|
|
|
![]() |
|
|||||||
|
||||||||
One of FIRST Team 45's off season projects. Any guesses?
18-11-2011 13:54
JesseKIf I had to guess, I'd say it's a for a drive train that's similar to what 1629 did in 2008. They had 2 wheel pods that could steer left right. When both turned opposite of each other, the effect was a very tight turn radius (front rotates left, rear rotates right, causing the robot to make a sharp left turn).
There may be other considerations that I'm not thinking of, but 3x1.5 C-channel (0.2" wall for the legs, 1/8-3/16" wall for the base) is rigid enough for what you're doing there and also fits the 2.5" super shifter into the channel very nicely. It would also provide extra aluminum to support the rest of an FRC frame.
18-11-2011 15:07
Aren SiekmeierAckermann steering, or at least simplified car steering. Are you planning to put a differential of some sort in the middle?
Cool idea. It would definitely change things since turning at high speeds is more straight forward, but turning on a dime would probably be harder (or at least different from what people are used to...)
18-11-2011 15:35
Peyton Yeung
18-11-2011 15:39
Ninja_Bait
18-11-2011 15:49
Peyton Yeung
|
However, I believe this is meant to be the first ever use of Necco Wafers as a lightweight sprocket replacement.
|
18-11-2011 16:18
Ninja_BaitHey, does that actually turn far enough for the robot to translate sideways? I just realized that it looks like it could - it would be an ackerman-crab.
18-11-2011 16:51
Peyton Yeung
|
Hey, does that actually turn far enough for the robot to translate sideways? I just realized that it looks like it could - it would be an ackerman-crab.
|
18-11-2011 19:00

18-11-2011 19:54
Peyton Yeung
|
Drivers: "This'll be interesting!" Builders: "Simple design, I love it!" Programmers: "f*ck." |
18-11-2011 19:57
thefro526
Does this use true Ackerman geometry or do the wheels turn in a parallelogram configuration?
There's a big difference between the two at high speeds.
18-11-2011 20:30
Aren SiekmeierOne thing I just noticed. It looks like on each side there are 3 miter gears, with two on the same shaft but on opposite sides of the wheel side gear. It seems like this wouldn't work and just lock up and/or shred the gears.
And I guess I'm a bit confused. It sounds like this is basically just a swerve implementation with wheels powered in pairs and perhaps steered in pairs as well. How exactly are you adding Ackermann functionality in a way that isn't already there mechanically?
18-11-2011 22:54
Ninja_Bait|
Aww...I was hoping more people would guess before someone got it right. Yes it is an ackerman-crab prototype.
|
19-11-2011 07:47
Peyton Yeung
In all this project was initially started as a swerve project but to reduce weight and cost of a traditional swerve/crab drive it took some designs from a car (ackerman steering). It isn't a true swerve/crab drive nor is it a true ackerman steering drive but somewhat of a hybrid for the sake of using less funds and weight.
19-11-2011 12:27
MichaelBick
There are a couple of problems with this design:
1) Compwiztobe is right; the miter gears will lock
2) If you try to turn this assembly, the wheels will try turn in opposite directions (think of a 2 wheel drive turning) and something will have to brake or slip.
3) (i'm assuming you are going to use two of these modules) When all the wheels are collinear, the robot will fall over. If you add casters though, you can fix this problem, but that defeats the point of make two wheel modules in the first place.
4) You won't be able to turn, because unlike regular swerve, you do not have the 4 modules to form a x with the "turn' chain. You could potentially add caster, but again, that defeats the point of making modules with two wheels.
Sorry for being so negative. On the bright side, this is the dedication that makes teams win. Nice job trying to reduce the price, weight, and complication of arguable the best drive train in first.
19-11-2011 12:41
MichaelBick
Sorry, I didn't realize what the pink sprockets were there for, so you can disregard comments 2 and 3. Just another comment, you will be limited in your strafing, specifically in strafes exactly left and right, or near those angles. This is countered though, by your easily achieved Ackerman steering, that needs very little programming. After my little revelation, I have realized that this is actually a rather viable drive train, once comment 1 has been fixed, to allow teams with very little resources and time, to achieve a drive train that acts like 6 wheels swerve, without the ability to strafe perpendicular the the robot, and to turn on the spot, like swerve and 6 wheel. With improvements, this could become something a lot like octocanum, giveing close to the best of both worlds.
19-11-2011 13:26
EricH
With the miter gears, I wouldn't be surprised in the least if the bottom one was free-spinning. It's set up so that that's all you'd need to do to keep the gears from locking up. Then you just have one gear acting as a spacer.
19-11-2011 22:29
Peyton Yeung
|
There are a couple of problems with this design:
1) Compwiztobe is right; the miter gears will lock 2) If you try to turn this assembly, the wheels will try turn in opposite directions (think of a 2 wheel drive turning) and something will have to brake or slip. 3) (i'm assuming you are going to use two of these modules) When all the wheels are collinear, the robot will fall over. If you add casters though, you can fix this problem, but that defeats the point of make two wheel modules in the first place. 4) You won't be able to turn, because unlike regular swerve, you do not have the 4 modules to form a x with the "turn' chain. You could potentially add caster, but again, that defeats the point of making modules with two wheels. Sorry for being so negative. On the bright side, this is the dedication that makes teams win. Nice job trying to reduce the price, weight, and complication of arguable the best drive train in first. |
20-11-2011 10:19
MattC9Awwwww I was looking for the new ball drive.
But it looks really good! I would also like to see how it interfaces with the frame and such.
20-11-2011 10:46
MichaelBick
|
1. Like EricH said, the bottom one is a spacer. 2. The wheels turn together so they always face the same direction. 3. There is no possible way for all 4 wheels to be on the same line. 4. Isn't it possible to turn a four wheel drive bot by giving a positive value to 1 set of wheels and a negative value to the other set of wheels? The fronts are linked together and the backs are linked together. |
20-11-2011 14:45
EricH
|
Like I said, disregard 2 and 3. For number 4, if you look at the CAD, the supershifter will be connected to both wheels. This means that you cannot power each side individually, so you cannot turn like a 6 wheel.
|
04-12-2011 08:49
CNettles11I hate to revive old threads but, this is a steering solution that will work, and this is how.

04-12-2011 08:55
Ninja_BaitOn that last diagram, the bottom arrow is reversed, right? Or does this have mecanum wheels capable of that kind of translation? 
04-12-2011 08:57
CNettles11That was definitely me slipping up, I should fix that.
04-12-2011 21:52
04-12-2011 21:55
MichaelBick
Not with this setup, as left and right wheels are driven together.
04-12-2011 23:28
RogerR
|
Not with this setup, as left and right wheels are driven together.
|
|
If you have 2 of these, each side is powered independently. If they're front and back, as stated, then it's a wide-oriented 4WD. Turns quite well, actually--maybe better than a six wheel drive
|
05-12-2011 01:57
Dillon Carey
|
I wonder if that's even more possible with a 6 wheel drop center swerve...
|
05-12-2011 09:38
JesseKHow well the robot will turn is also based on the placement of c.g., as described in this paper:
http://www.chiefdelphi.com/media/papers/1443
If anything, the zero-radius turn will be somewhat off-center because there is no symmetric place the center of gravity can exist while also maintaining stability for the other turning configurations AND being far enough offset from the center of area to reduce the torque/current needed. Off-centered-ness isn't a big deal, but it may muck with autonomous mode if zero-radius turns are heavily depended upon.