FRC T-bone-ing and Hexagonal drive

[nuclearnerd]

Originally Posted by Lil' Lavery

Only if they hit it right on.

If I understand it correctly, one of the effective things about T-Bone pins is that they are "self-engaging". Check out the attached sketch (in stunning powerpoint CAD!) When the defensive robot on the left (yellow) hits the offensive robot on the side (blue), the contact force is angled to the right of the Blue robot's turning center. This creates a torque (green arrow) that turns the blue robot into the front bumper of the yellow robot. Since the other corner of the yellow robot is on the left side of the turning center, contact there will try to turn the blue robot back the other way. The result is as long as the yellow bot pushes, the blue robot can't exert a turning torque greater than the one caused by the contact, so the two robots become locked together at a 90 degree T (not shown).

A hexagonal robot (shown in blue on the right) changes the contact location and angle in an attempt to move that contact force closer towards (or completely across) the blue robot's turning center, reducing the leverage and producing less "T-Bone torque". But this can be subverted by notching the front bumper of the defensive robot (shown in yellow on the right). The notched bumper contacts the blue robot at almost exactly the same place, and with as much T-Bone torque as the scenario on the left. (although it might be a little reduced if the hex angle is sharp enough).

An easier way to reduce the chance of getting T-boned is to move the turning center closer to the front or back of the offensive robot. If you moved the turning center on the blue robot all the way to the right of the diagram for instance, the yellow robot would produce almost no T-Bone torque (or even reverse or helpful) T-Bone torque. The turning center can be repositioned using permanent or drop-down omni wheels (as suggested earlier), or with swerve drives. Those features allow an offensive robot to "pick and roll" off a T-Bone pin.

That's my understanding of things anyway. I'd be interested to hear if other people of different theories or data.

[GeeTwo]

Originally Posted by nuclearnerd

An easier way to reduce the chance of getting T-boned is to move the turning center closer to the front or back of the offensive robot.... using permanent or drop-down omni wheels (as suggested earlier), or with swerve drives. Those features allow an offensive robot to "pick and roll" off a T-Bone pin.

That's my understanding of things anyway. I'd be interested to hear if other people of different theories or data.

Using permanent omni wheels to allow you to roll off of a T-bone pin may be jumping out of the frying pan into the fire. If you move your center of rotation very far towards one end of the robot, you are making yourself more susceptible to being steered from the other end by an opposing robot.

[JVN]

Originally Posted by GeeTwo

Using permanent omni wheels to allow you to roll off of a T-bone pin may be jumping out of the frying pan into the fire. If you move your center of rotation very far towards one end of the robot, you are making yourself more susceptible to being steered from the other end by an opposing robot.

This is entirely game dependent. Several teams were successful in 2014 using a 2+2 setup. Similarly safe zones in 2011, 2012, 2013 would have lent themselves well to this setup.

"Beat the defender, get into position and score."

In 2014, the secret was to score quick enough that a safe zone wasn't needed.

[GeeTwo]

Originally Posted by GeeTwo

Using permanent omni wheels to allow you to roll off of a T-bone pin may be jumping out of the frying pan into the fire. If you move your center of rotation very far towards one end of the robot, you are making yourself more susceptible to being steered from the other end by an opposing robot.

Originally Posted by JVN

This is entirely game dependent. Several teams were successful in 2014 using a 2+2 setup. Similarly safe zones in 2011, 2012, 2013 would have lent themselves well to this setup.

"Beat the defender, get into position and score."

In 2014, the secret was to score quick enough that a safe zone wasn't needed.

Exactly what I meant by "may be". If a defender can steer you away from where you want to be, that might be as effective as stopping you where you are. Note also that you may be opening yourself up to being effectively defended via steering by robots which cannot generate enough force and drivers who do not have quite enough skill to effect a T-bone against you.

[AdamHeard]

Too many people in this thread are speculating. That puts a lot of noise out there for teams looking for good information.

[EricH]

Originally Posted by AdamHeard

Too many people in this thread are speculating. That puts a lot of noise out there for teams looking for good information.

Question (somewhat rhetorical): As it relates to being the victim of a T-bone pin, does swerve (or mecanum, or other omnidirectional drives) help you get out, or make the problem worse? What about if you're trying to place the T-bone?

[Ari423]

Originally Posted by EricH

Question (somewhat rhetorical): As it relates to being the victim of a T-bone pin, does swerve (or mecanum, or other omnidirectional drives) help you get out, or make the problem worse? What about if you're trying to place the T-bone?

Wheels with rollers (mecanum, omni) don't get T-boned, they just get pushed. This could be a positive or a negative depending on your team's strategy. Swerve also can't get T-boned because you can turn the wheels 90* and drive away from the T-boner.

Wheels with rollers also can't cause a T-bone, since they can't push with any substantial force. Swerve drive can place a T-bone.

[Andrew Schreiber]

Originally Posted by Ari423

Swerve also can't get T-boned because you can turn the wheels 90* and drive away from the T-boner.

Source? You're operating under the assumption that the wheels can under the extra pinning load (which for some swerves isn't true) and that the wheels all retain normal loading making driving it sane. And that the driver thinks of that.

I've seen swerves get TBoned.


I've also seen mecanum wheeled robots shove 6wd robots sideways.. Physics is weird sometimes, this is why we prototype.

[Abhishek R]

Originally Posted by Ari423

Wheels with rollers also can't cause a T-bone, since they can't push with any substantial force. Swerve drive can place a T-bone.

Not true - I'm not sure about mecanum, but you can definitely friction pin other robots with omnis, we did it all the time in 2014.

Personally I'm a fan of active mechanisms to get out of T-bones. Using our offseason edition of our H-drive in 2014 we never got stuck in a pin unless it was against the wall. However, a hexagon is another way to achieve the same goal. If your team feels it's worth pursuing, go ahead and do it - speaking from experience, it really is an awful feeling to not even be able to control the movement of your robot just because someone's pushing you around.

[Jay O'Donnell]

Originally Posted by Abhishek R

Not true - I'm not sure about mecanum, but you can definitely friction pin other robots with omnis, we did it all the time in 2014.

1058 did it with mecanum against teams with weaker tank drives or other non-tank drives pretty consistently in 2014. However once we met teams with more powerful drive trains we were less effective.

[Wayne TenBrink]

We retrofitted our 2014 bot for a hex frame (Octoframe, actually) last fall and used it at an off-season event. There is a brief thread on CD about the experience:

http://www.chiefdelphi.com/forums/sh...30#post1406830

We weren't on the receiving end of a lot of T-Bone attempts at the event, but our driver did get a bit of experience with it. It helped a bit with T-Bones, but didn't appear to be a magic bullet. The biggest problem with bulged side bumpers came when we tried driving parallel to the wall. Once the tip of the bulge hit the wall, it turned the robot slightly toward the wall, which started an instant vicious cycle that sucked the bot tight into the wall. I don't expect we will try them again soon.

[AdamHeard]

Originally Posted by EricH

Question (somewhat rhetorical): As it relates to being the victim of a T-bone pin, does swerve (or mecanum, or other omnidirectional drives) help you get out, or make the problem worse? What about if you're trying to place the T-bone?

In regards to swerve, we didn't do enough testing with this in mind to make scientific statements. That being said based on the three we've run, I would think it would make it easier to get out of T-bones in a perfect world. I'd highly recommend not investigating swerve as a t-bone solution. There are so many points in a swerve drive where the increased load from being t-boned could cause your gearing to experience increased friction and make the situation worse.

Our experience running butterfly in 2013 offseason made it seem well suited for that style of safe zone to safe zone sprinting, but we also ran a normal 6wd that year and had similar performance.

It's hard to say for sure, there are many variables involved with the implementation (and how its' driven) that matter a great deal, so it's not as simple as X drive versus Y drive.



The best thing to do is to proto and test what you want to run if this is a performance advantage you are going for. Designing your setup to be a modification of the AndyMark or Vex kit drives would make this iteration easier for most teams.

[notmattlythgoe]

Originally Posted by AdamHeard

In regards to swerve, we didn't do enough testing with this in mind to make scientific statements. That being said based on the three we've run, I would think it would make it easier to get out of T-bones in a perfect world. I'd highly recommend not investigating swerve as a t-bone solution. There are so many points in a swerve drive where the increased load from being t-boned could cause your gearing to experience increased friction and make the situation worse.

Our experience running butterfly in 2013 offseason made it seem well suited for that style of safe zone to safe zone sprinting, but we also ran a normal 6wd that year and had similar performance.

It's hard to say for sure, there are many variables involved with the implementation (and how its' driven) that matter a great deal, so it's not as simple as X drive versus Y drive.



The best thing to do is to proto and test what you want to run if this is a performance advantage you are going for. Designing your setup to be a modification of the AndyMark or Vex kit drives would make this iteration easier for most teams.

Curious. What kind of butterfly did you run? This style? How did you have the controls set up? Were the front and rear wheels able to actuate independently?

[MichaelBick]

Originally Posted by notmattlythgoe

Curious. What kind of butterfly did you run? This style? How did you have the controls set up? Were the front and rear wheels able to actuate independently?

They ran the neutrino style with slightly faster gearing. Each module had it's own actuator and motor.

[MrRoboSteve]

Hexagonal and round robots have trouble fitting through doors in some venues. Williams Arena, I'm lookin' at you!