DEWBOT VI Drive Train
- Its superior maneuverability. This should be especially valuable in aligning with soccer balls and with the goal.
- Safety while crossing the bump. A 6-wheeled drive (6wd) robot would tilt over further during the climb and would also experience two tipping points. The second of these (coming off the flat top of the bump) will put the robot at a considerable tipping risk unless center of mass (CoM) is very low.
DEWBOT VI is able to safely cross the bump and also drive through the tunnel. For both of these actions, the robot will need to drive in its long (x) axis direction. Chassis orientation is important for us in BREAKAWAY. The robot will not be capable of driving along the top of the bump. For more information, see the analyses Over the Hump (reprise) - An updated analysis of DEWBOT VI crossing the bump in "portrait" (x) direction and Running the Bump - An analysis of DEWBOT VI climbing the bump in "landscape" (y) orientation, then pivoting wheels to drive along the bump in "Portrait" (x) direction (This doesn't work). .
At the start of build season, a prototype pivot was built using the design developed for the hypothetical 4-wheel pivot drive-train. Following prototype testing, a number of modifications were made in the design of the pivot assembly prior to machining parts for DEWBOT VI's pivots. Rather more substantive changes had to be made to the pivot mounting plates due to necessary changes to the motor positions.
The pivot pivots on a 1" pivot tube, within which runs a coaxial drive shaft. Roller bearings at the top and bottom of the pivot tube bear the drive shaft's radial loads. The coaxial drive shaft terminates with a 16T-16 diametrical pitch Miter Gear. A thrust bearing bears the axial loads developed by the miter gear. The pivot tube pivots within bottom and top pivot plates secured to the chassis frame using (2) 1" ball bearing races and (1) 1½" ID thrust bearing. The top pivot plate also serves as motor mount for the drive and steering motors.
The coaxial miter gear drives an identical miter gear on the horizontal transfer shaft. A 9T drive sprocket also mounted on the transfer shaft drives the wheel via a type 35 chain. The transfer shaft uses (1) thrust bearing (to bear the miter gear axial load) and (2) 3/8" ball bearing races.
View a drawing of the pivot assembly, including the cross-section.
Wheels are AndyMark 4" Plaction Wheels with Roughtop tread.
Pivot bodies are machined 6061 aluminum. Machining was graciously provided by Wamac, Inc., a Downingtown Area Robotics sponsor.
- The CIM drives a 9T Type 35 sprocket on its shaft
- driving the Pivot Drive Chain
- which drives the 28T Pivot Drive Sprocket
- which turns the Coaxial Drive Shaft from its top
- turning a 16T Drive Miter Gear at its bottom
- which drives the 16T Transfer Shaft Miter Gear
- thereby turning the Transfer Shaft
- and the 9T Transfer Shaft Sprocket
- driving the Wheel Drive Chain
- which drives the Wheel via the attached 24T Wheel Drive Sprocket
There is no Gearbox. All reduction is accomplished using chains and sprockets (the Miter Gears are 1:1). This makes an unusual noise. Overall reduction is (28/9)x(24/9) = 8.30. Maximum robot speed is 9.8 ft/s.
All chains are Type 35 steel. Chain lengths are kept to a practical minimum.
Unlike tank drives, all CIM motors drive in the same direction, so there is no left-to-right mismatch between motors' clockwise and counterclockwise performances.
AndyMark 4" Plaction Wheels with Roughtop tread were selected due to their:
- low weight;
- reasonable cost;
- high friction coefficient of the Roughtop treads; and
- the ability to replace treads easily in-situ by splitting the wheel by loosening the (6) hub bolts.
We faced two challenges with these wheels in practice. Both of these challenged negated the easy-maintenance feature team team desired when purchasing the wheels. We worked out effective solutions to both of these challenges, which are documented here. The challenges were:
- The 4" wheels split by loosening the (6) hub bolts. This is unique to the 4" wheel; larger wheels have separate sets of bolts for splitting the wheel and the hub. Since we are also using the hub bolts to secure the drive sprocket to the wheel, the challenge was to work out a bolt/spacer/nut arrangement which would not interfere with the wheel splitting. Our solution, shown in the exploded sketch at right, uses male-female hex standoffs to replace the nulock nuts on the wheel. The sprocket then bolts to the standoffs, secured with the nylocks. The Standoffs hold the sprocket far enough from the wheel to avoid chain-wheel interference.
- The treads need to be secured. As-shipped, treads on wheels have their ends stapled together. This works, but requires pivot disassembly to change treads (not easy). This problem was solved with Gary Deaver's Tread Mounting System.
Safety - chainguards
All drive chains between CIMs and Pivots are therefore fully guarded by vacuum-formed polycarbonate chainguards.
- (2) 4-40 tapped holes are made in the sprocket face
- a 3-layer polycarbonate magnet holder is made with a magnetic encoder magnet secured in the middle 1/8" thick layer
- magnet and polycarbonate holder are secured via (2) 4-40 FHCSs
A Cherry AN8 series magnetic rotary position sensor is mounted about 2mm from the face of the magnet holder.
The Window Motors drive the Pivots via type 35 chain and a 15T (1" bore) driven sprocket on the Pivot Tube. Since Drive and driven sprockets have the same tooth count, measuring the drive sprocket angle provides a good surrogate for measuring the pivot angle.
Wheelbase is 28” (x) x 21.5” (y). Wheel pivot centers are inset 3" from the frame rear and side edges; 6" from the frame front edge. The extra front inset provides a "contained" area for balls being either kicked or herded.
The center-front of the chassis frame is dominated by the mounts necessary for the kicker and kicker cocking and release mechanisms. Because of the energy stored and released by the kicker, these mounts are necessarily quite robust.
Bumper mounting points are designed into the frame gussets. These bumper mount points are at the bumper zone elevation midline (13" elevation).
A great deal of attention was paid to keeping the robot's center-of-mass as low and central (relative to the wheelbase) as practical. The completed robot's center-of-mass, with bumpers and battery, is 17.5" from the aft edge of the frame (which is the dead-center of the wheelbase) and 9.75" above grade.
A multi-mode control approach will be employed. Supported control modes are:
- Absolute Crab
- Stationary rotation around wheelbase centerpoint
- Stationary rotation around ball
For more on Pivot control, see:
- Mathematical Analysis of the Pivot-Wheel System.
- Programming a Pivot Drive Robot.
- Pivot Worksheet - based on 0-5vdc Potentiometer Output.
- Pivot Worksheet - based on 0-1024 Potentiometer Output.
- Mathematical analysis - Crab with a Twist.
- Pivot Worksheet - Twist Mode 1.
- Pivot Worksheet - Twist Mode 2.
Xerox Creativity Award
- Team Sab-BOT-age and DEWBOT VI received the Xerox Creativity Award during the Philadelphia Regional for it's innovative 4-wheel independent pivot drive-train.
(BR)2 Engineering Excellence Award
- We received the (BR)2 Engineering Excellence Award due to our innovative drive-train and the cool mirror we mounted to find balls behind bumps.
We noticed that one of the wheels was showing extreme tread wear when compared to the others. The tread wear discussion page has all of the comments and ideas that would cause the wear and how we would test to see if that was the problem.