provided a prepared surface for gluing the belts on. We found that using harder rubber (Shore A: Firm) provided better rear resistance. However, by gluing on the rubber on the timing belts we raised some maintenance issues. After a little while, the rubber layer would start to peel off. This required us to replace the treads shortly before the competition and be aware of their state. If the rubber was starting to peel anywhere it was possible to reapply the glue and thus salvage the system. The most ideal solution to this problem would have been to purchase a timing belt that was higher and has a rubber backing as one piece. Unfortunately, such a timing belt is uncommon and cost would have been prohibitive in having belts custom made.

As mentioned above, the main function of the tension bolts was to provide contact between the belt and the driving wheel. But we soon realized that the amount tension of the drive system would be critical to the operation of the robot, we made the tension bolts run in slots rather than circular holes. Adjusting the tension altered the torque on the motor and gave us a way to ensure that Brannigan went straight when we wanted him to go straight. The tension bolts consisted of four parts: a bolt, a spacer (to prevent the channel legs from compressing), a runner (which was a tube of brass that allowed the tread to run over the bolt freely), and the nut.

Motor selection is one of the major decisions of any moving robot. We decided on DC motors, rather than stepper motors because we believed that they could provide the accuracy we needed without the complexity of controlling stepper motors. We also wanted to purchase a motor with an gear head that would provide the necessary speed reduction and torque increase. The first motor set we tried, while suitably priced, was not powerful enough. We then decided to purchase higher quality but more expensive motors from Pittman's Pittman Express Catalog (PN GM9234S022). The motors we chose have 38.3:1 gearhead reduction ratio, a maximum continuous torque of 187 oz*in (1.321 N*m), a stall torque of 911 oz*in (6.434 N*m), and a no load speed of 911 rpm. The motors were rated at 12 volts, had a resistance of 1.36 ohms, and had a torque constant of 3.29 oz*in/A (0.023 N*m/A). While these motors were a little larger and more expensive than we hoped for, the performance was excellent. They allowed our robot to move at a speed about 1 ft/s and had enough power to take whatever damage we could give it. Because of the power it supplied we were able to use journal bearings instead of rolling contact bearings and thus saved some costs and time on other parts of the whole system. If we went with these motors first a lot of time would have been saved and the project could have gone smoother.

Coupling the motor to the drive system turned out to be one of the major challenges of the project. The final system, shown above, utilized a set-screw shaft collar. The driving machine screw has two holes: one axial hole at the end of the bolt where the motor shaft enters and one transverse hole where the set screw of the shaft collar goes through. The set screw clamped not on the bolt, but on the flat face of the motor shaft. This coupled the motor shaft to the wheel shaft in a relatively simple way. It was also a very low profile solution. As shown below, the motors are large - fitting two of them inline proved to be quite a challenge. Notice that the bearing is backed up against the shaft collar. This prevented the bearing and wheel from unscrewing off the bolt during operation.

The above coupler, which was not used, took the role of both coupling the motor and acting as a bearing surface for the drive shaft. It had two transverse holes, one to hold the bolt and one to hold the motor. The axial hole of the coupler was stepped in order to hold both the bolt and the motor shaft rigidly. There were several problems with this coupler. 1. It is a relatively complex part to machine, with many operations and three tapped holes, requiring about two hours to make. Mistakes in machining these couplers or small design changes in the drive mechanism required them to be remachined, which put a strain on time. 2. The coupler allowed for no misalignment of shafts. Many manufactured couplers that can be purchased have rubber or plastic inserts for a good reason. If the shafts are misaligned slightly angularly or linearly, the insert will eat up the difference with little stress put on either shaft. A rigid coupler like the one above has to be machined precisely and does not allow for shaft misalignment. 3. The coupler was not efficient with space. Locating the shaft of the motor within the shaft of the drive shaft saved about 3/8", which was critical.

Introduction      Mechanical System      Electrical System      Strategies