The most persistent and hindering problem i have with any of my automation tasks is presentation. I was wondering how everyone else out there solves this problem. The fact of getting part reliably in the same spot and in the same orientation. One of the only thing hindering me from tripping the amount of work load that is done by the robots. So, I just wondered if you guys had any suggestions, or just this that help with this issue.

We use a variety of methods from hard automation where parts are being presented to the robot automatically in the correct orientation to vision guided vibratory tables. We have one system where the humans load a large stack of parts and then a pneumatic gantry robot picks up a part and places it in a nest that the UR uses to assemble and pick up the part for the next operation. We have several vibratory feeders where we use vision to locate and pick the part. Some of these are multiple robot setups where one robot is picking random parts off the table and then orienting them for a second robot that uses vision to locate a feature on the part to place it in the assembly. We have two robots for cycle time, sub 4 second part to part pick times

Part presentation is definitely one of the biggest challenges we face on a day to day basis. The issue is finding a reliable way to locate parts that meets the requirements of the robot while not overburdening the operator. We have some cells where the operator loads a part or a few parts and cycles the robot. The robot is then able to process the work while the operator loads a part into another fixture or another robot. This is good for quick deployments as well as lower volume jobs where it’s hard to justify spending upwards of $25k on a feed system. 

@matthewd92
My problem is that i work at a small job shop, were the largest jobs are normally in the 10,000 range. The robots are moved from job to job and i have to find a way to hold and retrieve the parts for each job differently. They can be as small at 1/4 of an inch, and the largest job that is currently running is a 15,000 part order or 11" by 6" Running at 12 seconds a part. The question i have is what direction would be best for my company? In all reality has to be something that is flexible enough that the part presentation can be changed quickly and smoothly. The majority of the time i spend on automation is part presentation. I'm just looking for some good input from anyone. Thanks in advance -JS

@JSchena Have you thought about using vision to solve the problem?  You could use a Robotiq wrist camera if you have the cycle time needed to find a part before picking it.  Another option would be to use a camera from SICK or Cognex and fix mount the camera over the part table so that you can find the part in a background thread while the robot is moving, that way as soon as its ready to pick you already know where to go to get the part, minimal delays occur this way.  Both SICK and Cognex have URCaps to help make using the camera much easier, the Cognex one is new but is supposed to be really good.  Our preference is usually to fix mount the camera because we are usually pressed for time in the cycle, running pick and place as fast as 2.9 seconds, pick to pick with a UR3 and UR5.

We are working on a project currently with a wide variety of parts to pick and so will be going down the path of a backlit pick table that will allow us to put any variety of parts on the table and because they are backlit we are just looking for black blobs on the table.  Only disadvantage with this approach is we can only handle a single layer of parts.

@matthewd92 The problem is i have to prove the functionality and versatility to justify paying for something that expensive. My biggest problem with most vision systems is that the parts i am handling are normally complex small and randomly thrown into a large container. As of now parts have to ether get stacked, or be stacked to run on one of our universal robots. Most of the work in the shop could be integrated, but the parts are ether not stackable or too irregular of a shape to be presented in a manner that is process-able. 

You mentioned having the vision system over a part table, how do the parts get processed before hand? Is there a vibrating feeder hopper, or is it something more complex and more reliable. vision is something i have always thought would be the way to go, and i have been looking into t for some time. The only problem is i cant find the correct way to get it started, or to find the right system that is versatile enough to be use full in this job environment.