Let's Make Robots!

Mecanum Bot

Driving around

I bought these mecanum wheels from sparkfun and its manufacture by FingerTech Robotics. When I got them it came with four set of wheels. One pack is missing all the washers and other one has misplace the hole and other two pack was fine. I contact sparkfun about this issue and they don't response at all. I guess they don't want to deal with such "small" issue.

Anyway, I bought my own washers and fix the misplaced hole with epoxy. Then, that video us the result I got. The way I program it suppose to move horizontally(going left or right side way) but it's kinda driving strangely. I don't know it was originally not well designed or the way I fix it is not proper. But iI figure if I have more weight on it and it works a bit better.

Image below is the hole in wrong spot so I couldn't fit the shaft in.

I might put a cover on top of it later, meanwhile still working on the program. The most painful experience to program it is the driving direction confusion that I don't even how to explain how confuse that is since its base on 4 independent driving and all thses 4x4 comonation can generate at least 16 different results.


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.


Slowly collecting data and reading all I can before attempting to build my first bot.

I've been eyeing those Mechanum wheels because I dont really have large tile floor area...mainly carpet in this condo. For what those wheels cost, I would have had a major sad face based on that first vid you posted :) So I am interested in knowning if you tested those wheels on carpet as well and what your results were.

As far as FingerTechRobotics comments about adding more weight to the chassis: would that add more "load" to your servo motors and run down your power source? It's all new to me :)  

Anyway, nice Bot!  


You're welcome!  My biggest goal is to make robotics easier and accesible to everyone.  Can't wait to see yours driving around!

I have replied to hardmouse's email but I will post some tips and video links here in case anyone else is wondering too.

First, very knowledgable commentors!  I will confirm some of them.
- Yes, the wheels should make an X when viewed from above.  The instructions included with your set will have shown this, and they will assume you have it set up this way when describing how to spin motors to move in different directions.
- Weight distribution is important, and make sure that all four wheels have equal contact with the ground.  If one wheel is lifted you will not get sideways movement.  A flexible chassis (that can twist in the middle) is one way to do this.  You could also use flexible motor mounts, except that Tamiya gearboxes are rigid so this won't help your robot.
- Adding weight could help.  The test chassis we use weighs 3kg and we have run it up to 13kg on top.  If there is not enough weight the rollers will bounce and skip across the ground, especially at high speeds.

Personally, I think a compass IC module would give better results than encoders for these wheels (assuming the motors all spin the same speed to begin with).  The roller shaft holes are drilled by hand and though the company did a very good job, not all holes are perfect (as noted by hardmouse's roller that didn't fit [which is being replaced]).  This results in some slippage when driving, and using encoders assumes that each wheel rolls and slides excatly the same.  Using a compass to keep the robot always pointed in the direction you want should give better results.  This is just my theory though.  I plan to test it when I get some free time.

Here is the YouTube video of our R/C Mecanum chassis: http://youtu.be/DhNm0FOIiG0
In the video you will see how the wheels should rotate to achieve the different directions.  That chassis has no microcontroller and only used two channel mixers to make driving easier.  There is also a mecanum mixer available at Robot Logic but it is expensive.

SparkFun's Mecanum test platform (they had to verify that the wheels worked before selling them): http://youtu.be/j95nKHO6CKc

This video shows the difference between dry rollers and lubricated rollers: http://youtu.be/AbjzTDQtWcA
Lubrication is important!  It was using DAGU's Rover 5 chassis (video courtesy of OddBot).

And just for fun, here is the video SparkFun made when they started selling the wheels: http://youtu.be/hLy0_xzbSa4

Thanks for fast reply after I send the email last night. I think Sparkfun is way to busy to take care of all email requests.

Compass IC is a good idea too even I never use them before. Thanks!!

Have you tried swapping your wheels front for back? I had not been familiar with Mecanum wheels before, but found a number of videos of them on YouTube. It appears to me that your wheels seem to be working in opposition to each other, which made me wonder if they might be swapped front for back.

How grippy are the rollers on the meccanum wheels?

You may want to try a surface with more friction such as short carpet.  The rollers also have to roll freely with side-to-side pressure which it should if they are greased and have smooth edges.  Encoders will also get you the best performance since you can acheive symmetrical rotations, and you may want to use some small weights to balance it on center with some sort of balancing bar for X and Y axes.

As I recall, the top roller axles make an X pattern using the two different pairs of wheels (yours don't make an X).  The right front and back should oppose in one direction, and the left front and back should oppose in the opposite direction to get lateral movement (at least thats what I recall).

This guy printed his own and posted the stl files to share, download is at the bottom of the page.  I printed a pair and they look nice, but I've got to make the wheels grippier than the abs plastic surface.

Yeah, I notice my wheel looks like Dimaond shape instead of "X" and I reverse ir back to "X" but same result. Encoder might helps in this case but I don't have encoder handy to try it out. So cool you guys can print your own version out from 3D printer.

So you have the X shape, the correct motor directions for all four motors to achieve lateral motion, greased rollers, and enough friction between the rollers and driving surface?  Even without encoders this should still work much better than shown in the video, it just won't be as straight in the sideways direction as when encoders are used.  It may also help alot to center the weight if your bot is not already reasonably well balanced.

I think friction can be a large determining factor in performance, we may print additional rollers with grooves for orings to circumscribe to get more friction for lateral motion. 

It looks like they don't get good traction on the tile floor. Have you tried it on other surfaces? Maybe try slowing down the wheels so they don't spin so fast.

FingerTechRobtics posts on this site a bit. It might be worth contacting him and letting him know about the issue with SparkFun. If they are shipping defective product, it should be resolved.


Yeah, slow it down and have a bit more weight helps but still looking funny. Thanks for the link~