1/12/2014 Day 8: Beginning of Week 2 and Exiting the Conceptual Design Phase

Today we were back to a full group, since the VEX Competition was finished and emergency work was taken care of the day before.  Our big decision today was whether or not to go with a puncher/jaws intake and shooter assembly, or to go with the intake roller that pivots with a catapult.  After reviewing the spreadsheet, the decision was inconclusive (slight majority for the independent systems, but not entirely conclusive), so the "Committee of Public Safety" (the three members, including me, and two mentors the team voted on to become the official tie breakers) voted 4-1 (split mentors, unanimous students) to go with keeping the intake roller and shooter/launcher separate.  As we decided on the catapult, we split back into our original groups, so I went back to work on the catcher.

Shooter

Today a group of 4 students and 1 mentor designed the shooter, which would be a pneumatic catapult, rather than a surgical tubing catapult, due to the small amount of space a pneumatic (flat) catapult takes up (picture below), as opposed to the spring powered shooter, which would have a tower attached.

By the end of the day, drawings for all of the gusset plates and the shooter as a whole were finished, ready for manufacturing tomorrow.

Intake

Today the intake team figured out how to fold the intake system in and out, so it can start within the frame perimeter, but extend outward the 20" it needs to.  The intake basically just required several pieces of square tubing, ABS sewer pipe, two pistons (one for each side), and coarse plastic/paper material.  The coarse material on a plastic pipe roller has been shown to be very effective, especially with picking up somewhat compressible objects.

Drivetrain

Not a lot was done with the drivetrain today, and was mostly worked on by programmers to make it drive.

Catcher

Today was a major advancement for the catcher.  As we now have a general concept of how it can be actuated, we worked on possible material to use for it.  After thinking about netting for previous days, we took a slightly different approach.  As we had lots of conduit, we decided to play around with that.  As a bumpers discussion was going on previously, pool noodles quickly came to mind, as those are what are used for bumpers, so we slipped on a pool noodle on the conduit (shown below), and tested both conduit with pool noodles and without pool noodles.

As expected. the pool noodle worked well to dampen the impact from the ball, and much better than the conduit on its own.  This meant the idea with folding out sides would work, so that method of catching was now out of the conceptual design phase and well into the preliminary design phase.  The only other concept for catching was the folding out corners, which had to be prototyped a different day.

In addition to the new possible sides, the catching subteam worked on a frame that goes on top of the drive base, including the catcher.  Since there is not much space on the drive base to mount things, and it is preferred to not have to deal with mechanisms being mixed into electronics, the elevated frame is necessary as something to mount the subsystems onto, as well as keep everything above the drive base.  We designed the frame, with the main purposes of elevating everything and having something to mount things onto, and laser cut wooden gussets over the course of the day, as well as going out to buy parts (wood) for the frame.

Approximate times of working: 12:00 PM - 6:00 PM

End Of Week 1: Comparison Update

While the bulk of my posts are mostly about the what, and what happened so far, the blog and general project goal is not only how to build a robot, but rather the process being used for building a robot.  The goal of this project is to see and understand how the engineering design process functions and how the time factor makes an impact, and comparing the final product and process to past years, when we didn't use the engineering design process.  Because we currently have some time, but not lots of time, a lot of ideas, and solid machining ability, as this year we prepared everyone to use the now available machinery, we have been going in and out of the conceptual design and preliminary design phases using cheap parts (PVC, wood, etc.) for the different mechanisms, creating many ideas, then building and testing them.
The extra prototyping gives us alternate options before it gets too late in the season to change, in the case of something we thought would work turns out to not meet expectations and requirements.  While it does add extra work, this method of having "competing ideas" allows us to choose the prototypes and concepts that best reach and exceed the requirements and standards we made for ourselves and move on from there.  This, of course, gives a better chance for a more improved final product.
In comparison to the past years, we are much more prepared, as we have now eliminated any financial restrictions through heavily fundraising throughout the offseason, created a partnership with a growing organization that gives us access to a full machine shop for our purposes (3D printers, laser cutters, lathes, drill presses, table saws, etcetera), trained every member how to use any machinery, and showed examples of different mechanisms and structurally sound parts.  This means, the only major factor is the clock, as the other factors that may effect the build season are mostly taken care of.
In the past years, we only made at most one prototype scorer.  While we were somewhat lucky both worked, part of the luck was due to other teams posting their mechanisms, all giving us the concepts well into Week 2 and early Week 3, and spent most of Weeks 1 and 2 discussing strategy and learning how to use power tools.  With the ability now to quickly machine basic and cheap parts, we can be one of the teams that releases successful prototypes within the first couple weeks, as we have so far done.
As a summary of the conclusion to Week 1, the engineering design process has been much better than past years, and time has not been much of a factor.

1/11/2014 Day 7: A Smaller Team

Today was primarily a day for programmers.  As today was the VEX Robotics Regional Competition in Sahuarita, a solid chunk of the team was missing (6 members and 1 mentor), and another two gone for the weekend.  This left one of the mentors to work with the mass of programmers (almost everyone at today's meeting), and the other to work with the couple mechanical kids left.  As the elastic puncher was not yet eliminated from the jaws idea, that is what was worked on today, just using a PVC pipe, a 2' long piece of 2"x4", and lots of elastic tubing.  The puncher worked very similarly to a bow and arrow, but with the arrow being larger and attached to the bow.  This means when the bow/surgical tubing was pulled back, the puncher was "loaded" and ready to shoot, and the surgical tubing would be let go, thrusting the PVC pipe forward, punching the ball.
By the end of the day, the puncher was finished and tested, but was considerably weaker than the catapults.

1/10/2014 Day 6: Continued Integration Discussion

Today was similar to yesterday, but we began to optimize different methods for our intake and shooter mechanism/created a spreadsheet of important qualities of an intake and a shooter.  With this, we made a conceptual rank/effectiveness score for each differently styled mechanism for each quality.  The main question was whether or not to go with it all in one assembly/unit.  If it were one assembly, the passing would be precise and easy, but the shooting would be less powerful and inconsistent, as it would be the normal aiming, but with a puncher, not a catapult, which, again, we figured wouldn't be as effective.  With an intake roller separate from the catapult, it would be slightly less controlled (passing), but shooting would be great, whether a pneumatic catapult or elastic catapult is used.
The single unit would be called jaws, since it looks like they could be jaws, in that there are two crossing structures, the one crossing to the top with rollers, the other static and used as a track for the ball, and a puncher by where they intersect, so the shooter pivots with the intake.  On the other hand, there is the assembly where the shooter and intake mechanisms are independent of each other.

Approximate times of working: 3:30 PM -7:30 PM

1/9/2014 Day 5: Integration

Today was a bit of a slow down day, just to ensure everyone is on the same pages and the different mechanisms are still compatible with each other.  Instead of continuing the preliminary design phase for each of the individual mechanisms, we went back to the conceptual design phase for the robot and it as a whole, something we hadn't done yet.  After spending the day in small groups and conceptualizing, the general wants were all the same, a roller intake, a shooter/catapult, and a fold-out catcher.  This meeting, again, was primarily for getting the group on the same page, and was beginning to mark the end of the conceptual design process.

Approximate times of working: 4:00 PM -7:30 PM

1/8/2014 Day 4: Success!!!

Shooter

Today we built our first prototype catapult (that actually worked!!!), which can be seen in this video.  After figuring out the shape of the "flinger" was the problem, we changed from the bent chair legs to just a piece of conduit bent twice.  Both "flingers" can be seen in the video, the conduit mounted on the cart being used and the chair legs sitting on the bottom of the cart.  The biggest problem was the general geometry.  The flinger has a constant angular acceleration, but that translates to linear acceleration based on how far away the ball is from the pivot point.  Because the chair legs are bent at half the length, the end of the chair legs prevent the ball from getting to the back of the chair legs, which would be the same as using a straight "flinger" that is half the length, meaning half the linear acceleration. This is the main reason why the conduit works so well, since the ball can get to the back of the conduit, meaning larger linear acceleration for the ball since it is further away from the pivot point.

This was found after an increase in tension with the chair legs had no significant affect on the ball's flight.  This change made the ball go from being shot 3' high and 5' far to 7' high and 25' far, with the same amount of tension.  This was now a proven design, although we did want to see whether or not an air powered launcher/puncher would work better or be more compact.

It is safe to say this design of the shooter has now moved out of the conceptual design and is in the middle of preliminary design.  We have now shown the concept works and works well, and are in prototyping phases and have made basic CAD models and other drawings.

Intake

Today the intake group looked at ways to build an intake, and have now decided to use information we found from the Robot in 3 Days groups, and past intakes we have built.  As expected, roller intakes appeared to work, so the intake team began building a roller intake as a prototype, powered by a drill motor. This was much more desirable than a claw due to the speed of the intake and outake.

Drivetrain

Today we began to get the drivetrain working, but it was primarily modifying software, rather than making hardware changes.

Catcher

Today I created a CAD model for a potential catcher, as we came up with an idea to use a funnel.  Since robots have to start within their own frame perimeter, but can expand outward after the start of the match. The catcher is similar to a cardboard box with 40" flaps that extend 20" outward.

We had two ideas for actuation (moving the stuff), either directly controlling the sides using pistons, or make the edges like levers, and actuate the levers with either pistons or motors.  After discussing both, we made prototypes of each, the first actually using a cardboard box, and the second using PVC pipe.

Now, most individual mechanisms were out of the conceptual design phase, and into the preliminary design phase.

Approximate times of working: 3:30 PM -7:30 PM

1/7/2014 Day 3: First Day Off

This year, my weekends will be on Tuesdays and Tuesdays only (will most likely do work on those days also anyways).  As it is Day 3, all (5) of the "Robot in 3 Days" teams released their release videos of their robots. The main robot I looked into was Team O-RYON and their robot "Draconis".  As expected, they used a roller intake to pick up the balls, which was effective.  Unlike our thought's though, they decided to use some sort of kicker or puncher, which fired, but not overpoweringly.  This helped show us a catapult is definitely the way to go.  Also, the robot didn't have a catching device, so it required a very accurate throw from either another robot or a human player.

1/6/2014 Day 2: On to the Next Steps, Time to Apply Strategy

Today (first day after school) was the beginning of the splitting into subgroups, which are shooter/launcher, intake, drivetrain, and catcher.  Each are definitely at different stages, but all still going according to the Engineering Design Process.

Shooter

After much discussion involving the shooter and playing with the ball, a couple concepts were thought of.  As there was a similar game in 2008, we mostly pulled from the knowledge and past experience of the mentors, who were on a team in 2008, as well as videos from robots in 2008.  As seen by punching the ball in the air, the ball could be launched with a puncher, but only for a fairly short range.  Instead of a puncher though, the shooter group began prototyping a spring powered shooter using our old robot cart, surgical tubing, a small pole, hose clamps, and some 2"x4".



By the end of the day, we finished and launched twice, but both were ineffective and fairly weak, although this was most likely due to the lack of tension in the surgical tubing.  Overall, today was a very promising Day 2 for the shooter.  Day 4 will consist of adjustments to the shooter, and the shooter is now already out of conceptual design and burying itself into preliminary design.

Intake

Similar to the shooter, the intake was quickly leaving the conceptual design state, but was stuck in a debating state.  As we have worked with different types of intakes, the intake group had several concepts down, and it is more so a matter of the catapult group to finish first.  Similar to the other past designs, the two primary ideas are a roller intake and a claw intake.  The roller intake would need at least one motor to power it, while the claw would need at least two motors or three pistons to power it.



As roller intakes are simpler and almost immediate in the intake process, while a claw is a little slower, the idea of using a roller intake is being leaned toward by the intake team.  Other options are being considered as a "just in case" a different type of intake might have other advantages.  Also, we observed all of the "Robot in 3 Days" teams are using roller intakes for their bots, and are extremely effective and consistent in picking up the ball.  The biggest question, though, as brought up previously, is "What is most compatible with the shooter?"

Drivetrain

As we decided to use a basic tank drive, we plan on using a modified kit bot, 6 wheels and 6 CIM motors (simplest thing that meets and exceeds our requirements).  The size of the kit bot is 32"x32", which is too large, but the launcher is needed before modifying the drivetrain.  Although it will be modified, the drivetrain team finished putting together the large kit bot and started mounting the electronics to it, so we can test any prototypes we build.



Catching

While the other three are essential and absolutely required for the robot, the catching mechanism is what I decided to work on, due to a previous lack of concepts, other than increase surface size, and lack of interest on the catching mechanism from other members.  Today the catching team worked on possible ideas of catching shapes, with a funnel as the easiest, since it is passive and the most logical shape for catching and passively directing an object.  By the end of the day, we had a general idea of how it would work and be actuated, but was not yet sure what material it would be made of.

Approximate times of working: 3:30 PM -7:30 PM (Half hour break in between)

1/5/2014 Day 1: Strategy, Game Analysis, and Requirements Document

1-5-13

A more lengthy and more in depth document of Day 1.

Scoring:

            Low goal - 1pt

            High goal - 10pt

            Truss toss - 10pt

            Truss catch - 10pt

            Bonus:

                        1 Assist: 0

                        2 Assists: 10

                        3 Assists: 30

            Penalties:

                        Foul - 20pts

                        Technical - 50pts

            Autonomous:

                        Mobility bonus: +5

                        Any goal: +5

                        Hot goal: +5

Game notes

            Assists: # robots in possession of ball in unique zones

            Score all auto balls before cycles begin

            Up to 3 assists per cycle

            10 second autonomous

            2 minute, 20 second teleop

            20” outside frame perimeter after start of match

            Truss = 5’2” , top = 6’2”

            goal = 6’10” at bottom, 3’ tall, 11’ wide

            low goal = 7” lip, 29” up, 28” sides

            field = 24’8” x 54’

Robot:

            112” perimeter

            120lbs weight limit

            60” tall starting config

            6” cylinder above robot in goalie zone

            Bumper zone = 2”-10”

Technical fouls:

            Human reaching into field

            Possessing opponents ball

            Ball entered not from side of field

            Ball inbounded does not contact robot or carpet first

            Human player impeding opponents ball

            Ball grabbed or removed if pedestal not lit

            Ball inbounded if pedestal is not lit

            Impeding flow of match

            Drawing fouls

            Contacting robot or ball in contact with robot

            Touch OI during autonomous

            Intentionally picking up other robots

            Robot extending outside field

            Contact inside another robot’s frame perimeter

            Pinning

           

Strategerizing

            Overall game:

                        One robot cycles (1 assist)

                                    Pros:

don’t rely on other robots

other robots are free to play defense

less risk of ball loss

less scouting required

fastest cycle time

                                    Cons:

miss out on assist points

no assist points = no ranking points

tell other robots to not touch ball (rude)

no catch points

            Two robot cycles (2 assists)

                        Pros:

                                    Not missing out on assist points

                                    3rd robot can play defense

                                    Allows weak offensive partner to be useful

                                    Catch points

                                    assist points = ranking points, so higher seeding

                        Cons:

                                    lose ball, cycle takes longer

                                    not getting aaall of the assist points

                                    lots of assist practice

                        Three robot cycles (3 assists)

                                    Pros:

                                                max points possible per cycle

                                                whole alliance involved

                                                good for scouting

                                                assist points = ranking points, so higher seeding

                                                weaker robots can score points

                                                most flexible, can fall back to one or two robot cycles

                                    Cons:

                                                slow

                                                heavy reliance on partners

                                                high risk of ball loss

                                                lots of assist practice

                                                may lose assist points if takes too long

                        High goal scoring

                                    Pros:

                                                points! yay

                                                can also score truss points

                                                used in autonomous

                                                looks good :D

                                                pass to human player

                                                used for passing

                                    Cons:

                                                harder to achieve

                                                can lose ball more easily

                                                takes resources

                        Low goal scoring

                                    Pros:

                                                simple

                                                easy to make, less resources

                                                possibly more cycles?

                                                harder to miss, less risk if you do

                                                easily pass to other robots

                                                easy to adapt mechanisms to accomplish

                                    Cons:

                                                fewer points

                                                not cool bro : (

                        play goalie

                                    pros:

                                                block shots

                                                awesomesauce :D ie look like a pro

                                                possible in autonomous

                                    cons:

                                                6” wide blocker

                                                not scoring points

                                                limited application (only effective against high goal scoring)

                        play defense (pushy or fast)

                                    pros:

                                                preventing points from being scored

                                                prevent other teams from seeding high

                                                combined with other strategy, ie assists

                                    cons:

                                                not scoring

                                                not necessarily preventing scoring

                                                risk of fouls (pinning, etc)

                        Throw balls over truss

                                    pros:

                                                points

                                                simple with launcher, compatible with high scoring

                                                reliable, immediate points

                                    cons:

                                                need to catch or retrieve ball

                                                makes ball bounce or exit field : (

                                                may miss, would be bad

                                                slower than driving under

                        Catch thrown balls

                                    pros:

                                                points

                                                same mechanism for passing or receiving from human player

                                                compatible with launcher

                                                easy to build

                                                immediately in possession of ball (assist)

                                                avoid defensive robots

                                                looks cewl

                                    cons:

                                                very hard, better drivers or maneuverable drivetrain

                                                relies on alliance partner

                                                missed ball = bad

                        Fine control of ball on field

                                    pros:

                                                very good at assists

                                                improved cycle time

                                                easier to chase down ball

                                                cool

                                    cons:

                                                mucho practice-o

                                                harder to build, more resources

                                                reach out, therefore bigger target

                        Herding ball

                                    pros:

                                                its easy

                                                assisting

                                                less resources

                                                less risk of trapping

                                    cons:

                                                slower cycles

                                                everyone else can do it

                                                its easy - to lose control

                                                need agility, maneuverability, good drivers

                        Pass to other robots

                                    pros:

                                                assist points

                                                compatible with other mechanisms

                                                looks good :D

                                    cons:

                                                rely on partners

                                                takes time

                                                requires good drivers

                                                possibly lose control of ball

                                                requires mechanism

                                                potentially incompatible designs between robots

                                                bad for ego : (

                        Catch passed balls (floor)

                                    pros:

                                                same mechanism as floor pickup

                                                works well with other strategies

                                                can be the scorer

                                                good for ego : D

                                                relatively easy?

                                                very attractive feature

                                    cons:

                                                requires mechanism to gain control of ball

                                                kinda sorta a little required

                        Pass ball back to human player for easy assist

                                    pros:

                                                fast maneuver

                                                potential assist points

                                                more options (against defense)

                                                fairly easy maneuver

                                    cons:

                                                requires a launcher

                                                can only do in four locations

                                                risk of losing ball

                                                risk of penalties

                        zoned offense (passing zone to zone)

                                    pros:

                                                less robot movement

                                                simple strategy

                                                requires one good shooter

                                                fastest passing strategy

                                                compatible with defense

                                                easy with rookie teams

                                                compatible with most robots

                                                can be done by only two robots

                                                oh yeah. : D

                                    cons:

                                                risk of losing ball

                                                reliance on partners

                                                restrict drivers

                                                potentially easier to defend

Autonomous:

go for hot goal

            pros:

                        points

                        easier than most vision tasks

                        have programming resources

                        experience

                        cool : D

            cons:

                        requires vision

                        less points for amount of work

                        go for second ball

                                    pros:

                                                more points

                                                awesome

                                                have the resources

                                    cons:

                                                hard

                                                chase down missed ball before cycles begin

                                                limited time

                                                harder to get the hot goal twice ?

                                                harder to maintain position of ball

                        drive forward into zone

                                    pros:

                                                super, super easy

                                                points

                                    cons:

                                                anyone can do that

                                                requires a teensy bit of programming

                                                further from human players at end of auto

                        bulldoze into corner goal

                                    pros:

                                                mobility bonus

                                                easier than shooting

                                                less risk

                                                reduce risk of two ball auto (second ball)

                                    cons:

                                                fewer points

                                                only two goals, alliance partners might both do this

                                                makes two ball auto harder

                        Start in goalie zone, play defense

                                    pros:

awesome if works

close to human players

            can be easy to program

                                    cons:

                                                hard

                                                cannot score in auto

                                                harder to do it well

                                                cannot defend corner goals

                                                requires 6” paddle

                        don’t shoot, start with one less ball, start close to human player

                                    pros:

                                                less risk

                                                really, reaaaaally easy

                                                Cal gets fired

                                    cons:

                                                no ball, no shirt, no points.

                                                looks not cool : (

                                               

Want to do	Controversial	Do not want to do/No discussion
One robot cycles	play defense	play goalie
Two robot cycles	Catch thrown balls	go for second ball
Three robot cycles	Herding ball	Start in goalie zone, play defense
High goal scoring	Pass ball back to human player for easy assist	don’t shoot, start with one less ball, start close to human player
Low goal scoring	zoned offense	go for hot goal
Throw balls over truss		bulldoze into corner goal (autonomous)
Fine control of ball on field
Pass to other robots
Catch passed balls (floor)
drive forward into zone

Decided defense does not dictate the design

Catching thrown balls can be divided into two groups:

            catching from human players

            catching from other robots

Everyone is for catching from human player

            To what degree? close or further away?

            All want to catch from close-up

Catching from other robots will not drive design, is not a requirement

Herding balls

            herding is a feature, preferably using intake mechanism

Pass ball to human player

            nice to have, won’t drive design

Zoned offense

            possible design: intake on one side, spit out on other

                        not a design feature

Required Robot Attributes:

            Defense:

            Catching thrown balls:

           

                       

Drivetrain discussion

            Omnidirectional useful, but do not want

            Performance for standard skid-steer

                        Cal and Linus: speed over force

                        Alex: speed and force, two-speed shifter

                        Brittany: speed is good, force might be nice for defense

                        Gary: fast but with more torque

                        Danny: two-speed, shifted a lot last year, use better transmission, lower speed for finer driving

                        Alex: 3-CIM gearbox alternative to two-speed for speed and power with fine control

                        Gary: cheap and easy to add another CIM to each side of current drivetrain

                        Brittney: drive of CRUSH, two-speed very useful last year

                        David: functionally want speed AND force

                        Danny: putting 6 CIMs on drivetrain limits mechanisms

                       

Launching ball:

            Trajectory

                        flat as it enters goal

                        determined by whoever designs the darn thing

Spin

            Grant: spin probably not needed

            Cal: spin not necessary

            David: spin takes a lot of angular momentum

                        consensus: not needed

           

Robot Requirements

Type:

            Functional - what the robot needs to do

            Performance - how well the robot needs to do something

Priority:

            Must - cannot live without

            Desired - would like to have

Num.	Description	Priority	Performance
0	Win	Must	Yes
1	Drive	Must	Speed - ~10 actual fps Force - a lot
2	Omnidirectional	None	Does not exist
3	Launch ball	Must	Height - 100” Accuracy - 95%-100% Spin - no
4	Floor pickup	Must	Speed - <= 1sec Offset - ~50% ball diam. On-the-fly pickup - yes
5	Remove ball (no launch, no bounce)	Must	Time - <= 1 sec Speed - >= 5fps Control length - <= 50% ball diam.
6	Human player reception	Must	Distance - 5 to 10 feet Horizontal disp - +/- 8” Accuracy - 100%
7	Herding mechanism	Desired	Works. Not a driving feature