DEWBOT X Ball Handling

From DEW Robotics
Jump to: navigation, search
Loading ball in autonomous
The team designed a very effective Roller Frame to control ball movement. While originally envisioned primarily for the purpose of picking the ball up off of the floor, the finished system performed the following functions:
  • Rapidly gaining control of the ball on the floor and loading it into the robot's shooter
  • Holding the ball securely in the shooter during aggressive play and maneuvers
  • "Packing" the ball into the shooter to facilitate accurate shooting
  • Rolling the ball out of the shooter and passing to an alliance robot
  • Passing directly to an alliance robot as an inbounder
  • Rolling the ball out of the shooter and scoring into the low goal
  • Loading the ball into the shooter at the start of autonomous
  • Facilitating catching the ball - from a truss shot (rare) or from a human player (very common)

A lot of utility for a single, simple mechanism! In fact, the Roller Frame was used a lot during competition; far more frequently than anticipated. This led to pneumatic system performance issues which became apparent at Hatboro-Horsham.

As the only part of DEWBOT X to extend beyond the frame perimeter, the Roller Frame was also subject to an enormous amount of physical abuse. The design had to accommodate this.

While the Roller Frame itself is a simple mechanism, a great deal of mind-time went into the details of its design; especially in terms of materials selection, roller drive, specific geometry and mounting arrangements. Minding those details ended up making a huge difference, especially in reliability & durability.


Roller Frame & Boot
The system consists of two components: the Roller Frame and the Boot.

The Roller Frame is a U-frame open and anchored to the chassis with a pivot joint at the bottom. There are three main structural elements: the sides and the bottom (at the top) of the "U". These three structural elements are connected to each other via (4) bolted 6061 Al gussets. One of these gussets (left-rear) also serves as a motor mount. When retracted, the Roller Frame is superimposed outboard of the shooter frame.

The Roller Frame is anchored to the chassis weldment via a custom welded 3/8" steel roller pin secured with an e-clip. The roller pin passes through a custom pillow block with a bronze bushing securely bolted to the chassis weldment (via ¼"-20 SHCS's and rivet nuts). As originally built, the pillow block was a purchased COTS item (McMaster-Carr 5912K3), but the sintered aluminum proved to be brittle and failed repeatedly during testing. The COTS pillow block was replaced with a CNC-machined 6061 Al block which has held up to competition service.

A pair of 1-1/16" bore x 6" stroke Bimba cylinders drive the Roller Frame's articulation. The cylinders' connections are rotated 90° to keep the fittings inside the frame perimeter and prevent interference with the shooter frame. The cylinders' clevises are replaced with ball-joint rod ends to allow extra flexibility in the assembly.

The business end of the Roller Frame is the roller, of course. The roller is a ½" keyed 4140 steel axle shaft with four wheels: (2) outer 6" AndyMark old kitbot wheels; and (2) inner 3.25" VexPro VersaWheels. Hubs, spacers and clamp collars kept all wheels in position. A 7075 Al shaft was tested, but this proved to be too flexible. The roller axle is mounted on two ball-joint bearing blocks bolted to the front face of the U-frame. These bearing blocks allow the u-frame to deform without bending the roller axle.

The roller is driven via type 35 chain and sprockets by a AndyMark 9015 motor with a 2-stage 16:1 BaneBots P60 planetary gearbox. A spring clutch between the gearbox shaft and the 15T drive sprocket allows the roller to stall if jammed, thereby preventing motor burnout. As of this writing (4-May), only one 9015 motor was smoked in this service and that was during testing, not competition.

For personnel and motor safety, the chain and motor are protected by a polypropylene guard.

The Boot is a formed piece of polypropylene attached to both the chassis and shooter weldments which forms a ramp for both loading and unloading the ball. The Boot helps keep the ball in the shooter.


Filling pultrusion ends with carbon fiber/epoxy composite (two-by-two; hands of blue)
More than any other part of the robot (save perhaps bumpers), the Roller Frame was absolutely in harm's way during competition. In spite of this, as of this writing, it has sustained no damage other than a loss of some paint. How was this accomplished? By design, of course. And in this case, design began with materials selection.

The three main structural elements making up the Roller Frame are lengths of 1" x 1/8" wall fiberglass pultrusion square tubing. Fiberglass pultrusion is very tough and flexible. It has high tensile strength along its major (glass fiber) axis. It deforms under load but recovers elastically when that load is removed.

But pultrusion tends to crush (inelastically) under heavy through-bolt (cross-axis) compressive loading. To avoid this crushing, the ends of the pultrusion elements where they were connected via gussets and where the roller bearing blocks were through-bolted were filled with a carbon fiber/epoxy composite prior to machining. This composite filler then took the compressive load of the gusset and bearing block through bolts, not the pultrusion walls. 1640 had first employed this construction with DEWBOT VII's arm.

At the bottom ends of the Roller Frame, 3/4" x 1/16" wall steel square tubes were machined at one end to serve as pivot joints and then inserted (snugly) into the open pultrusion ends and bolted into place. At this end of the pultrusion, the steel square tube inserts prevented through-bolt compressive collapse.

As mentioned above, the sintered aluminum COTS pivot pillow blocked failed rapidly during testing and were replaced with CNC machined 6061 Al blocks with the same critical geometry. Bronze bushings were used here, together with a custom welded 3/8" steel roller pin.

Drive & Clutch

Roller drive clutch
An AndyMark 9015 motor mounted to a 2-stage 16:1 BaneBots planetary gearbox. This drove the Roller via type 35 chain & sprockets (15T drive and 22T driven), yielding an overall gear reduction of 23.47:1.

Under normal circumstances, this arrangement would have brought an enormous risk of motor failure due to jams. 9015's lack thermal protection and rapidly burn out if stopped while driven. Any sort of roller mechanism (especially with these big balls which may be captured well off-axis) is subject to jamming.

Sab-BOT-age addressed this by fitting a home-grown, spring-loaded clutch to the P60 gearbox output. When torque exceeded the clutch setpoint, the clutch would slip, saving the motor. As soon as torque dropped to below the setpoint, the clutch re-engaged. Simple, but effective design.


  • Really the team's first truly successful roller capture mechanism (kinda' sad, actually, for a ten year old team)
  • Effective - if we touched the ball, we got it!
  • Durability (no repairs needed during competition season)
  • Reliability (no failures)
  • Really dug the balls out in tough defense
    CNC machined bushing block and pivoting connection
  • Synergy with pivot drive
  • Far more useful than originally intended
  • With the Roller Frame closed (retracted), the ball was secure and could not be removed (we were able to drive through a demolition derby without risk of losing the ball)
  • System started and ended each match de-energized and safe
  • We smoked a total of one 9015 motor, and that was during testing/practice with the clutch set tight. Never lost function during competition.

Could have done better (in retrospect)

  • Our failure to appreciate ahead of time how often the Roller Frame would be pneumatically deployed during real competition led directly to a crisis at Hatboro-Horsham. Our shooter failed to shoot. Pre-competition testing and practice had failed to reveal this problem. This turned out to be a result of low pneumatic system pressure (the shooter release being the first system to fail to function due to low pneumatic pressure). Following Hatboro-Horsham, we were able to correctly diagnose the problem's cause and augmented the pneumatic system's capacity to meet the actual competition needs.
  • We were really slow to correctly diagnose the Hatboro-Horsham shooter failure, blocking an effective recovery during this event
  • We could only shoot by extending the Roller Frame first, which gave defenders an opportunity to shut us down
  • Energizing and loading the ball at the start of the short 10s autonomous period took a significant portion of this time
  • Clutch design is effective, but can probably be improved upon
  • Should have used a modified (drive-train steering motor) EZ-change bolt arrangement on the Roller Drive motor - without this, replacing the motor would have been very difficult in competition


  • Senior student Lucy designed this system in SolidWorks following prototyping
  • Senior Mentor Gary Deaver brought this composite technology to the team
  • Senior Mentor Ben Kellom and sponsor Edgetech turned around the CNC pillow blocks in record time