|
|
|
![]() |
|
|||||||
|
||||||||
![]() |
| Thread Tools |
Rating:
|
Display Modes |
|
#46
|
||||
|
||||
|
Re: pic: 2015, Year of swerves?
Quote:
Wow, I didn't expect this to generate this much philosophical discussion of the objectives of FRC and the COTS vs. Custom argument, but i'm not complaining. Can't wait to see what 2015 brings. |
|
#47
|
||||
|
||||
|
Re: pic: 2015, Year of swerves?
Quote:
However, there is no way to avoid overcoming the friction of the wheels if the swerve is locked in such a formation. Unless you push with enough force to rotate the modules out of the formation (extremely unlikely), you are basically dealing with the whole weight of the robot times COF. If you can overcome that, then bully for you. If you push only the corner of the swerve, then it might be easier to push versus a head-on assault. As an aside, putting the wheels in such a formation makes it very quick to rotate the modules into their next position, making it an ideal rest position as well. |
|
#48
|
||||
|
||||
|
Re: pic: 2015, Year of swerves?
I'd like to see a graph like this made for other topics. It seems to me in general post quantity has gone up in general, and more teams are posting things they work on in the off-season then ever before. While I think people posting/working on swerve has probably increased, I think if you removed the growth of the program, the growth of presence on the forum, and the growth of openness about what teams are working on, I think this graph would probably be a lot less dramatic looking.
|
|
#49
|
|||
|
|||
|
Re: pic: 2015, Year of swerves?
That's a beautiful graph to me. I see no point in discouraging the trend, it means teams are going beyond what they did the previous years. While next year I'll bet a lot of those teams will stand still their first match or 10. 2,3 years from now many will be experienced "swerve teams". Just saying, variety makes first more exciting and educational. 50 "swerve teams" can teach all of us more about drive trains than 10.
|
|
#50
|
|||
|
|||
|
Re: pic: 2015, Year of swerves?
I enjoyed reading through this thread. It shows what a lot of people think about during competition and throughout the season. A lot more then actually talking about swerve lol.
Where I think most of the discussion is coming from is a difference in what teams are trying to accomplish during the season. Everyone is trying there best to accomplish what they feel is the best goal. That's the wonder of FIRST. There are many goals to shoot for. Take for example, My Senior year on 1565. I was team captain on a team that had never produced a good quality robot (Competitive or Engineering wise). Our team was small and had little machining capabilities. I decided to spend all summer prototyping a drive train that FIRST has never seen before. The game happened to work very well for this type of system so we built it that year to compete. Here is the drivetrain: http://www.chiefdelphi.com/forums/sh...806#post706806 The main objective of this drive train was not, however, to win an event. It was built entirely to "put our team in the spotlight" so to speak. To generate buzz and interest in the community of FIRST as well as in our own community. The robot did exactly what I wanted it to. We won some awards for it as a bonus and got a lot of attention at the events we attended as well as here on Delphi. But the biggest thing that happened was tripling our team from having about 10 people to have around 30. So, Long story short, Some teams build robots to win a regional, some build to win a championship, and some build to generate interest in science, as well as many other reasons. Swerve can be built to fit any of those goals, or built to fit none of them. Its really the choice of the team that wants to build it. |
|
#51
|
||||
|
||||
|
Quote:
Sorry, I disagree. As a teacher my responsibility to my students is to teach them new things. Now I'm sure that someone could make different iterations of the same type drive each year and learn something. I however try to have my team do at lease 1 thing every year that is completely new. That may or may not be a swerve. As a mentor my responsibility is to inspire the students. How many times have we heard that it's not about the robot? If my students look at the game and say "a swerve would be really cool this year" we are going to give it a shot. We may fail. We have before. In that case we will try a contingency plan. It may fail too. In the situation that we become a hindrance in someone's team for the 2 min match, honestly we'd be sorry, but we aren't going to try to limit our ambition during a 6 week build season because of it. I think doing that would reflect more poorly on us. |
|
#52
|
||||
|
||||
|
Re: pic: 2015, Year of swerves?
A skid-steer drivetrain does not necessarily have to have a tank drive interface.
|
|
#53
|
||||
|
||||
|
Something does get lost in the competition aspect of frc! It is a fact that can't be argued. I take that back... It can be argued for the sake of arguing it.
How many students join a robotics program with their sole intention is to win a competition?? I'd be willing to bet that the number is near zero. |
|
#54
|
|||
|
|||
|
Re: pic: 2015, Year of swerves?
Quote:
Another thing, you talk about swerve having great pushing ability because they can direct their wheels in line with where they are pushing. I don't think it is as simple as this. Most times we are trying to push someone, we are trying to push them in the direction our robot is pointed, so our wheels are largely aligned anyways. However, we will likely have more torque behind our robot because all of our drive wheels on a side are ganged together. When a robot pushes, the weight distribution on the wheels shifts (likely back, depending on how the bumer contact goes). Also, robots typically are not balanced front-back. If you do not have a 50-50 distribution on your wheel, you will not be having as much power to the ground as a 6WD. |
|
#55
|
||||
|
||||
|
Re: pic: 2015, Year of swerves?
Quote:
Assuming F is the force one motor exerts and two of the swerve robot's wheels are off the ground 4Fcos(theta) = 2F cos(theta) = 1/2 Theta = 60* So the "break even"point is 60*. The model is inaccurate because the 4 motor robot is actually traction limited, but the point stands. Independent swerve isn't very good at pushing matches, and its best bet is to avoid pushing altogether. Last edited by MichaelBick : 01-01-2015 at 14:44. |
|
#56
|
|||
|
|||
|
Re: pic: 2015, Year of swerves?
Quote:
Edit: Travis beat me to it Also, if a team is using PID control on the modules it's nearly impossible to rotate the caster with outside forces. Quote:
Last edited by Dunngeon : 01-01-2015 at 13:18. |
|
#57
|
|||
|
|||
|
Re: pic: 2015, Year of swerves?
An aside from the swerve vs. tank and "build to win" vs "build because it's cool" debate (all of which have arguments for):
Regardless of the drivetrain, something I have found common among the best of the best teams is that they have a drivetrain that they have iterated so much that it is pretty much a science for them. Examples include 16 and 1717's swerves or 254's WCD. I don't know what timeframe it takes the former two teams, but from 254's build blog, aside from minor changes, the luxury of having a drivetrain they know works seems to allow them to be done with it maybe 2 weeks into the build season. This allows them to focus on the other parts of the robot for the majority of the season and really produce a highly sophisticated machine. I don't think my team finished our prototype robot's drivetrain until week 4 of the build season last year (although some of that has to do with no CAD to be able to pull from for a new design - especially as it was our first year with true WCD, and the fact that we do all machining in house which as sanddrag pointed out really puts some limitations on your time and resource usage). On another note regarding the competition debates: I am firmly in the build to win camp. From Kickoff to the end of the the last competition, I am thinking about how we can build a robot that will maximize our score in any given match, and all my strategies and design revolve around one goal - winning the match. If that results in a "simple" robot (which it never does, I wish it would for the sake of consistency and ease of maintenance) then that's completely fine with me. However, I understand that everyone has different goals, and I recognize the different viewpoints of different people. Something I don't agree with at all is the notion that a few people seem to have that if you're not building a swerve drive you're not really experimenting or pushing yourself to the extremes of drivetrain innovation. There's a few things wrong with this, the first being that some people regard swerve a bit too highly in this thread - having a swerve drive, even a really good swerve system with great programming and great drivers, will not necessarily give you a very large advantage. I saw some of the best swerve drives from this year still get bogged down by defenders. A lot of people make an assumption that if you have good drivers, you will be able to navigate through defense - however, the people across the field are not idiots, those teams probably have good drivers as well. We build a prototype last year that a few teams have begun iterating called the Grasshopper drive, which combines the advantages of butterfly with WCD and uses only COTS items or easy to manually mill parts. It was a successful experiment for us and we will likely use the system again, game permitting. We learned quite a few things, and it wasn't swerve. It doesn't even come CLOSE to the complexity of swerve. We were able to compete with and against some of the highest level swerve drives this year. I have yet to see a team that makes me think, "if we don't have swerve, we're gonna get rolled over in competition." This is all just my opinion however, and I do agree that swerve is a great technical challenge. But to say it is the best drivetrain hands down would be a little misguided. |
|
#58
|
||||
|
||||
|
Re: pic: 2015, Year of swerves?
If FRC is supposed to be a real-world engineering "simulation" of sorts for students, as it sounds like some people are arguing (and fairly, because it is), then I think the build-to-win vs. build-to-learn debate comes down to one big thing.
In the real world, your efforts to be the "best" engineer should be those which produce the most effective product and best experience for your customer. Success in accomplishing a customer's goals for a product is analogous to success in accomplishing the GDC's goals for the game. Success in playing the game is defined by on-field performance, namely winning. Hence, by building with the intent to win the game, you have created the robot which most closely resembles the solution you should pursue in a real-world engineering project, and thus begun to learn how to be the best engineer you can. |
|
#59
|
||||
|
||||
|
Re: pic: 2015, Year of swerves?
I say we just go ahead and make it the year of the half track.
You know rack and pinion front two wheels the back uses tank treads. |
|
#60
|
||||
|
||||
|
Re: pic: 2015, Year of swerves?
I'll say that swerves are an example of doing something unnecessary yet competitively beneficial to gain experience and if you do happen to do it right an edge in competition. I never said I don't want to win I said I wanted to do something I haven't done before whenever I can and so long as it could possibly add to my competitiveness.
For instance in 2012 you didn't need to cross the middle barrier but we designed a drive train that could do it just in case. We barely used it but it came in handy when we needed it. And should a similar need arise I have a good way to do it. The other angle is a lot of teams can't afford to make off season projects or on season prototypes. We get 1 full robot a year, 1 shot, and I'm going to at least try to do something interesting with it, doesn't need to be a drive train. It can be using encoders for the first time or using pneumatics for the first time doesn't matter. Try to do something that you aren't sure about but can help you now or later both competitively and educationally. A rookie can learn a lot from a kit bot, a team that's made three kit bots can learn a lot from making a custom WCD drive, a team that has always made a custom WCD can learn alit from buying some 221 modules and a team that's made a swerve 3yrs in a row can learn a lot from making a octocanum. And no I'm not referencing anyone specifically or saying drive trains are the main point of all this. Say it with me... Examples. Another consideration is how likely would any given team be to win a regional doing only what they are comfortable with? That is to say a team that has won worlds one or more times probably is comfortable with many advanced things, so for them their comfort zone is miles ahead of the curve meanwhile a newer team my only be %90 sure with the basics. If they at least try to go for more they get one of two outcomes. #1 it didn't work but they were able to get much more knowledge to improve later or #2 it did and they are performing far better than was even just barley technically possible before. Last edited by jman4747 : 01-01-2015 at 14:35. Reason: grammar |
![]() |
| Thread Tools | |
| Display Modes | Rate This Thread |
|
|