Build Timeline

[idahorobot]

Does anyone have a build timeline for this year? Specifically I am looking for a timeline that outlines what you expect to have done each week during the build season.


Thank You


Ed Katz

Team 2130

[Tim Baird]

We typically follow this:

week 1: brainstorm and mockup
week 2: finish brainstorming and mockups
week 3: finalize designs and start building
week 4: build
week 5: build
week 6: finish building, tweak, test, tweak, break, fix, ship.

week 7: check to see if the pets are still alive, ask significant others to take us back, do 6 weeks worth of make-up homework, and ask my boss for another extension.

[alex1699]

this is something from first..

http://www.usfirst.org/uploadedFiles...0Timeline3.pdf

[Billfred]

Historically, our first week involves little to no robot fabrication (we may get up to putting the kitbot frame together on the Saturday following Kickoff).

From there, the schedule gets fuzzy from year to year depending on the robot and game, mainly focused initially on mobility before any worrying about manipulators. We got burned on that last season; even with a marginally-functioning shoulder, we didn't have a good method of grabbing tubes to score them. While many veteran FIRSTers will argue that mobility should be your first priority, don't let it become your focus to the point that it's Week 5 and you're just now starting to think about manipulators. The first few attempts will almost always show issues ranging from pesky to show-stopping; doing that part early will allow you to have a much improved manipulator come the actual competition.

Or, as the now-defunct 1083 had as their motto, fail often to succeed faster.

[lukevanoort]

Originally Posted by Billfred

While many veteran FIRSTers will argue that mobility should be your first priority, don't let it become your focus to the point that it's Week 5 and you're just now starting to think about manipulators.

I would like to highlight this point. As I watched competitions and walked through the pits, one thing became clear. In most cases tube robots were not lacking because they lacked arm capability, it was because their gripper was not too good. Tubes either fell out, were hard to grasp in the first place, or were hard to align to the goal. (the worst cases being all three) Most of the arms were similar. There were lots of jointed arms and elevators out there, and very few arms were unique (heck, we had a pretty unusual design, but 1251 had essentially the exact same concept just with different proportions, motor selections, and execution). What differentiated the good, the bad, and the average was mostly in the grippers.

I think Billfred is too modest and doesn't present the gripper situation with his own team's bot as drastic enough. Our pit at Palmetto was next to 1618's and so I saw an awful lot of their robot (including the incident that earned it its name). The shoulder was fast, one of the fastest I have ever seen. I was constantly telling my team that 1618 was the dark horse of Palmetto; if they had had a working gripper, I think they would have been a very strong robot (#1 seed material, in fact). The drivetrain was very effective, the shoulder was very fast, but the gripper just didn't work. If I remember their stats from our scouting correctly, 1618 scored one (maybe two) ringers at Palmetto and I feel with a better gripper design that they could have been scoring 4 per match.

In contrast, consider 68, Truck Town Thunder. I never saw their robot in person, but I have seen a number of their matches on video. Their arm was a good deal slower than 1618's, but their gripper was very effective. They were a well above average scoring machine, not only at the Championship, but also at the toughest event of the year, IRI. (of course, their drive system was solid too; nobody pushes Truck Town, nobody (except 177))

The drivetrain determines if you can get to the goal, the arm if you can reach it, but the gripper determines if you can even get the game piece in the first place, much less score it. Don't shortchange the gripper.

[MrForbes]

Originally Posted by lukevanoort

Don't shortchange the gripper.

I'll second that, you need to have all the parts working...mobility probably is most important because a "flower pot" isn't much fun to watch, but if you intend to score, then you have to work on building and testing scoring gizmos right at the beginning and keep after it until you have one that really works. If you just think it'll work,and don't try it out until the last few days when the rest of the robot is ready for it, things might not go so well.

[AndyB]

I believe Andy Baker did a nice presentation on this in Atlanta a few years back for the championship. I'll see if I can find a link.

It seems we've always gone with this on 269.

Day 1

• Review Game
• Analyze Game by determining all available options to play the game
• Decide how we want to play the game

Week 1

• Individuals sketch ideas
• Mentors help break down the ideas and determine the possible success of each
• Decide on drive system
• We usually have a mentor that kind of settles into their part of the robot. For instance, we had a mentor that helped the kids with the lift system. We had one that helped with the elevator for tube placement, and we had two that really just focused on the drive system the whole time. Of course we have our electrical and programming mentors as well.
• Mockups
• Prototyping

Week 2

• Frame Design
• CAD Designs of Individual Components
• Optimization of Designs
• Constraining Each Component (Weight, Size)
• Continued Game Analysis
• Frame Manufacturing
• Drive Prototype Done

Week 3

• Part Manufacturing
• Drive Assembly (Depending on drive, could be later)
• Manipulator Assembly
• Refining Parts and CAD Drawings

Week 4

• Component Testing
• Robot Assembly
• First drive test
• Usually completely taking half the robot off and putting it in the dumpster and redesigning it. Usually the most crucial component...
• Driver practice near end of week if lucky

Week 5

• Programming optimization
• Programming of the individual components
• Driver Practice
• Fixing problems

Week 6

• Fixing
• Optimizing
• Practicing
• Shipping
• Finally, sleep... But probably not.

This definitly varies from year to year, but this seems to be the best starting point. Some years like last, we worked on the drive all the way into week 4 and really didn't finish assembling the robot and uploading the finished code until people were sitting there yelling at us that we needed to crate the thing.

[EricH]

You guys are fast! It only takes two weeks to build a robot, IF you have planned carefully. A typical season for 330:

Kickoff--see the game, read the rules, discuss with as many people as possible.
Week 1: During the weekend of Kickoff, simulate the game at full scale if at all possible. Analyze options. Figure out a winning strategy. (Repeat last step as necessary--If you can beat a winning strategy, you probably have a better winning strategy).

Week 2: Finish any loose ends from Week 1. If the requirements list isn't finalized yet, finalize it ASAP once the loose ends are tied up. Mockups are made and evaluated. Start design if possible.

Week 3: Design the thing. Get materials.

Week 4: Build.
Week 5: Build.

Week 6: Finish build and get to the programmers early. (At least, that's the plan--it has happened on occasion!)

End Week 6: SCRRF Preship Scrimmage.

Ship.

We do fudge it a bit, notably in 2005, when we cut our first metal at the end of Week 4 with the robot completely designed and an arm being evaluated, and 2007, when we had our strategy in about two days and a practice robot by Week 3--but part of the robot hadn't been designed yet (something about "that's a secondary function, wait for the primary to be complete first")! We still had the whole thing just about ready by ship.

[AndyB]

Huh? Are you referring to us? We try and get ours done in week 4, but it never works out that way...

If you're referring to the prototype, yes, we do try to get a prototype done by the end of week 2. You can see last years prototype below in my signature. (3rd Picture). Its linked to view the full size pic.

[EricH]

Originally Posted by AndyB

Huh? Are you referring to us? We try and get ours done in week 4, but it never works out that way...

Not specifically. You've actually got a pretty decent schedule. It seems that many teams try to have a design by the end of Week 1 or before. You can get away with this, but only if you know exactly what you are doing. (And do a good analysis of the game right after Kickoff.)

And prototypes do need quick completion for a more complete analysis. 330 has been known to have provision for two drive types up to Week 6 on the competition robot--we just planned for both and made room. One system was onboard, the other on Kitbot until we had finished testing. (Not recommended unless you are sure you know what both options will look like.)

[AndyB]

The two years I was involved with the mechanical team (the last two), we tried new drives that we had never done. (06 Treads, 07 Crab). That could explain the reason for us prototyping a drive earlier. Then we can usually mount anything else we need onto it.

[dtengineering]

A typical team 1346 build schedule:

Week 0: (Pre-Kickoff) Carefully plan build schedule so that the programmers and drivers get two weeks to practice. Promise to "keep it simpler this year" and "spend more time on design, less time in the shop".

Week 1: Discover what the game is. Ask "how the heck are we supposed to do THAT?" Throw promise to keep it simple out the window.

Week 2: Start building. "If we spend more time in the shop now, we might have a week for the programmers and drivers to practice."

Week 3: Start re-building. Usually something is moving around at this point that bears some resemblence to the final drive train. Non robotics students are amazed to discover that the shop suddenly has a carpeted floor. Positive karma and good feelings balanced by impending sense of doom when we realize that we're halfway through and have only done "the easy part".

Week 4: Finish first draft of design. CAD team learns what "as-built" drawings are, build team learns that sometimes at 9:30 at night when something just isn't working right, that teachers have a slightly larger vocabulary than they hear during class time.

Week 5: It is all together but doesn't work right. Promise self to never do this @$##@ robot project again. Embrace panic in order to banish despair and frustration. Discover the part that doesn't fit, but has to, that local suppliers are all out of much needed parts, and that we have exceeded the build budget by 20%. Realize that Valentine's day passed without notice. Ooops.

Week 6: Weigh robot. Take it all apart again for speed holing. Put it all back together again. Repeat as needed. Programmers come through with a miracle yet again. Apparently they haven't been sitting on their duff doing nothing for the past month, despite build team's previously expressed opinions to the contrary. Drivers demonstrate that the robot can, indeed, dent lockers and move large tables. The more tools and instruments on the table, the more likely it is to move.

Ship Day: Last minute panic to get everything into the crate and out the door. Much skipping of classes is involved. Remember important shipping documents and customs forms, but still manage to screw something up. Contemplate rest, but realize that Portland is a Week 1 regional and there are still details that need to be attended to.

Promise to do it "differently" next year.

Repeat as needed.

Jason

P.S. That was our "old" build schedule. We're going to keep it simple and be more organized this year so that the programmers and drivers can have two weeks to test and practice.

[Karthik]

Many teams have different philosophies when it comes to managing the build period. Before you can decide on a schedule, you need to evaluate the resources at your disposal. How often and how long is your shop going to be open? How many students and mentors will be participating? Are your suppliers local, or will you be shipping parts in? What's your team's level of engineering experience? Do you need to prototype extensively, or can you trust yourselves to start building sooner? These are just some of the many questions that need to be asked.

The following is a presentation on running a FIRST team, which has a section on managing the build season. A proposed timeline is given, which has been used with much success by many Canadian teams.

http://firstroboticscanada.org/site/...ps/runteam.pdf

It's an aggressive schedule, designed for use by low to medium resource team with a simple robot design. It also works well for a high resources team with a more complicated design. The schedule stresses finishing earlier, to allow more time for practice and tweaking. I've found that the amount of time spent practicing and iterating is directly proportional to a team's success. Teams who are less aggressive during build, often end up using their regional as practice and tweak time.¹ To me, that's an expensive way to spend $4000. (And if you're only going to one regional, like 75% of teams, this is not a good situation to be in.)

The design freeze is much earlier than most teams use, but this a conceptual design freeze. What that means is that you're settling on "double jointed arm vs. linear elevator" as opposed to the specific implementation. This can be determined during the prototyping phase, and modified throughout the practice and tweaking phase. The 1114 robots always undergo significant changes during the final two weeks.

I hope this helps.

1. In 2006, when we were testing our robot in week 5, we realized that our hopper had major issues with ball jams. We played around with it for a few days until we finally determined we needed a new subsystem, an agitator, to clear jams. The agitator worked like a charm, and was a major part of our regional success. At our regionals, we helped many teams design and build their own agitators. These teams discovered their ball jam issues during these regionals, because of a lack of practice and tweaking time. As a result, these jams drastically decreased their scoring ability. Just an example of how valuable tweaking time can be.

[Kyle]

Check out MOE's MOEmentum website. It has a time line for each week during build and even covers pre-season and post season.

[JBotAlan]

Originally Posted by lukevanoort

In contrast, consider 68, Truck Town Thunder. I never saw their robot in person, but I have seen a number of their matches on video. Their arm was a good deal slower than 1618's, but their gripper was very effective. They were a well above average scoring machine, not only at the Championship, but also at the toughest event of the year, IRI. (of course, their drive system was solid too; nobody pushes Truck Town, nobody (except 177))

Well, obviously you didn't see our performance at Great Lakes Regional...ugh...This gripper wasn't the most inspired idea... But I digress...

First 2 days: enter conference room. Eat. Talk. Draw. Eat some more. Talk some more. More drawing.

Mon-Sat of first week: get to it. Nothing at this point is set in stone; we had prototype ball flingers (two years ago) in our shop that we never used (2 curved arms side-by-side welded together, put the ball in it and fling it outward...the ball *shoots* off the end of that...up and over the cage wall and into a very expensive GM project... )
Build chassis, get it ready to be welded, etc. As soon as it's done, start building manipulator. Since I spend my time on the other side of the shop, I'm not so tuned into the progress...I'm trying to change that for this year.

Somewhere between 4-5weeks in I get the practice bot; this past season it was nothing more than the prior year's chassis, with the new arm mounted on top. If it hadn't broken so easily I would've been done on time...that's my story and I'm sticking to it...

The final bot is always just being wired on the night before ship date. There's always something stupid--wired backwards, not tight/loose enough, etc. that makes our first competition not so great...I hope to make my part better this year.

So testing and practice time are at the first regional usually. That, and on our practice bot, but the practice bot is so ghetto-rigged that we were wiggling wires to get the pneumatics to trigger...

That's about it.
JBot