pic: Swerve^2 Prototype

[CD47-Bot]

[cdm-description=photo]16653[/cdm-description]

[Skabana159]

Are you using potentiometers or digital encoders to track the orientation of your wheels?
Also, can I get a close-up shot of your drill's gearbox? It looks neat.

[Stormhammer]

k, now THAT is interesting... though... hm :rolleyes:

[Greg Needel]

very nice system....have you had any problems with drawing too much current or stalling the drill motor since you are only using 1 to power all 4 wheels?

[WakeZero]

We use two pots on both steering chains to determine position, so four total. As for the one drill motor issues, we have noticed that problem to a small extent... but they will go away once we install the second drill :yikes:

[WakeZero]

http://stickysituation.westhost.com/crush/mikepix/X%20gear%20box.jpg

http://stickysituation.westhost.com/crush/mikepix/20031217%20xbox%20600x800.jpg

We lovingly refer to this little contraption as the X-Box, since it's chains form an X that go out to drive each wheel. What it allows is for two drill motors to be mounted horizontally on the robot to drive the wheels. It also has a nice feature in that it makes the drill motors spin in the same direction (Clockwise and Clockwise, or CCW and CCW) :ahh:

We may be taking out the four shafts around the middle one and just putting four or two sprockets on the top of the middle shaft... we would only do this if we thought it was too inefficient, or we were hurting for weight :rolleyes:

[Madison]

We lovingly refer to this little contraption as the X-Box, since it's chains form an X that go out to drive each wheel. What it allows is for two drill motors to be mounted horizontally on the robot to drive the wheels. It also has a nice feature in that it makes the drill motors spin in the same direction (Clockwise and Clockwise, or CCW and CCW) :ahh:

We may be taking out the four shafts around the middle one and just putting four or two sprockets on the top of the middle shaft... we would only do this if we thought it was too inefficient, or we were hurting for weight :rolleyes:

It's clear that it can translate along the ground -- but how does it turn? With two motors powering a single gearbox that drives all four wheels, is differential steering possible?

[edomus]

It's clear that it can translate along the ground -- but how does it turn? With two motors powering a single gearbox that drives all four wheels, is differential steering possible?


looks like they also are using crab drive with different motors powering the turning

[WakeZero]

http://stickysituation.westhost.com/crush/mikepix/steering%20motor%20and%20sensor%20assembly.jpg

Here is what we use to turn ONE set of wheels. We have one of these on the back two wheels, and one on the front. You can't see them in the picture because they are mounted upward, and hidden by the frame ;)

I explained it in the picture description, but that doesn't show up in the discussion. So here it is again :)

Yes, it does drive on it's own power up the small steps. Right now we have one drill motor driving all four motors, this will eventually be two. To steer we use two globe motors, one for the front wheels and one for the back. This allows us to do two things:

A) Tank-turn by turning the front wheels 180 degrees from the back wheels. This allows us to change the robots orientation easily.

and

B) Change to a car drive mode where only the front wheels turn.

To drive the swerve, right now we have one joystick. In programming we convert the x and y coordinates to polar by using lookup tables. This gives us the angle we want the wheels to face, and then the magnitude for how fast we want to drive.

[Travis Covington]

From what I can tell, it looks like there is no support on the other side of the globe motor shaft. It may just be the picture you showed, but in that configuration there is a significant side load that (typically) results in the globe motor releasing its magic smoke ;)

If I am incorrect, sorry about that... Just thought I would note that... its an easy fix.

Otherwise it is VERY original! Looks good!

[WizardOfAz]

From what I can tell, it looks like there is no support on the other side of the globe motor shaft. It may just be the picture you showed, but in that configuration there is a significant side load that (typically) results in the globe motor releasing its magic smoke ;)

If I am incorrect, sorry about that... Just thought I would note that... its an easy fix.

Otherwise it is VERY original! Looks good!

Well, you're right about that. I've worried about it some, but since steering is such light duty cycle, I don't think we'll have a problem. Also, the side load is very close to the motor - it would be much worse with a longer shaft. I'd guess in a typical game we'll turn the wheels an aggregate of a few revolutions, maybe a few tens at the upper bound. In our setup, the globe motors rotate 2.8x as fast as you steer the wheels, so maybe we'll get on the order of 100 total turns of rotation in a game. I'm thinking the motor won't suffer much with that much use. Does this disagree with your intutition or experience?

Bill

[Travis Covington]

Your assumption would be correct for almost all motors in the kit except the globe motor (in my honest opinion).

For some reason (maybe someone knows specifically) the globe motor is VERY sensitive to side loads. In past years when my team used these motors, if the shaft was not supported on both ends we would ruin the motor, even if it was used rarely.

If it is possible (it looks failry easy) you might want to try and put a bearing supporting the shaft on the other end... even though you might not really need it... it seems like it serves an important enough function to deem the extra work beneficial... even if it only gives you peace of mind.

Hope that helps :D

[WakeZero]

Thanks Travis, we will surely be looking into this now :yikes:

[Paul Copioli]

I am going to have to back up Travis on this one. We used the globe motors for swerve steering in 2002 and we hade to add a bearing to support the load. We ruined 3 globe motors on our prototype unitl we figured it out. Once we added the bearing, we did not fail any globe motors.


-Paul

[WizardOfAz]

I am going to have to back up Travis on this one. We used the globe motors for swerve steering in 2002 and we hade to add a bearing to support the load. We ruined 3 globe motors on our prototype unitl we figured it out. Once we added the bearing, we did not fail any globe motors.
-Paul

OK, I hear the growing concensus. Our original design had a top plate and a bearing. Just "simplified" it to a single plate to the the prototype going quick, and it worked pretty well, so I was hoping to leave it that way. The prototype in the photo will be getting some road time in the next few days, so we'll see how that goes. What are the first symptoms of impending failure? Noise? Smoke? Just locks up?

[Paul Copioli]

The first indications on our robot were that the wheels were not getting to the position we commanded. At first, we thought it was the potentiometer going bad, but it was actually the motor not having enough torque at the lower gains (as it gets closer to the commanded position). The motor slowely loses torque capacity as you fail the front bushing. The planetary gears inside the globe transmission start to wear and bind up, thus losing torque capacity. The motor will eventually lock up, but the swerve steering will suffer long before lock up.

-Paul

[Joe Johnson]

The main reason the Globes need support is that they have only one bearing if you don't. The second bearing becomes the planetary gears on the last stage if you don't give it an external one. This is very bad for efficiency of the gearbox.

Joe J.

[ajlapp]

i noticed you have clearance for the steps....any thoughts about going up the sides of the platform. there isn't much room around that goal once you go up the steps.

[WizardOfAz]

i noticed you have clearance for the steps....any thoughts about going up the sides of the platform. there isn't much room around that goal once you go up the steps.

Yes, we're working on that. Essentially raising the body bottom for more clearance and putting front and mid-body skids. Will post another picture soon.

Bill

[Paul H]

How much does that prototype weigh?

[WizardOfAz]

How much does that prototype weigh?

It's about 100 pounds, but about half of that is the wood base which will be aluminum for the competition guy. But it is heavy. Lot of hardware in each wheel. We're a tad concerned about it...

Bill

[FotoPlasma]

Do you think you might be able to post a CAD or two of the wheel modules? I am very interested in how you implemented the coaxial swerve system.

[WizardOfAz]

Do you think you might be able to post a CAD or two of the wheel modules? I am very interested in how you implemented the coaxial swerve system.

Here's a photo of a pre-season prototype wheel, #0 as it were. Though the new ones have minor improvements (e.g. better spacers), they are essentially like this one.

http://crush1011.wookus.net/pix/2003_1203/swerveWheelPrototype.jpg

The motive drive comes down the center shaft, through the miter gear pair, then by chain to the wheel. The ratio is 1-1 all the way from the S-B motor up to the last sprocket, where it gets geared down by 1-2.8 (10 to 28 tooth sprocket). The vertical shaft is spinning in and supported by a half inch bore in the large teflon bushing.

The steering drive comes in the big sprocket on top. That sprocket is bolted through the large teflon bushing to the top plate of the yoke. Steering is accomplished by turning the teflon bushing in the top plate. The top plate mounts to the robot frame. The steering drive is also a 1-2.8 geardown, using the globe motor.

For size reference, the top plate is 6"x6". The large teflon bushing is 3" OD with a 3/8" groove where it rides in the top plate. The teflon splits in two pieces within this groove in order to make assembly possible.

We have some CAD drawings, but a few changes were made to the design post-CAD. Hopefully this picture makes it all reasonably clear. If not, please ask.

This system is not without its challenges, but we're working through them. First, it's a lot of gears and chain, and resulting end-to-end losses. Second, there is some coupling from the motive drive to the steering drive, which makes good steering tracking code more important.

It's very fun working on it though [;-)

Bill

[bjammin64]

hey, i was just wondering how exactly the driver will control all of this. would it mbe one person, or would both drivers have to work together to achieve topmost maneuveurability

[WizardOfAz]

hey, i was just wondering how exactly the driver will control all of this. would it mbe one person, or would both drivers have to work together to achieve topmost maneuveurability

We have a few different driver control options in the works, but all are single driver, at least for the "driving" part. Another operator will probably manage other gadgets on the 'bot.

We think the "preferred" driving mode will be a single joystick, where the joystick displacement from center is speed, and joystick angle is the drive direction. All four wheels steer in tandem to the same angle and always roll the same direction (never have to run the drill motor backwards).

We'll also be able to lock the back wheels and steer just the front, which makes it drive like a car more or less. This is one way we could rotate the body of the 'bot. Notice that in the "preferred" mode above there's no way for the driver to rotate the body. In "car" mode we'll enable the motors to go backwards for more maneuverability.

Another variation will be to steer the wheels in opposite directions and do a turn-in-place (a tank turn), then return to swerve for moving around the field. This gives a quick way to turn the body to a needed orientation.

Bill