Assembly entry point: Drive Train Right With Gearbox.SLDASM
As we came back from Kickoff, we immediately started to formulate the game strategy for this year’s competition. Right away, we ruled out being a hopper-dependant robot. Early on in the match we figured the hoppers would either be purposefully activated to dump the balls, and there’d be a surplus of fuel on the game floor, regardless. So we created an Intake in order to pick up balls from the floor. We also made that intake vertical, transporting the balls in an elevator-like path into a ball reservoir We also knew we wanted to try and do gears for autonomous. That way, we could guarantee ourselves a significant amount of points during the autonomous period. We debated about being a low goal shooter, but in the end we decided to go for the high goal on the boiler, because of the value discrepancy between each type of goal. We thought of it this way, we would have to put 3x as many balls into the low boiler boiler goal for the same amount of points and 1 shot for the high boiler. With a little bit of fine tuning we could shoot for high goal and get more value for the same amount of time. In our design, one of the upperclassmen suggested having a ‘3-headed shooter’ that would not only go for high goal, but would shoot 3 balls at a time. This would make us highly efficient, and although we might not be as accurate, dishing out as many possible points as possible would lead for us to be desirable to other team’s strategies.
For gears, although we initially wanted a passive gear system, but soon turned into an active system of which two doors would push the gear out, onto its pin. We wanted to try and use the gear assembly in the autonomous period in order to rack up as many points as possible.
We ended up designing the climber later, and created a spool to coil up the rope, which was relatively narrow in respect to the rest of the back of the robot. To fix this, we implemented another system above the climber to move the rope into the narrower mouth of the climber. In terms of our drive train, we finalized on a torque-heavy robot to be hard to defend and block against, especially in the final rounds.
As we competed at NAU, we noticed something during competition. Gearbots, everywhere. Almost no robot had tried for fuel, opting instead to specialize in gear, and climbing, which is the most efficient way to score points. Although we had anticipated this, our strategy for fuel soon became too inconvenient in this Regional match. The way to be competitive was to race for gears, or block the other team from doing so. Since our team was built sturdily like a tank, we soon became the defensive robot in the game, as agile and faster robots would race around getting gears, and we would aid our alliance by hindering the opposing team. With some tricky and strategic driving, we would block one of the two areas around the DAVIT, and be an immovable wall to get around. This proved to get us far in the match, as we became the third pick for the 2nd highest ranked alliance in the finals.
However, going on from this competition, we realized a few things. Doing fuel is useless until you get to finals matches or the world competitions. Fuel is only viable when it will be the tie-breaker between teams with robots that are so good at gears, they can get 2-3 rotors turning on their own. So, our fuel strategy will be doing fuel in autonomous, or doing it simultaneously when putting gears into the gearpin closest to the boiler. That way, our shooting will be as efficient as possible. And with a revamped climber and gear mechanism, we should be competitive with the other robots in our next regional.
This year's fastener choice was heavily influenced by that of previous years with a few tweaks. We made the desicion to have less hardware diversity. This means that we are less likely to not have the correct size of something in case of breaking. For instance, last year, our climber was had a hook which was fastened to an arm with one 1" countersunk #10 made of special hardeded steel to be able to take the load. in competition, the load from the 120lb robot made the screw very dificult to repair when it came arround to routine arm maintenance. The screw was stripped due to someone useing a very small #2. When it came to needing a new screw, we didn't have any extra because of how specific the screw was, and no other teams around us has any. This made it very dificult to repair the arm while using a stripped screw. This year, we are using generic #10 screws almost exclusivly. This means that we can just bring a surplus of generic screws and we will never have any problems with specialized hardware breaking. Instead of ungodly ammounts of nylock nuts we used in previous years, we are increasing the frequency of rivnut on the robot. The use of rivnuts makes removing and attatching modules from the robot quicker because the nut doesn't need to be heald in place in order to unscew a fastener. One of the reasons that rivnuts were not used as frequently last year is due to their easy stip torque. In previous years, we used aluminum rivnuts, but by using a mix of aluminum and steel this year, we can have fasteners that support over tightened screwes.
*When Manufacturing, safety is our top priority. Proper PPE are training are of the upmost importance.*Our team manufactures most of the parts by hand. We use 1:1 drawings made by our design team in Solidworks. The manufacturing process beigns with the design team, they must develop a cutlist and drawing list with the correct number of parts that need to be manufactured. We then print out the 1:1 drawings, and cut the sheet metal to bounding rectangles of the size.
The drive train was built for versatility, robustness, and modularity. Design was heavily influenced by our manufacturing capabilities. Our primary manufacturing facility has tooling for sheet metal, so we designed most of the robot, including the drivetrain, in sheet metal. The decision was reinforced by the sheet metal design tools provided by Solidworks. We used 6061-T6 alloy 0.0625” sheets for a majority of parts.
We chose to use 8” OD pneumatic wheels for our robot this year. The wheels allow us to regulate the speed and traction by setting the pressure of the wheels.
[TODO: Insert tire presure image comparison]
We used these wheels for their high traction. This will provide a higher grip when being defended against. A six (6) wheel skid drive train is used; three (3) wheels on each side of the drivetrain. These wheels are belted together using HTD 15mm belts. Although belts are more prone to slipping, they are lighter than chain and gears. The weight of the wheels was a concern; all six (6) drive wheels weigh approximately 10.8 lbs. The wheels alone would be nine percent (9%) of the total robot weight, assuming a 120lb robot.
In the 2017 season, speed and agression is a must have for a drive train. The game is less about what can be accomplished and more about how efficient the task is accomplished. When designing the drive train, the motor power and reduction is one of the most important features of the drive train. The dirve power determines the speed of the robot, which is directly proportional to scoring rates, and how much your robot can push and avoid pushing. Pushing is most important in finals. When the motor is geared for torque over speed, the robot will be more consistant because it cannot be blocked as easily. This defense feature in tandom with high traction wheels make a robot that is harder to defend, which is a very good feature for the finals. In the beginning of the gearbox design, we were considering the possibility of the "two and a half cim" drive train, where each side of the skid drive is driven by two cims and a minicim. This would provide us an edge in the speed category. We were looking for a single-speed-single-reduction (SSSR) gearbox. The best fit for our needs was on the VEX Pro store.
The issue came when the highest reduction the gearbox supported was a 7:1. We did the math against the Tough-box mini's which we used in pervious years. If we were to use the 7:1 SSSR gearbox, we would lose seven pounds (7lbs) of pushing power, and only increase our speed by 30% in a perfect world. This was an unacceptable loss of torque. The SSSR would be heavier and harder to do finely controlled programming for. Because of this, and the overall cost of ordering four (4) new gearboxes for both the robot, and the practice robot, we decided to go with the Tough Box Mini's. The mini's are a tried and true gearbox with built in encoders and we had enough in supply for both robots.
Early on in the game strategy process we decided we wanted to be a shooter robot. Our source of fuel was to come from the floor, as the hoppers are expected to be depleted early in the match . The intake on the robot is located in the center between drive trains, shifted back a little over a foot towards the back of the robot. That way, we also have a reservoir between drivetrains where we can pick up and store a significant amount of balls without having them in the fuel tank. The frame is made of TS 6061 Aluminum, it houses two shafts which are tethered together with surgical tubing bands.
[TODO: Insert picture of banebots intake wheel pattern CAD]
The shafts include a pattern of T40 2- 3/8” and 1-3/8” Banebots wheels which alternate: The smaller diameter wheels sits in between two larger ones to create a slot of which the surgical tubing to go in. There is also a ‘C’ curved backplate in order to smoothly feed the fuel into our storage tank. We decided to use a 775 Pro motor, with a 10:1 gear ratio.
[TODO: Isometric view of Intake CAD]
Unique to many game strategies we have seen this year, we designed around being a ‘triple shooter’. Our shooter, in premise would shoot 3 balls at once, being more efficient when scoring into the high goal of the Boiler. So, we made a shooter that is similar in premise to the intake, being modular and a pattern of wheels. There are four identical pieces of sheet metal which end up creating 3 slots for fuel to be shot out of. We actually utilize a copy of the shaft with the Banebots wheel pattern (featured on our intake) to feed the shooter which is located towards the top of our robot. Our Shooter’s Intake is powered with a two 775 Pros into a dual gearbox with a 7:1 gearbox. The balls are then directed towards larger flywheels, which, turning at around 1500 rpm, hurl the ball towards the boiler. The back plates which support this system for the shape of an “S” Curve and the fuel is fed from the middle of the robot and the balls are shot from the front.
[TODO: insert cross section of shooter showing wheel pattern and s curve.]
The area of which we store the bulk of the fuel we intake, is the tank area between the Shooter and Intake. It was initially just empty space with two pieces of bent polycarbonate on the sides, but we soon ran into an issue: the fuel would create a lattice-like structure which would keep the balls from being fed to the shooter. We solved this by created a displacer, which uses the vibrations of a mini cim to unclog the tank area. We included a piece of steel as a counterweight, and attached the mini CIM under a ramp we already had mounted to feed the balls.
[TODO: insert picture of physical robot fuel tank, and vibrator]
[TODO: Insert picture of gear assembly]
Although we were initially hoping to make the gear system passive in the beginning of the season, we soon ran into a problem of how the inner rings of the gear could randomly position itself. In one of two configurations it would work, in the other it would not. So we used a servo to open 2 ‘doors’ and push the gear out. We also included a ramped top of which the gear could fall into. It’s made out of TS 6061 Sheet metal and we used TSX Servos. This is one of the systems of which we utilized the pixi, an imaging sensor which we use to tell our distance, position relative to the gear pin. The gear is mounted on some linear slides as it was initially going to move side to side, but ended up being static when it came to being manufactured.
One of our latter builds, the climber ended up having a limited space in the back of our robot, in between the CIMs of the Drive Train, and the Intake, already constructed on our final robot. So our Climber ended up having to fit in a smaller space than initially intended. We ended up with a climber concept that the rope would end up being coiled inside of a cut piece of tubing, as it would grab the rope and keep winding.This system has been dedicated its own CIM motor, as it is going to have to lift a 120+ lb. robot. We are supporting it with a structure also being used to mount the bumpers, and an added bellypan in the back to keep the robot from buckling, or folding in two.
However, our climber ‘winder’ ended up being very small, and we had to create a way to put the rope inside of it. So we created this slot that would slide back and forth and put the rope into the winder, enclosing it. But, this slot also had a really area of coverage in the back of the robot. We added to servos on top that have two arms on it, that will pivot towards the slot at the same time, pushing the robot with a armspan of around a foot, feeding into a 2 inch slot, increasing our range tremendously.
Throughout our qualifying matches we kept having problems with the reliability of our climber. Our system had relied too much on too many systems, happening one after another.Two servos would swing and bring the rope in, then a linear actuator to bring the rope into the winder; that winder then having to spin and coil the rope up. In at least one match we couldn't climb because the PWM cord for the climber’s linear actuator had gotten unplugged, rendering our climber useless. In retrospect, we had over-complicated the climber. Additionally, we had to adapt with our pilot, as well as our alliance partners, in order for the rope to be dropped at a reasonable time to climb, successfully touch the pressure plate, and get the corresponding points.
We noticed at NAU and on youtube and overwhelming number of teams were successful using a drum spinner, that included velcro taped onto it. People we using it to snag the rope, so it could be wound up the spinner. We took out the linear actuator, the railed system to hold the rope, and the rope's previous winder. We opted instead on filling the shaft with 1 7/8" Banebots T40 wheels. We've kept the linear guide system, however, as it has been a great asset and convenience for the driver, but in general, Climber v2 has a much more simplified design.
[To Do: Take pictures of the new Climber as of 3/22 and showcase the new parts of it.]