[DFTF] The Classic Blunder...

[Joe Johnson]

This is part of a series of posts called Drinking From The Firehose on getting Dr Joe back up to speed on All Things FIRST.



Today's topic:
The Classic Blunder...

No, I didn't get involved in a land war in Asia. Nor did I go against a Sicilian when Death was on the line. No, my blunder was even more classic:

I designed a robot that SOME team could finish in less than 45 days, just not MY team.

As the picture above implies, I have been trying to make up for it by working myself to death for the last 3+ weeks, thinking I could set things aright by din of effort.

It has worked to some degree. Our chassis was looking awesome when I turned out the lights in The Box* this morning and headed to my day job. With more hard work between now and Tuesday Midnight, we will have a pretty amazing chassis, one that no rookie team has any business building: We'll cross the barrier with ease, we should be able to balance a Box of Rock on the Coopertition Bridge, we may even be able to help pull off a 3 robot balance after lunch on Saturday. It has other features as well that I won't go into (hint: we've named it, The Rook)... ...this sounds like a list of positives, but actually it works equally well as a list of negatives. You see... ... I am out of gas, we're almost out of time, and we've got nothing to handle balls... ...at all.

We've got weight for it. We've got space for it. We even have a reasonable concept or two we've sketched on various napkins over the last few weeks. What we don't have is TIME.

I can hear myself saying the words to my Rookies:

"It doesn't matter what Cheesy Poofs can build in 45 days, or Thunder Chickens, or Swamp Thing, or Miss Daisy, or WildStang, or CHAOS, or Beatty, or Buzz, or Pink, or ANYONE ELSE can build in 45 days...

...it only matters what WE can build... ...in 45 days... ...because that is the only option available to us."

I said it. I believed it. I meant it. I just didn't do it...

and there are only 106 hours left...

Joe J.

* Where else would you expect a team named Schrodinger's Cat to work?

[AlecMataloni]

Don't forget about that 30lbs withholding allowance. Try building a way to mount your ball handling mechanism in the time from now until stop build day. then, while waiting for competition, try to figure out how you're going to score. If it has to be a stationary dumper that requires you to be a fender scoring bot, then that's not a bad thing. just make sure you can get balls to place on your robot where an end effector can be used, and build it before you bag the robot. It sounds tough and it definitely will be, but what else do you have to lose?

[Tom Line]

Always remember, that 30 pounds withholding is huge. In 2009, we withheld our entire turret. In 2010, our lift system. In 2011, our minibot deployment system.

This year, I think we may get lucky and be able to bag an entire robot (for the first time since 2008).

Use that withholding allowance to your advantage.

[RogerR]

Originally Posted by Andy Baker

... but from my experience, it is really fun to play defense really well.

A big fuss has been made about making 3 point shots from across the arena, but no one has really talked about 'stealing' opponents missed shots. An agile, low CG robot that isn't slowed by the bump, and can quickly acquire missed shots from the opponents zone and ferry them across the field to friendly robots could provide a much more effective defens that a high traction box on wheels...

With the materials you have on hand, how hard would it be to create a minimalistic ball acquisition and storage system (a la 1902 in 2006)?

[wireties]

Yeah, the 30lbs is HUGE! We added enough length to our arm to use the 3rd row in between regionals last year. Concentrate on the drive train and the gatherer, fix the shooter and bring it to the regional. You'll sacrifice most of Thursday but it is better than nothing.

[Al Skierkiewicz]

Originally Posted by Joe Johnson

I can hear myself saying the words to my Rookies:

"It doesn't matter what Cheesy Poofs can build in 45 days, or Thunder Chickens, or Swamp Thing, or Miss Daisy, or WildStang, or CHAOS, or Beatty, or Buzz, or Pink, or ANYONE ELSE can build in 45 days...

...it only matters what WE can build... ...in 45 days... ...because that is the only option available to us."

Joe,
What makes you think we're finished building? We have a proto that moved onto the bridge last night. We think we see a light at the end of the tunnel but it could be that fatigue is putting up spots before our eyes. I will know more by tomorrow night I hope!

The reality here is you can prototype between stop build and your regional. Plan on what you can bring to manufacture and work like crazy on the practice day. I bet everyone of those teams you mentioned has taken their robot apart on the the practice day at least once having figured out that some parts won't work the way the game is played or the functionality is not needed due to a different strategy.

[Andy Baker]

Fellow Rookie,

We just got our powder coated parts back from the paint shop today, and we have all weekend to build. We do have a practice robot mostly finished, but we too have a bunch of work to do.

Keep in mind that you can walk in to the event with 30 pounds of custom parts.

Have a great weekend,
Andy

[thefro526]

Originally Posted by Andy Baker

Keep in mind that you can walk in to the event with 30 pounds of custom parts.

Added Emphasis here.

The 30lb withholding allowance does not apply to unmodified COTS parts. This means that you can with hold in excess of 30lbs of your robot assuming that you only have 30lbs of fabricated/custom/modified COTS parts (Andy's had me thinking about this for a bit). Not sure what you're looking into withholding, but depending on how it's built and how much is COTS you could withhold darn near half your robot - or at least I know that we could.

[sanddrag]

Joe, your situation is not unique to being a rookie team. On 968 one year, we shipped nothing more than a completed base, which we had not even driven. In 2008, the year we made it to Enstein, we did not ship it to its first event with a working ball mechanism. In fact, we had the last parts of the ball mechanism shipped to the hotel in San Diego, where they arrived at 9:00 AM on Thursday morning. Talk about cutting it close. There's a lot that can be done very quickly if you have a solid plan and the dedication and tools to implement it.

Currently, in 696's 12th year, we just now completed the base and bridge/bump crossing mechanism. It has not been driven yet. It looks stellar, but none of it is done or tested.

You're not alone.

[onecoolc]

I'm seconding what everyone else has said - 30lbs is enough to do a lot.

In 2010 we got to our regional, and our kicking mechanism... didn't work at all. It was a nightmare, and dangerous. So between the regional and Champs we completely redesigned the kicker from scratch, in a single unit that we could bolt into our chassis and plug in, and be done.

It's definitely do-able!

The fact that what you have done is solid is fantastic. Better to have one thing completely finished than two things half finished, correct?

[frasnow]

I have a question for all the people talking about the 30 pounds and how great it is. How can a robot truly be competitive if the electrical team doesn’t have time to test their wiring, the software team doesn’t have time to test autonomous routines, and the drivers don’t have time to practice? I’ve seen this type of thing occur many times at work, and it always delays projects. Mechanical engineers sometimes think the deadline for the project is their deadline and don’t leave any time for integration and debug.

Say you have a hypothetical team with very few resources, so they can only go to one regional and they can’t afford to build two robots. Let’s say the mechanical part of the team decides the shooter doesn’t need to be completed prior to bag-and-tag. They build a prototype shooter, but they never put it on a robot. The plan is to work on the shooter for a few weeks and attach it at the regional. It won’t go onto a robot until practice day, and the team will miss much of the practice session working on the build.

System integration is the hardest part of engineering. In the real world the game balls have different densities, so your precisely controlled mechanical system is not so precise. In the real world things rarely work the first time and you have to study the problem to decide if it is mechanical, electrical, or software. This is why teams should be doing system integration throughout the season. Don’t wait until the end to test the system or you may quickly run out of time.

I don’t believe a team with low resources can decide to just finish 30 pounds of the robot after bag-and-tag and truly be successful (OK, a low percentage will). These types of teams need to finish early giving the electrical team time to work out wiring issues, the software team time to test code and write autonomous modes, and the entire team time to make an objective decision about the drivers. You can build the most perfect mechanical system in the world, but without electronics, software, and drivers it will just sit there. Yes, we couldn't do it without the mechanical team either.

Remember there are 3 robots on every alliance. Many alliances will likely pick one defender/feeder robot, and many will look for a good balancer. You don’t have to have a way to score balls. Make a basic feeder before bag-and-tag. Give your drivers plenty of time to practice. If you show yourself to be a great defender that doesn’t get penalties at the regional, you will likely get picked.

[wireties]

Originally Posted by frasnow

I have a question for all the people talking about the 30 pounds and how great it is. How can a robot truly be competitive if the electrical team doesn’t have time to test their wiring, the software team doesn’t have time to test autonomous routines, and the drivers don’t have time to practice? I’ve seen this type of thing occur many times at work, and it always delays projects. Mechanical engineers sometimes think the deadline for the project is their deadline and don’t leave any time for integration and debug.

I don't think anyone here is advising this as a primary strategy. It is a backup when in desperate straits and running out of time.

[frasnow]

Originally Posted by wireties

I don't think anyone here is advising this as a primary strategy. It is a backup when in desperate straits and running out of time.

Then why not take the Apollo 13 approach? "Let's look at this thing from a... um, from a standpoint of status. What do we got on the spacecraft that's good?" Find a strategy that will work for the robot you have.

Originally Posted by wireties

fix the shooter and bring it to the regional. You'll sacrifice most of Thursday but it is better than nothing.

Sacrificing the practice time on Thursday for a high risk addition is rarely going to work in the scenario I described. There are three robots on every alliance.

[Basel A]

Originally Posted by frasnow

Then why not take the Apollo 13 approach? "Let's look at this thing from a... um, from a standpoint of status. What do we got on the spacecraft that's good?" Find a strategy that will work for the robot you have.

Let's say there's 3 potential subsystems, a drivetrain and two scoring mechanisms. The drivetrain and one scoring mechanism are completed. Are you suggesting that they do not work on a second scoring mechanism? I'm of the opinion that it couldn't hurt and if it doesn't work at the competition, as you've suggested is very likely, they can just take it off. Yes, they may sacrifice practice time, but if the addition does work, it could double their point output. A team that expects to win in qualification can't rely on playing defence because they will have matches paired with robots that can't score.

Yes, it is a high-risk situation. But it's also high-reward and it seems most people in this thread support taking the chance.

[frasnow]

Originally Posted by Basel A

Let's say there's 3 potential subsystems, a drivetrain and two scoring mechanisms. The drivetrain and one scoring mechanism are completed. Are you suggesting that they do not work on a second scoring mechanism? I'm of the opinion that it couldn't hurt and if it doesn't work at the competition, as you've suggested is very likely, they can just take it off. Yes, they may sacrifice practice time, but if the addition does work, it could double their point output. A team that expects to win in qualification can't rely on playing defence because they will have matches paired with robots that can't score.

Yes, it is a high-risk situation. But it's also high-reward and it seems most people in this thread support taking the chance.

Yes I am suggesting the team use what they know will work. Having a robust robot with well practiced drivers is far better than a half working robot with unpracticed drivers. Robustness is vital. Most teams will mark a team down on their scouting sheet for breaking repeatedly on the field. I'm saying a team doesn't have to expect to do well in the qualification rounds to do well at a regional. Our third pick for eliminations at the 2010 Autodesk Oregon Regional was ranked 55th after the qualification rounds. During eliminations they never scored a point, but they played some monster defense. We couldn't have won the regional without them.