I finally have my mecanum bot up & running and did some testing today (http://youtu.be/81F_xTLLb2w). I found the bot drifting due to a lack of traction (not too bad on the video, but erratic behavior elsewhere in the living room ;-). The alloy frame is pretty rigid causing not all four wheels always to be firmly on the ground. Looking for solutions I encountered the idea of dividing the chassis in two parts that can rotate (twist vertical) relative to each other (see pictures). My question regards the "rotation component". What is this for component and were can I buy it. Any other suggestions resolve my suspension problems?
I would use the rear wheels, I would use the rear wheels, if they are casters, like that, rather than the front wheels. It works on wheel chairs…
I would love to know of a pre-made component, but for now I would make one with a bearing and a sturdy piece of rod mounted between the two part of the robot. I yield, however, to the more mechanically minded among us. This could be a question Roxanna.
I had contact with the constructor of the alloy chassis and would like to share his views with you.
For a meccanum robot, I’m not sure that the rocker joint I used in the Scout would be the best option, but it’s worth a try. You might need something with more independent suspension - so each wheel can adjust itself to the ground on its own spring/damper system. You can look at what RC cars use as a reference.
The Scout system is a custom pillow block - just a small block of aluminum with a precise hole drilled in the middle, and some tapped screw holes for mounting. Your research is correct in that you don’t want a self-aligning bearing. For my design I did not even put a bushing or a ball bearing inside. A direct metal shaft to metal hole is called a journal bearing. These can be lined, or in the simplest case unlined, as I have implemented. To reduce part count and provide easier assembly, I use a shoulder bolt, which has partial threads at one end, a precise smooth diameter in the middle, and the head at the other end. Because the bolt will squeeze the two pillow blocks together, we need someway to reduce this friction. I used a thrust washer (like http://www.sdp-si.com/web/images/product_images/Complete_Thrust_Bearings.jpg) which is designed to take an axial load and still permit rotation.
My robot was small and lightweight, and used a 1/4" (6.35mm) steel shoulder bolt for the pin, and 6061 aluminum for the pillow block. Depending on how much weight you will place on your robot, this could still be sufficient. The exact materials aren’t too critical as this is for a prototype. If you are making a product, you will want to be careful about rubbing aluminum and steel together - you can research this to find out why. I also place a little grease in the joint so it turns smoother.
I did prototype a spring mechanism that would always try to keep the two halves of the robot aligned/centered, but ultimately found it to be unnecessary. What I did need to add was rotation stops to prevent the halves from turning too far past each other. This was just a simple hard stop by a pin or small bump on the robot chassis.
I now understand the mechanics of the joint and will look into his RC-car suggestion. After that I will decide how to proceed from here. Keep you posted.
