23-07-2014 05:51

Richard Wallace

Re: pic: #NotSwerve

#muitotorque
#talestilo

(Trabalhando com os geeks brasileiros.)

23-07-2014 19:40

audietron

Re: pic: #NotSwerve

Nice Job! Are you still planning on improving the swerve drive throughout the next couple seasons?

23-07-2014 19:44

thatprogrammer

Re: pic: #NotSwerve

#actuallymoveduringautonmous

Planning to use these in a WCD set up?

23-07-2014 19:52

Boe

Re: pic: #NotSwerve

Whats your plan for these? Are you just making a offseason drivebase, or are they being integrated into your 2014 robot for Minni-Mini, and Gitchi Gummi this summer/fall?

23-07-2014 20:15

ClockworkGold

Re: pic: #NotSwerve

Heck no

23-07-2014 20:19

ClockworkGold

Re: pic: #NotSwerve

1. Yes, we are doing an offseason drivebase. We finished part 1 of the CAD today.

2. We will likely be using this for the Minni-Mini, but as for the Gitchi Gummi, we would be cutting our time quite close.

23-07-2014 20:42

Ginger Power

Re: pic: #NotSwerve

From what I remember autonomous wasn't the problem. . . The rest of the match, however, was a different story

23-07-2014 20:46

AKSoapy29

Re: pic: #NotSwerve

For the swerve, we don't even have it working. The shadow bot was taken apart (Never really got it working) and the main bot was converted to tank during the MSHSL State Competition. We will probably be done with swerve for a very long time.

We are planning on making an 8 wheel WCD for offseason practice. We have had them in the past, but that experience has left the team and wasn't really past down..

I personally think it would be nice to incorporate this drive into the Gitchi Gummi or Minne-Mini offseason regionals (And redo the robot while we're at it ) but that probably won't happen. We might have to compete with our barely working 2014 thing that could fall apart any minute. Although we did redo all of the electrical and pneumatics and add lights

23-07-2014 20:48

AKSoapy29

Re: pic: #NotSwerve

We have had multiple autonomous' with multiple drive trains this year. We never really could stick to one thing, or at least something that worked it seemed. During the MSHSL State Competition, during lunch and any other time we got, we managed to scrap together a working autonomous.

23-07-2014 21:25

Boe

Re: pic: #NotSwerve

You've got half a build season til then plenty of time

23-07-2014 23:29

BBray_T1296

Re: pic: #NotSwerve

I see you are well into development of the innovative 2 wheeled #segwaydrive

24-07-2014 14:35

JohnFogarty

Re: pic: #NotSwerve

They learned the most important lesson. Anything that can move sideways is a waste of time where you could have been moving forward with more important things.

24-07-2014 14:44

AKSoapy29

Re: pic: #NotSwerve

Swerve to begin with was a waste of time. I think we got everything to build our drive 2-3 weeks in, built both drive trains, and the rest of the time was fixing everything and getting the other robot working. The robots weren't exactly alike, so when we got the code kind of working on the main bot, we deployed it to the shadow and nothing worked. Oh, and I think it took our programmers forever to fine tune everything on one robot, then fine tune it on the other robot. Everything was a pain, especially when the robot wouldn't even move on the field. We wanted maneuverability, but instead....

Oh yeah, and the first shooter was 100% pneumatic, and we were constantly working on it. Tip for anyone in the future: Don't use a pneumatic shooter unless you get it working REALLY well.

24-07-2014 15:05

llamadon

Re: pic: #NotSwerve

What he meant to say is "No most likely not, swerve was a challenge that we did not have the proper planning or knowhow to accomplish. Though it was a valuable lesson."
After analysing which drivetrains made it to einstein, and the kind or drivetrains our team has built in the past we decided to give this a try. It will be implemented in a west coast fashion, and is currently being CADed

24-07-2014 15:07

Re: pic: #NotSwerve

What spreads and speeds are you using on the transmissions, alongside what wheels? Also, will you have any active way of measuring your total current draw with the 6 CIMs?

24-07-2014 15:17

llamadon

Re: pic: #NotSwerve

Hey Brian, these gearboxes are going to be implemented on a summer drive project aimed towards testing the benefits of a 6 cim WC drive. We hope that future games will still allow for west coast drives to be feasible designs, and that our work this summer can be used for seasons to come.

24-07-2014 15:27

llamadon

Re: pic: #NotSwerve

We plan to use 4" hex bore Andymark performance wheels on live axles with blue nitrile tread. We will be testing a 2.92 spread ratio, set to roughly 7fps/18fps, a 3.64 spread at 5.5fps/18fps, and finally a 2.06 spread geared to 7fps/13.5fps. The 2.06 spread is being tested specifically to see what kinds of acceleration we will see with a lower top speed, but more torque. As for the current draw, we have not thought of any ways to measure that yet.

24-07-2014 15:33

Re: pic: #NotSwerve

Excellent idea of testing the different speeds and spreads. I'm not entirely certain what method they used, but Team Spectrum 3847 had a method of measuring their current draw on their robot this year. While you are testing spreads, I think it would be interesting to add this current meter to see how many amps you're pulling in certain situations. It could help you better assess what main breaker limitations you may have in the future.

24-07-2014 15:42

llamadon

Re: pic: #NotSwerve

Thanks! I will definitely inquire of this testing methods. I have read up a small amount on the effects of this number of CIMs on the breaker and have heard of a few blow outs. These blow outs tend to happen in an immediate change in voltage. IE. driving the robot forward and then immediately going backwards. Teams have implemented a part in the code that does not allow this immediate jump from 0 to 100 percent, and instead ramps it up steadily.

24-07-2014 15:46

AKSoapy29

Re: pic: #NotSwerve

Well we only tripped the breakers a few times on our shooter, lol. Load testing would be neat. I think we have a load tester in the cave actually, I saw it the other day.

25-07-2014 12:00

Arpan

Re: pic: #NotSwerve

We used these gearboxes on our robot this year. They were fantastic, esp. considering how we abused them. A few notes :

1. LOCKTITE like the instructions say. We did not do this and it caused several screw failures throughout the season, resulting in worn teeth on some of our gears. The gears themselves are overbuilt - they survived the extreme chipping a CIM coming unmounted caused.

2. The 22 fps calculated speed is too fast to drive without proper code, and will blow your breakers whenever you hit anything. For FRC applications , I don't recommend going faster than 16 FPS.

I'd be very interested in seeing the results of your tests of different ratios.

25-07-2014 12:09

AKSoapy29

Re: pic: #NotSwerve

Thank you, this is very useful to know! 22 fps does seem too fast. I think I have only driven with 16, and it was plenty, and easy to control. I am really interested to see how our new drive train will handle!

25-07-2014 12:14

Boe

Re: pic: #NotSwerve

There were a fair number of very successful teams this year that ran 6 CIM drive trains geared to >20fps before losses. Its possible to do but you have to be careful with you dont blow your breakers.

25-07-2014 12:27

RKazmer

Re: pic: #NotSwerve

This is very true. It took three weeks of prototyping to get the shooter right on our robot this year. We even had to tweak it after that due to other pieces for better ball control and stability. However, having a pneumatic shooter was great since we didn't have to have stored energy on the bot versus a spring or motor driven shooting system.

Main lesson for using pneumatics to power a catapult: Start early and have multiple designs that can be tested quickly.

25-07-2014 16:45

llamadon

Re: pic: #NotSwerve

Thanks Arpan! I will definitely post the results of our testing for you to veiw

I will remember the locktite! As for the speed I was actually having an mental debate with myself, and that is: Why gear for a high speed like say 22 fps? How often will you actually attain that speed? Being able to accelerate quickly is much more beneficial (in my mind) then being able to reach a high speed after spending a notable amount of time accelerating. Just a thought, and one of the reasons for the testing

25-07-2014 17:06

Re: pic: #NotSwerve

Teams that gear for that speed use it in cases where you would want to cross the entire field extremely quickly in one straight shot (think 2011 or 2013). That being said, it's still something interesting to test. Most robots that run these speeds are on the lighter side (90-100 lbs), which helps them abuse the higher speeds with greater acceleration.

25-07-2014 18:26

Travis Hoffman

Re: pic: #NotSwerve

These were sufficiently evil enough for our purposes in 2014.

4" Colsons on corners, 4" blue nitrile performance in the middle, 2.92 spread, max robot weight, roughly 5 fps / 15 fps split

And yes, Loctite everything per instructions.

25-07-2014 21:38

nathannfm

Re: pic: #NotSwerve

It looks like I missed the back story here. If it's not too painful would someone from 2169 mind recounting their season of swerve and what problems you had with it for other teams who are dead set on at least trying swerve as an off season project so they may learn from your season?

26-07-2014 23:46

llamadon

Re: pic: #NotSwerve

We used the "Wild Swerve" modules from Team 221 robotics parts. We did not order them prior to build season. I think they were ordered by the end of week 1. Arriving mid week 2, we assembled them and attached them to our drive base. Our programers were coding prior to receiving the modules, confident in their code. We implement the code, not up to snuff. We keep editing and revising, problems arise left and right (using 3D printed encoder mounts was one of said issues.) We focused so much on our swerve that other mechanisms suffered. We over complicated the code, and the start up routine for each match (we had a "hard zero" button to "reset" the encoder values to 0) and in the end we didn't move for half our regionals. My advice is do your trouble shooting and learning before the build season, and only implement swerve for the build season if you already have a polished code for it, and have designed and planned the mechanical implementation of the drive before kick-off.

27-07-2014 10:40

Arpan

Re: pic: #NotSwerve

They do. However, in my experience, the risk of blowing the main breaker here is not worth it. It also necessitates a drive system such as split arcade for decent control.

We're considering going to ~16 calculated high and ~7-9 calculated low next year to avoid blowing the breaker.

02-08-2014 04:53

asid61

Re: pic: #NotSwerve

Main breaker blows can be easily avoided by implementing a "crash detection" current-sense code. Use a current sensor to check if it is a spike. If the current spikes for too long, you shift to low gear or put the CIMs in brake mode.
If you are using a shifter, I think going 18-20 fps is not all bad, as long as your driver can control it.