6 wheel swerves [Archive]

Andrew Lawrence

06-12-2011, 21:31

Does anyone know anything about 6 wheel swerves? Has it been done?

Is it possible to do a 6 wheel drop center swerve? Where the center 2 wheels are dropped the normal 1/8 of an inch, enabling easier turning with the same swerve capabilities? Is that bad? Do you not need it to turn easier?

Any pics/vids would be great! Gonna go look at CD Media some more and see what I get. ::rtm::

Thanks!

Ankit S.

06-12-2011, 21:33

I believe 1625's 2010 robot was a 6 wheel swerve. Heres a vid: http://www.youtube.com/watch?v=U7sXWhHfULs

Wouldnt it just be easier to program the robot for turning in a "wide" orientation?

thefro526

06-12-2011, 21:34

Team 1625 built a 6WD swerve in 2010. It ended up playing on Einstein.

If you search through Aren Hill's recent posts, he's got some solid information and pictures out there. Or PM directly if he doesn't post in this thread.

I believe that 1625's 6WD swerve is the first and only 6WD Swerve.

Good picture in this post: http://www.chiefdelphi.com/forums/showpost.php?p=1086307&postcount=21

O'Sancheski

06-12-2011, 21:40

Like everyone has stated, 1625 has done it. (I believe that they are the only ones so far.)

Here are some pics that I got from CD-Media. I'm sure that 1625 has many more on their website.

http://www.chiefdelphi.com/media/photos/35824

http://www.chiefdelphi.com/media/photos/35087

http://www.chiefdelphi.com/media/photos/35854

Also, here is a great whitepaper that 1625 published about their swerve drives throughout the teams history.

http://www.chiefdelphi.com/media/papers/2578

Peyton Yeung

06-12-2011, 21:40

1625 also posted this paper that references the 6 wheel drive swerve.

http://www.chiefdelphi.com/media/papers/2578

Andrew Lawrence

06-12-2011, 21:44

Aren_Hill

06-12-2011, 22:59

If you want to know more about that message Dillon Carey, he's the one who made the thing

Akash Rastogi

06-12-2011, 23:16

If you want to know more about that message Dillon Carey, he's the one who made the thing

People tend to forget that 1625 isn't solely Aren Hill. :rolleyes:

Lil' Lavery

07-12-2011, 04:08

Thanks all! Just finished watching their video on youtube from 2010. Seems nice. The only real implementation I can see is extra pushing power in all directions, which is in itself useful, depending on the robot.

See the litany of other threads on the issue, but more wheels does not mean more "pushing power."

thefro526

07-12-2011, 09:59

See the litany of other threads on the issue, but more wheels does not mean more "pushing power."

This is not always true, especially with small rough-top tread wheels. I believe someone posted Data on the subject that showed a ~20% increase in Pushing Power/Tractive Force/Force of Friction when a 4" wheel was widened from 1" to 2". (For all intents and purposes, the same as adding another 1" wide wheel to the DT.)

This seems to be caused, at least in part, by the way rough-top tread interfaces with carpet.

Andrew Lawrence

07-12-2011, 10:24

^ What Dustin said.

Lil' Lavery

07-12-2011, 14:10

Aren_Hill

07-12-2011, 14:13

234 has not posted the Data online, as it is their choice. But Chris graciously sent me the info and test setup they used to determine traction differences. They compared 4",6", and 8" wheels, 1" and 2" wide. 4x2 wheels won out by a very significant margin

thefro526

07-12-2011, 14:19

Please provide a link.

I don't have a direct link to the data as it was given to me second hand.

If all goes well, I will have some data soon on the relation between pushing force and contact patch once we complete testing with our prototype chassis.

JamesCH95

07-12-2011, 15:21

This is not always true, especially with small rough-top tread wheels. I believe someone posted Data on the subject that showed a ~20% increase in Pushing Power/Tractive Force/Force of Friction when a 4" wheel was widened from 1" to 2". (For all intents and purposes, the same as adding another 1" wide wheel to the DT.)

This seems to be caused, at least in part, by the way rough-top tread interfaces with carpet.

Please provide a link.

I think that the principal is sound from a real-world physics standpoint. I think that it is very possible to see significant traction gains from wider wheels.

From my race car experience (both designing, setting up, and driving) I have learned that polymers' coefficient of friction is related to their contact pressure. A larger contact area therefore leads to a higher coefficient of friction.

They way it was explained to me is that as a polymer, i.e. roughtop or a tire tread, is pushed onto a surface small parts of the two surfaces become interlocked. As normal force between the two surfaces increases, the two surfaces interlock less and less per unit normal force. At some point the polymer/material interfaces are completely saturated and there is not much more grip to be had.

JesseK

07-12-2011, 15:24

Quick aside -- if 234's data is true then wider treads also make turning more difficult due to extra skid traction (if implemented across all 6 wheels).

Dillon Carey

07-12-2011, 16:31

The only real implementation I can see is extra pushing power in all directions, which is in itself useful, depending on the robot.

The six wheel swerve actually had basically no extra pushing power past that which a four wheel swerve would have supplied. This is because it was a dropped center 6 wheel. Therefore it effectively only had four wheels on the ground at any given time.

The main plus to the 6wd swerve was maneuverability.

Our team has done a few 4 wheel "swerves" (people call different wheel configurations different names, so I will just group them all under the term "swerve"). We had always done 4 wheel swerves before, but didn't have the programming capability to create a swerve with independent steering and power for each module. without doing this a four wheel swerve will almost always lack the ability to rotate about itself well. (a notable exception is 1717's steering configuration)

In order to get past this dilemma without using programming, we decided to solve it mechanically, with the 6wd swerve.

Ether

07-12-2011, 20:07

For any math-oriented students who might be curious, here's (http://www.chiefdelphi.com/media/papers/2426)1 the inverse kinematics for 6-wheel swerve.

I can't imagine any team actually wanting to build one like that, but who knows :-)

1look for "swerveN.pdf" near bottom of list

Brandon Holley

08-12-2011, 09:31

Quick aside -- if 234's data is true then wider treads also make turning more difficult due to extra skid traction (if implemented across all 6 wheels).

While I'm not certain of the exact testing setup 234 has created and tested, we setup our own system.

I can verify that 4X2 won out in our setup as well.

-Brando

Ian Curtis

08-12-2011, 17:03

While I'm not certain of the exact testing setup 234 has created and tested, we setup our own system.

I can verify that 4X2 won out in our setup as well.

-Brando

Brandon/234/Aren,

Did you use a flat plate of tread of varying sizes or use actual wheels and some kind of load cell? Did you do both? It probably doesn't matter for the final result, I'm just curious how much of that increase is due to straight up surface area or if you get "bonus traction" due to the changing radius and interactions of the wheel/carpet interface under different loads.

Lil' Lavery

08-12-2011, 19:01

Andrew Lawrence

08-12-2011, 19:03

Ok. Now a new question for those who have used swerve:

Wild swerve, Revolution swerve, or custom swerve?

Ninja_Bait

08-12-2011, 19:13

Ok. Now a new question for those who have used swerve:

Wild swerve, Revolution swerve, or custom swerve?

We haven't implemented it yet but we have a design for a custom swerve. I feel like COTS swerve is less impressive than one that a team actually spent time designing, and adding their personal style to the design.

Marc S.

08-12-2011, 20:33

Ok. Now a new question for those who have used swerve:

Wild swerve, Revolution swerve, or custom swerve?

If you know how to design or make it equally as effective than i would recommend making your own. If you try at it you will find some simple solutions in making it lighter and more suited to your design style.

However if you are not confident that you can make one then go with the revolution kit.

Peyton Yeung

08-12-2011, 21:40

While this is our first year trying swerve, I would suggest staying away from a custom setup if you are new to swerve drives. If you try it make sure you have plenty of time to get it all up and working.

We are trying to create a custom swerve drive and it has taken quite a while to design and build what we have so far (not quite done yet).

Wild swerve is a little cheaper than revolutionary and both take less time than a custom setup to design because they are already designed.

Brandon Holley

09-12-2011, 08:40

No offense to 234/125, but until I can see the testing set-up, parameters, and data first hand I reserve my right to be skeptical.

Additionally, I'd like to see if they evaluated a greater quantity of wheels compared to wheel diameter. While I don't doubt there would be a correlation there, I'm curious if it's a 1:1 relationship. After all, this point did originate with having 6WD swerves providing more "pushing power."

That is indeed your right, and I too would be skeptical if the situations were reversed. Until we have our testing results in a more organized format, we're going to withhold them.

If I am understanding your second question, are you asking if we tested more than 1 wheel at a time?

About the testing setup- the wheels are mounted on a pivot arm, weight is applied at the end of the arm thus giving a standard "normal" force to any wheel placed in the rig. A torque wrench was utilized, and torque was applied to each wheel (each wheel being affixed to a shaft). We then recorded the torque each wheel experienced right up until it slipped. Pretty simple.

-Brando

Ether

09-12-2011, 09:00

About the testing setup- the wheels are mounted on a pivot arm, weight is applied at the end of the arm thus giving a standard "normal" force to any wheel placed in the rig. A torque wrench was utilized, and torque was applied to each wheel (each wheel being affixed to a shaft). We then recorded the torque each wheel experienced right up until it slipped. Pretty simple.

What precautions did you take to make sure the force on the torque wrench handle did not change the normal force?

Brandon Holley

09-12-2011, 09:24

What precautions did you take to make sure the force on the torque wrench handle did not change the normal force?

The torque wrench utilized a ~5" extension so that the physical handle of it hung off the edge of the table. We then had an aluminum block that had a pocket cut out of the top to accept the extension on the torque wrench. A top piece would then screw into the block to captivate the torque wrench extension. The entire torque wrench assembly would then be self-standing even when disconnected from the wheel testing rig. Due to various size wheels, we used this block in combination with a series of spacers to keep the extension on the torque wrench fully supported for all wheel sizes.

Ether

09-12-2011, 10:03

The torque wrench utilized a ~5" extension so that the physical handle of it hung off the edge of the table. We then had an aluminum block that had a pocket cut out of the top to accept the extension on the torque wrench. A top piece would then screw into the block to captivate the torque wrench extension. The entire torque wrench assembly would then be self-standing even when disconnected from the wheel testing rig. Due to various size wheels, we used this block in combination with a series of spacers to keep the extension on the torque wrench fully supported for all wheel sizes.

I can't quite picture what you're describing. If the torque wrench handle is captured in a (stationary?) aluminum block, how do you push on it to exert a torque?

Do you have a picture of the fixture?

Brandon Holley

09-12-2011, 10:44

I can't quite picture what you're describing. If the torque wrench handle is captured in a (stationary?) aluminum block, how do you push on it to exert a torque?

Do you have a picture of the fixture?

I don't have a picture offhand, but can get one at our next meeting.

The block is supporting the extension of the torque wrench, not the handle itself. The handle remains free to rotate as it would normally.

The aluminum block is not acting as a clamp, more of a bushing. A loose enough fit that the wrench can spin, but snug enough to provide support to the extension. The block ensures the torque wrench is in the same position every time, and also ensures that the torque load is being exerted as axially as possible to the wheel.

PS: Maybe we should consider resurrecting/starting another thread as we've pretty much hijacked this one.