Pic: Easy to make mecanums..

[Tatsu]

Okay, as an offseason project, part of 1072 is making a mecanum drive.
first part (and probably the hardest) is the wheel.

here's our scheme for making one.. we'll make at least one over this break (its spring break for us) and report back how it is.. but if you see any glaring problems, please post asap.

Tools -
A Mill (Bridgeport in our case)
1/4 in end mill / drill bit
1 in , 3/4 shank end mill
Clamp set
Indicator
Vise
Edge finder

Materials -
3/4 plate 6 in x 36 in
Lab-stops (Mcmaster - 9545K15)
Generic 2 3/4, 1/4-20 bolts
1/4-20 nuts
retaining rings.

First, Take the plate, cut into 6 plates of 6 in x 6 in.
Next, bring to our friendly machinist, and have him turn the plate into the following pattern..
http://i27.photobucket.com/albums/c199/ATAT23456/Part1.png


Now, take back to the brigeport, tilt the head 45 degrees, and true the vise.
Then after we have that, we can just vise the plate using the octagon.. that will ensure true-ness on one plane (ie the y is now true) and then find the center of the octagon, and then use a 1 in endmill, then a 1/4 in end mill through to get..
http://i27.photobucket.com/albums/c199/ATAT23456/Part2.png

Then, since this form is quite awkward, after tramming the head to 90, we make some ornamental cuts.
http://i27.photobucket.com/albums/c199/ATAT23456/part3.png

Now there's the hub, the rollers should be equally easy. originally, we were planning on begging for someone with cnc lathe to cut us this profile..
http://i27.photobucket.com/albums/c199/ATAT23456/Part4.png

But, our student mentor realized that a lab stop, bored with a .25 in hole would be nearly identical in terms of circularity (the one and maybe even only key measure of wheel performance)
here's comparison of theoretical optical versus lab stops
http://i27.photobucket.com/albums/c199/ATAT23456/Part5.png

Finally, it should look like..
http://i27.photobucket.com/albums/c199/ATAT23456/Final.png

So. Any suggestions, comments, concerns?

[sanddrag]

Looks like a plan. It's clear you've thought quite a bit about it. Good luck with making it. I'd like to see the finished product!

[Bill_Hancoc]

So i thought about this for a lil while and thought of something. My expeirence with lab stop it that it was a very soft compound rubber that seemed to peel and cut very easily. I think i wouldnt be durable enough for driving around on carpet and pushing other bots around.

[Andrew Blair]

Looks good, but how I hate to see all that Al burned through! Seems pretty possible with no CNC stuff. Though a rotary indexer is a must.

[Madison]

It's interesting to read others' ideas for producing these sorts of wheels with minimal headaches. I, too, have been giving it some thought lately as a potential off-season project for our team and was trying to design a wheel that could be made using what we have available -- a 3 axis CNC mill, drill press and lathe.

Forgive my lack of step-by-step illustrations :)
http://img80.imageshack.us/img80/974/xwheel20ma.th.jpg (http://img80.imageshack.us/my.php?image=xwheel20ma.jpg)

[Bill_Hancoc]

Ok o searched this one and i couldnt find anything difinative. What is the difference between omni wheels and mercanum wheels. i know the power used i different depending on the direction moving but i thought i read somewhere the it is hard to push bots with mercanum wheels but easy with omni wheels.

any help i appreciated thanks

Bill

[EricH]

What is the difference between omni wheels and mercanum wheels. i know the power used i different depending on the direction moving but i thought i read somewhere the it is hard to push bots with mercanum wheels but easy with omni wheels.
The difference is the angle of the outer rollers. Omni wheels have their rollers round sides facing at 90 degrees to the wheel's roundness. Mecanums have their rollers at about 45 degrees (though this is variable). If a team has mecanums, you will have a harder time pushing them (due to the angled rollers and the power--four motors available versus two at any given time for omni's). However, it is still possible to push mecanums around, it's just a little harder.

[MikeJ675]

Two quick comments on the first design-

With the support for the rollers in the center, they may not turn as efficiently.(not too big of an issue)

Having the rollers positioned like that will make it easier to bend the supports(leverage and all that) if they get hit. Protect them well.

[Veselin Kolev]

Horay for lab stops and ghetto-ness. Hey, if it works, we'll let you guys know. If it doesnt, we'll make it work, and let you know :P

[mechanicalbrain]

I almost wonder if the design by NeilR (http://www.chiefdelphi.com/forums/showthread.php?t=40926&highlight=mecanum+sell half way down) wouldn't be easier to make.
http://www.chiefdelphi.com/forums/attachment.php?attachmentid=3745

[Tytus Gerrish]

I cant see your pictures while im here at work but from your description your manufacture method is very similar to the latest i've come up with (http://www.chiefdelphi.com/forums/showthread.php?t=46016&highlight=tytus+mecanum). the term "easy to make" has to be viewed very realitively. because as far as the basic wheel goes there's no real easy wayt to make it without the right tools.

[Tatsu]

update - Disks are done.. mill is cleaned and ready to go.. now all we have to do is make the jig (for accurate indexing) and start cutting... we should have a prototype soon.

Anyway, this could easily be done with a rotary table and a manual mill (no CNC needed.)

[Tatsu]

Victory!

Went milling today, got the wheels done. Here are some pictures. They still need to be cheeseholed, but meh. we've 7 more to go!
(sorry for low-res, taken on my camerafone)

Blanks
http://i27.photobucket.com/albums/c199/ATAT23456/Blanks.jpg

Initial setup
http://i27.photobucket.com/albums/c199/ATAT23456/MillSetup.jpg

Closeup of the jig
http://i27.photobucket.com/albums/c199/ATAT23456/SetupCloseup.jpg

First cut
http://i27.photobucket.com/albums/c199/ATAT23456/FirstCut.jpg

Second cut
http://i27.photobucket.com/albums/c199/ATAT23456/SecondCut.jpg

Assembled
http://i27.photobucket.com/albums/c199/ATAT23456/Assembled.jpg

http://i27.photobucket.com/albums/c199/ATAT23456/Final.jpg
(NB - the steel rule is a 6" scale)

We didnt really NEED a mill, we could have done this with a drill press =p. As long as you can get the blanks, you can make these mecanums with a drill press and bandsaw. but a mill is nice =p.

[Tytus Gerrish]

cool. nice to see them done. what machine are these destined to be a part of?

[Veselin Kolev]

These wheels are for prototyping. We are going to have two such wheels, and two mirrored (and of course spares) on a robot. They will be run direct driven off four KOP transmissions with encoders on a KOP chassis. This is to perfect the wheel design and code behind mecanum drive. If we deem the drive reliable and useful enough, we might just use this next year.

[Greg Perkins]

Do you have some cad files? If so, you should share them with us; I'm sure a lot of people would love to make their own.

[Tatsu]

we'll whitepapre and upload it once we get all the problems worked out.. we've found a few, we're in the process of fixing them.

[DonRotolo]

but a mill is nice =p.
Sweet! Very nice process.

What, if any, bearings are used on the shafts that hold the lab stoppers?

I'd also be interested in hearing more about the durability of the stoppers.

Don

[Veselin Kolev]

Sweet! Very nice process.

What, if any, bearings are used on the shafts that hold the lab stoppers?

I'd also be interested in hearing more about the durability of the stoppers.

Don


As of now the lab stops are on bolts. This was just for the picture :P
We are in the process of ordering a bunch of .250 ID 1" length bushings to put in all the rollers. They will then run on a ground steel shaft that is pressed into the wheel housing, with E-rings on the edges to keep the rollers from sliding off. We know that the wheels are slightly fragile with the rollers outside the wheel like that, and the steel roller shafts would be prone to bending. But like most robots, we plan to have our wheels well protected from all sides. The wheels can take huge forces radially, because if pushed hard enough the rollers will compress and the wheel bottoms out against the aluminum housing. This takes about 150-200 lbs per wheel to do.

A side note, efficiency is not reduced by weight on the wheel (other than more friction in the bushigns) because there is enough clearance on the sides of the rollers to allow for expansion due to weight compression. Also, these wheels get very good traction because the rollers are made of gum rubber. This means that all the rollers will have to be replaced every regional, but at 25 cents a piece from mcmaster, I dont think that will be a problem.

Our goal from this is basically to have an agile mecanum robot with lots of traction and pushing power, something not common with omni-directional robots. Look forward to designs for ball shifting transmission for each wheel...

[Greg Perkins]

For some reason I'm contradicting myself, I've designed a set of these before, but I think I might have done them wrong; what angle are the rollers supposed to be offset to the hubs?

[Madison]

For some reason I'm contradicting myself, I've designed a set of these before, but I think I might have done them wrong; what angle are the rollers supposed to be offset to the hubs?

Nearly any offset will work in theory, but 45* is most efficient.

[Alex.Norton]

Its not that 45* is most efficient. It is that 45* results in the max speed to the side is equally to the max speed forward. They are actually less efficient than say 30* because the rollers need to turn faster when moving in a direction other than forward. Of course 30* also means that the max side speed is lower than the max forward speed and the pushing power is greater.

[Kyle Love]

A little birdy told me that free mecanums are always the easiest, but if you don't win them these are pretty darn cool!

[Madison]

Its not that 45* is most efficient. It is that 45* results in the max speed to the side is equally to the max speed forward. They are actually less efficient than say 30* because the rollers need to turn faster when moving in a direction other than forward. Of course 30* also means that the max side speed is lower than the max forward speed and the pushing power is greater.

By 'most efficient,' I meant simply that a 45* offset allows for the fastest perpendicular drive speed. It's true that you can optimize the offset angle to better suit the conditions under which the drive will operate.

[Dillon Compton]

Question:

Most mecanum wheels have the rollers at a 45* angle on 2 axes (I think you all know what I mean, at least, I hope it's clear: the rollers are at a 45* angle to the wheel and then angled upwards, correct?)

Is this necessary, is it simply for increased traction, or could one theoretically design a mecanum wheel more in the style of an omni wheel, simply rotating the rollers to a 45* angle instead of a 90* angle?

[Simon Strauss]

Is this necessary, is it simply for increased traction, or could one theoretically design a mecanum wheel more in the style of an omni wheel, simply rotating the rollers to a 45* angle instead of a 90* angle?

Yeah, take a look at this (http://www.araa.asn.au/acra/acra2002/Papers/Diegel-Badve-Bright-Potgieter-Tlale.pdf) report (where i got most of my Mecanum information from) specifically the image on the first page titled the conventional Mecanum wheel. I think that is what you are talking about and although not as cool looking as the others, probably the easiest Mecanum to build if you don't have a CNC mill or, mill master and don't want each wheel to be a multi piece wheel.

[Dillon Compton]

Yeah, take a look at this (http://www.araa.asn.au/acra/acra2002/Papers/Diegel-Badve-Bright-Potgieter-Tlale.pdf) report (where i got most of my Mecanum information from) specifically the image on the first page titled the conventional Mecanum wheel. I think that is what you are talking about and although not as cool looking as the others, probably the easiest Mecanum to build if you don't have a CNC mill or, mill master and don't want each wheel to be a multi piece wheel.

Is the oblong roller necessary as well, or could one do something more along the lines of an AM trick wheel with the roller at a 45* angle?

[Alan Anderson]

Is the oblong roller necessary as well, or could one do something more along the lines of an AM trick wheel with the roller at a 45* angle?
The shape of the angled roller is intended to make the overall profile of the wheel a smooth circle. If that's not done, the wheel will be somewhat "bouncy" in operation, and it won't have consistent traction as it turns. If the inconsistency is severe enough, it will be difficult to maintain a desired direction of travel.

[rachal]

The finished mecanum! (Sorry about the glare)
http://i70.photobucket.com/albums/i100/wysiwyg90/finished-mecanum.jpg

A comparison with the spares.
http://i70.photobucket.com/albums/i100/wysiwyg90/mecanum-comparison.jpg

Much lighter now :D

[techtiger1]

Excellent job 1072 very nice machining process. I also love what you have done on being able to lighten the wheels. I like the fact that 1072 makes most of the stuff they design and has some great initiative about things like that. Finally, amazing pictures.

-Drew

[rachal]

Wow...we've been extremely busy with other things for the past few months, but we're finally getting close to wrapping this project up (the mechanical aspect, at least...)

http://i70.photobucket.com/albums/i100/wysiwyg90/final-mecanum.jpg
Looks just like the CAD, doesnt it?

http://i70.photobucket.com/albums/i100/wysiwyg90/assembly.jpg
We're using a kitbot to test the wheels for now. (two of the wheels are still uncheeseholed, which is why they're hiding in shame under the other two =P)

[Kingofl337]

Team 40 used these last year. The biggest thing I can stress is make sure the chassis is a square. If the chassis is rectangular you will have a difficult time turning. If you check out checkmate40.com we have some video's of the robot running. It looks like its floating when it drives but doesn't stand up well to being pushed.

Bottom line, if you are looking for a robot that will push well or stand up well to being pushed you are looking at the wrong type of wheel. See the 6 wheel drive thread.

We do not plan to use them again this year unless the game dictates we should.