PDA

View Full Version : pic: Mecanum Musings


CraigHickman
11-21-2008, 08:58 AM
[cdm-description=photo]32059[/cdm-description]

MrForbes
11-21-2008, 08:59 AM
neat design! although it looks like there's quite a stretch betwen the end of one roller and the beginning of the next....might be a bumpy ride

Andrew Schreiber
11-21-2008, 09:03 AM
neat design! although it looks like there's quite a stretch betwen the end of one roller and the beginning of the next....might be a bumpy ride

I like my robots shaken, not stirred.

Looks very nice, simple, and significantly cheaper than the AndyMark ones. How much does it weigh and what will the parts cost you?

CraigHickman
11-21-2008, 10:05 AM
I like my robots shaken, not stirred.

Looks very nice, simple, and significantly cheaper than the AndyMark ones. How much does it weigh and what will the parts cost you?

The weight is about .6lbs completed. I'm not sure about the total cost per wheel, but it should cost around $120-200 for enough stock to make 6 wheels.

M. Mellott
11-21-2008, 10:09 AM
Very interesting design! It's so simple, enough so that almost anyone could make these in a small shop! I'll be really interested to see these in action.

What size wheel would this equate to? Maybe 4" or 4.5" mechanum?

Andrew Schreiber
11-21-2008, 10:10 AM
Very nice, I will be interested in how these work out should you use them. At a nearly 85% cost savings over AndyMark (6 wheels X $96 = ~$600) and lower weight these could be useful. Keep us posted.

Stuart
11-21-2008, 10:16 AM
I like this my team does another comp other than FIRST called BEST and has been looking at how to implement mecanums with the limited kit you are given, I think this might be the solution. Really beautifully simple.

CraigHickman
11-21-2008, 10:33 AM
Very interesting design! It's so simple, enough so that almost anyone could make these in a small shop! I'll be really interested to see these in action.

What size wheel would this equate to? Maybe 4" or 4.5" mechanum?

The goal was definitely to have minimal impact on machining resources (gotta save all that complex stuff for the manipulator!), and keep cost down.

The effective diameter is about 4.375. It is a little rough ride, as the design goes, so I'll be doing some incremental changes to address the problem.

MrForbes
11-21-2008, 10:38 AM
What would happen if you tried to use a smaller hex for the central hub?

dtengineering
11-21-2008, 11:04 AM
Nice design! It is apparent that you took practical considerations into account as the primary design factor. There was a recent thread on CD, in contrast, discussing a mecanum wheel that the designer took aesthetic considerations into account as the primary factor... but as a result the wheel needed to be made on a 5 axis CNC mill in his opinion.

I can guess which one is more likely to show up on a robot!

As for the bumpy factor in the ride, you could always (at this cost) build 8 wheels, and attach them in tandem, with one rotated 30 degrees from it's partner... much the same way you can use two omnis to smooth out the ride.

Jason

Andrew Schreiber
11-21-2008, 12:33 PM
Nice design! It is apparent that you took practical considerations into account as the primary design factor. There was a recent thread on CD, in contrast, discussing a mecanum wheel that the designer took aesthetic considerations into account as the primary factor... but as a result the wheel needed to be made on a 5 axis CNC mill in his opinion.

I can guess which one is more likely to show up on a robot!

As for the bumpy factor in the ride, you could always (at this cost) build 8 wheels, and attach them in tandem, with one rotated 30 degrees from it's partner... much the same way you can use two omnis to smooth out the ride.

Jason

I remember that thread, IMHO this wheel looks almost as nice as it. Not as flashy but it has an elegance that certainly makes it beautiful.

Jason, would mounting two of them like that really work? I mean I suppose it would but has anyone ever done it with Mecanums?

JesseK
11-21-2008, 12:34 PM
For even less necessary machining, you could not mill out the inside of the hex stock, and instead thread-tap the holes where the counter-sunk screws go in. This makes it a bit heavier, but you wouldn't need a mill to make the wheel.

Then, you could also drill the KOP hub bolt-circle pattern into the side, eliminating the necessity of having to make a key way.

To take this even further, instead of counter-sunk screws you could use half-head-height machine bolts, eliminating the need to create the extra dimple after the initial hole is drilled.

CraigHickman
11-21-2008, 12:56 PM
What would happen if you tried to use a smaller hex for the central hub?

Not sure. Next time I boot my Mac into windows to do some CAD, I'll throw together a different version with smaller hex. But to be honest, I'm not sure I'd feel comfortable making it any smaller. The team I'm mentoring has an affinity for Banebots transmissions, so I think the goal is to keep to wheels that can be direct driven off a BB tranny.

For even less necessary machining, you could not mill out the inside of the hex stock, and instead thread-tap the holes where the counter-sunk screws go in. This makes it a bit heavier, but you wouldn't need a mill to make the wheel.

Then, you could also drill the KOP hub bolt-circle pattern into the side, eliminating the necessity of having to make a key way.

To take this even further, instead of counter-sunk screws you could use half-head-height machine bolts, eliminating the need to create the extra dimple after the initial hole is drilled.

There are definitely many revisions that could be made to seriously reduce machining time, but this is really just a 1.0 wheel. I'm debating changing the milling operation on the hex to be entirely one sided, so the "dishing" of the wheel would be done in a single Mill op. I'll have to do some FEA tests to see how the strength would stack up.

Also, for the rough ride concerns: Once I make the wheel wider, and thus the rollers longer, the wheel actually smooths out a good bit.

Akash Rastogi
11-21-2008, 01:52 PM
Woah, the assembly for these looks incredibly easy!:) Great design.

M. Mellott
11-21-2008, 02:13 PM
There are definitely many revisions that could be made to seriously reduce machining time, but this is really just a 1.0 wheel. I'm debating changing the milling operation on the hex to be entirely one sided, so the "dishing" of the wheel would be done in a single Mill op. I'll have to do some FEA tests to see how the strength would stack up...

I think what you have here is just fine. If you're using a CNC mill, the same program would be used for both sides, so there could be time savings there. Another option would be to incorporate a spoke pattern something like our 4" tread wheel design (http://www.chiefdelphi.com/forums/showthread.php?t=52944). It might help if you plan on making the wheel wider.

jwfoss
11-21-2008, 08:23 PM
that mecanum looks really good, reminds me alot of the WPI wheel from 2005 except a little smaller. Do you have a view directly from the side so we can see the complete arc of the rollers?

CraigHickman
11-21-2008, 08:46 PM
that mecanum looks really good, reminds me alot of the WPI wheel from 2005 except a little smaller. Do you have a view directly from the side so we can see the complete arc of the rollers?

Sure, let me grab a render real quick. I'll edit this post in a minute with it.

There's two renders here: My 1.0 ultra-cheap design, which is with 6 rollers. It's outer diameter is around 4, 4-5 inches. Next is the 8 roller one, a bit smoother, and a little bit larger with a effective diameter of 6 inches.

http://i37.tinypic.com/16k5kxe.jpg
http://i36.tinypic.com/3482jon.jpg

Alan Anderson
11-21-2008, 09:00 PM
Jason, would mounting two of them like that really work? I mean I suppose it would but has anyone ever done it with Mecanums?

I've driven a platform with doubled AndyMark mecanums, and it works just fine. It almost looks like they were designed to be able to be mounted that way.

Andrew Schreiber
11-21-2008, 09:35 PM
I've driven a platform with doubled AndyMark mecanums, and it works just fine. It almost looks like they were designed to be able to be mounted that way.

Ok, thanks, Ive just never seen anyone do that.

lbarger
11-22-2008, 06:02 PM
Great design. Do the nuts and bolt ends miss the carpet? You may want to look into tapping one end of the c-section and countersinking the clearance on the other. Then you could use a flat head screw and match the length of the screw to the c-section width. Then you should be able to avoid any hardware overhang and further reduce the weight. Typically you should have at least three full threads for any bolted joint.

Keep up the great work and please post your final version!

CraigHickman
11-22-2008, 06:31 PM
Great design. Do the nuts and bolt ends miss the carpet? You may want to look into tapping one end of the c-section and countersinking the clearance on the other. Then you could use a flat head screw and match the length of the screw to the c-section width. Then you should be able to avoid any hardware overhang and further reduce the weight. Typically you should have at least three full threads for any bolted joint.

Keep up the great work and please post your final version!

No, there's plenty of clearance for the hardware holding the rollers to the C channel. I'd really prefer to not take any more material out of the C, as the goal here was to go cheap and easy (hence going with aluminum from McMaster. If I go to something stronger (like a higher strength alloy), Then possibly I'd work on lowering the profile a bit.

I'd love to post the final version, but for me all designs are a work-in-progress. Maybe I'll post some renders side by side with a real thing if these actually get built? It's all really up to the team for what they'd like to build. I am just a mentor, anyway.

rutzman
11-22-2008, 06:43 PM
I am just a mentor, anyway.

Just a mentor? You say that like it's unimportant...

CraigHickman
11-22-2008, 06:49 PM
Just a mentor? You say that like it's unimportant...

I understand the importance of the position, but I see the limits in the spot. I'm here to teach students about engineering and smart decision making, not win them a competition. If that were my job, you'd be seeing a lot more designs up here that I "planned" on making, and justification on why they're the best.

Instead, I spend my time designing and planning parts and mechanisms with the intent of showing students that there are more possibilities than the most complex (high school students seem to be good at finding the most difficult solution first...) and expensive option.

A mentor should teach the team, not BE the team in my opinion.

rutzman
11-23-2008, 05:41 PM
You're completely right. My apologies for getting off topic.

Also, for the rough ride concerns: Once I make the wheel wider, and thus the rollers longer, the wheel actually smooths out a good bit.


How wide do you think the wheel would have to be to give a smooth ride?

CraigHickman
11-23-2008, 06:06 PM
How wide do you think the wheel would have to be to give a smooth ride?

About 3 inches wide, then it's pretty much smooth.

Here's a render of the latest revision. It's little bigger (effective diameter of 8ish inches), and uses more machining resources. to shave weight off.

Once I hit a revision that i feel comfortable putting on a robot, I'll be hosting the CAD for anyone that wants it, with only one rule: if you make it or modify it, you also have to put the CAD up. It's time to bring Open Source to robotics, folks.

http://i34.tinypic.com/2h4kco8.jpg

MrForbes
11-23-2008, 06:15 PM
can you render it with a viewpoint that's much further away, so the view of the side of the wheel is "flat"?

also....open source is nice! we did some this past season, I know we inspired a few teams

CraigHickman
11-23-2008, 06:21 PM
can you render it with a viewpoint that's much further away, so the view of the side of the wheel is "flat"?

I'm not understanding this, sorry.You're asking for a view that's in orthographic, not perspective, and from a "front" of the wheel view, or a "side" (front being looking at the narrow view, what I have in the middle there, "side" being the one on the left)?

MrForbes
11-23-2008, 07:25 PM
side view, as in looking at the side of the robot. The view in the left of your picture above. Change the perspective to infinite, so you can see how "round" the wheel is.

Brandon Holley
11-23-2008, 07:27 PM
About 3 inches wide, then it's pretty much smooth.

Here's a render of the latest revision. It's little bigger (effective diameter of 8ish inches), and uses more machining resources. to shave weight off.

Once I hit a revision that i feel comfortable putting on a robot, I'll be hosting the CAD for anyone that wants it, with only one rule: if you make it or modify it, you also have to put the CAD up. It's time to bring Open Source to robotics, folks.


I've spent some time designing mecanums (on a smaller scale), and I have my own method, but I was wondering if you would share the method you used to create the curvature of the roller. Specifically in CAD terms (I'm very familiar with any CAD system), what you did to ensure your roller curve was as effective as possible.

Thanks in advance!
Brando

D_Price
11-23-2008, 07:56 PM
The only question i would ask would be if they were to be bumped into would they break easier or would they still have enough strength.

CraigHickman
11-23-2008, 09:12 PM
I've spent some time designing mecanums (on a smaller scale), and I have my own method, but I was wondering if you would share the method you used to create the curvature of the roller. Specifically in CAD terms (I'm very familiar with any CAD system), what you did to ensure your roller curve was as effective as possible.

Thanks in advance!
Brando

Sure thing. Start with a center line that will serve as an axis to revolve the roller around. Next, draw your edge lines on the left and right edges of the axis. Put a third in the center, higher than the first two. Draw two arcs connecting these outer lines to the center line. Let me know if this is confusing and I can upload a drawing for you.

The only question i would ask would be if they were to be bumped into would they break easier or would they still have enough strength.

They're quite solid as they stand, but I don't like leaving a wheel uncovered for any reason. For that reason I'll probably enclose the wheels, most likely in 1/8th aluminum sheet on which the bumpers will be mounted.

Brandon Holley
11-23-2008, 09:28 PM
Sure thing. Start with a center line that will serve as an axis to revolve the roller around. Next, draw your edge lines on the left and right edges of the axis. Put a third in the center, higher than the first two. Draw two arcs connecting these outer lines to the center line. Let me know if this is confusing and I can upload a drawing for you.




I believe I understand what you are saying, it is similar to what I do as well, but a picture would not hurt... thanks!

CraigHickman
11-23-2008, 09:35 PM
I believe I understand what you are saying, it is similar to what I do as well, but a picture would not hurt... thanks!

Well, here's a Paint version... I don't entirely feel like booting into windows right now, so I can swap this image out for a screencap of the sketch some other time.

http://i37.tinypic.com/358c94n.png

Then you revolve it around that bottom line. After that you put a hole through the center for the bolt, then put recession into the roller for bushings.

Brandon Holley
11-23-2008, 10:45 PM
Well, here's a Paint version... I don't entirely feel like booting into windows right now, so I can swap this image out for a screencap of the sketch some other time.

http://i37.tinypic.com/358c94n.png

Then you revolve it around that bottom line. After that you put a hole through the center for the bolt, then put recession into the roller for bushings.

Yes i'm completely with you there.

To be a little more specific, I was referring to how you designated the actual arc on your sketch. Was it simply a large radius? A spline? If so, what did you use to define that spline, etc... Any more info would be awesome! Thanks again

CraigHickman
11-23-2008, 11:12 PM
Yes i'm completely with you there.

To be a little more specific, I was referring to how you designated the actual arc on your sketch. Was it simply a large radius? A spline? If so, what did you use to define that spline, etc... Any more info would be awesome! Thanks again

Gotcha. It's drawn as a Center Point Arc. The arcs were each 5 inches in radius (I think. I might be off a little.), and the entire roller is 2.75 wide. I'm still doing some math to see if really using an Arc is the right move. I might end up moving to a spline, who knows right now. I do know that we'll be pouring our own wheels into molds that we get 3d printed!

Brandon Holley
11-23-2008, 11:13 PM
Gotcha. It's drawn as a Center Point Arc. The arcs were each 5 inches in radius (I think. I might be off a little.), and the entire roller is 2.75 wide. I'm still doing some math to see if really using an Arc is the right move. I might end up moving to a spline, who knows right now. I do know that we'll be pouring our own wheels into molds that we get 3d printed!

Funny you say that....I use splines when I've designed mine

I actually have my wheel 3d printed already and am making a silicone mold of it, and then using that mold to cast my own hubs and rollers!

Tom Line
11-24-2008, 07:54 AM
I would strongly suggest that you make your rollers out of some non-lubricating plastic, then dip them in a rubberized coating to make them sticky. You can get the coating dirt cheap at any hardware store (it's what you dip tools in to rubberize them).

I say that because machining softer grade rubbers can be a real headache for the machinists. They tend to tear - even the harder durometers. So they end up having to freeze them then rush them to the lathe to get them done before they thaw. Like I said - real pain in the neck.

Plastic would be easier to machine, and in the end probably it would probably be cheaper too.

A step up from that would be to have a machine shop machine a female mold that you could pour your own rollers from. That'd be cool and not too hard - heat plastic in pan till liquid and pour in. A bit like pouring those old lead army men into casts.

Brandon Holley
11-24-2008, 09:20 AM
A step up from that would be to have a machine shop machine a female mold that you could pour your own rollers from. That'd be cool and not too hard - heat plastic in pan till liquid and pour in. A bit like pouring those old lead army men into casts.

From what he said, I believe he is going to do some sort of casting.

willson.thomas
11-24-2008, 02:37 PM
A step up from that would be to have a machine shop machine a female mold that you could pour your own rollers from. That'd be cool and not too hard - heat plastic in pan till liquid and pour in. A bit like pouring those old lead army men into casts.

In the past our team has had molds 3d printed by the local community college. We would probably do the same this year.