5370 L2 Controller

Ok so here is a recent project I have been tasked with which includes using a Fanuc robot. My questions are just for the PLC programming portion in that as the robot is running it will be picking and placing parts into a plate that holds 30 parts. Now I can't have 30 part present sensors to detect when each station is empty or has a part.



My current train of thought is that I will have to use the sensor where the parts originate from (one part at a time)as not being there, as well as the reed switch on a cylinder indicating that the cylinder is forward and the robot gripper closed as having grabbed the part. Then when the robot gripper switch is open, that would indicate that it has placed a part into the 30 part tray. That condition is what I will use to set a latch that there is a part in position X on this tray. As such the parts will then be removed from the tray based on FIFO, but each part has to sit in this tray for 5 minutes before it can be removed. Once the part is placed into the tray that will start a 5 minute timer, when the timer is done that will indicate that the part is ready to be removed from the tray.



Is there an easier/more efficient way of tracking parts as they are placed into this tray? I feel like the way I am currently thinking about is far more complicated that it needs to be, but I don't know of another way. Any assistance will be greatly appreciated, conversely if more info is needed please feel free to ask! Thanks

Mr. bbishop108,

While you can certainly use logical FIFOing schemes to track parts, it is always good advice to use physical sensing (in your case, all thirty (30) locations) instead. Why, because the sensors will give you a real-time representation of your systems' conditions allowing you to react (stop, debug, and possibly correct situations that are going to come up). Otherwise, your FIFO logic only allows you to assume what is going on and in the event of a misload, jam, etc., you will have no way of reacting until your assembly line is a mess of jumbled and possibly damaged equipment and product. Use as many sensors as you can, I know it is a pain, and the costs in I/O and the resulting channels go up, but in the long run, I believe you will be happier with your assembly line's performance.

I hope this comment allows you to make the best decision for you.

If you don't want to, or can't because of cost, go down the route of 30 sensors indicating "part in tray", then an easier alternative is a "part in robot gripper" sensor.

1. It will tell you that the robot can pick up a part (i.e. the last one is not "stuck" in the gripper).
2. It will tell you that the robot has successfully picked up a part (the sensor will come on).
3. It will tell you that the part is present all the way from pickup point to release point (i.e. not dropped on the way!).
4. It will tell you that the robot has successfully delivered the part to the tray (the sensor will go off).

daba: This is how I am currently planning to do that as the robot gripper will have indications for 3 positions: Open, Closed on Part, Closed no Part. I am considering using the FSC instruction to know where the next open position is on the 30-part tray, so after the gripper closes on a part I will trigger the FSC to search for the next open position and send that info to the robot. After the gripper changes from closed on part to open, that is what I plan to use to indicate that a part has been placed into the 30-part tray. This will then trigger a 5 minute timer, after the timer is complete that will indicate that the part is ready to be removed from the 30-part tray.

daba said:

4. It will tell you that the robot has successfully delivered the part to the tray (the sensor will go off).

Click to expand...

What if a partial part is stuck in the tray and the part is dropped off but not correctly? Or there is an allignment problem and part is released but doesn't sit correctly? I agree that this is better than nothing, but being able to measure the entire process is what you should always push for.

Too many companies are willing to save money up front in return for paying much more in labor down the road. If you can convince the bean counters that it will cost less overall to have all the sensors, I would go that route. You could also examine using a camera vision system to track load progress. It would be especially easy if the contrast between the part and the tray is large (black and white for instance).

LoganB said:

What if a partial part is stuck in the tray and the part is dropped off but not correctly? Or there is an allignment problem and part is released but doesn't sit correctly? I agree that this is better than nothing, but being able to measure the entire process is what you should always push for.

Click to expand...

I don't think that "not sitting correctly" is an issue here. From the description given, the part just has to remain in the tray for 5 minutes, presumably before another (or the same) robot picks it up again for the next part of the process.



The critical thing is that the robot has it "located" in the tray correctly, doesn't "drop" it in, and releases it cleanly, then it will be where it left it when it comes back to being picked up again.


You can throw an awful lot of money at something when all is needed is proper programming of the robot..


my 2c

Why have I suddenly started get double line-feeds ?


I normally press enter twice to start a new paragraph, once to end the line I've just finished, and once again to insert 1 blank line, but it's now producing 2 blank lines... as above.

If a robot is going to be picking the part up after it sits for 5 minutes, then there most certainly could be issues with it not being placed correctly. Unless the robot is using a vision-guided pick, it will return to the correct location to pick up the item as it would sit correctly on the tray. Any variation from that position can cause issues.

Obviously this is just theory based on my history working on jobs. This particular application is mostly a mystery at this point, so I'm not saying it's required. I would need to know more about the application. I don't know if the orientation of the part is important, or just the position, for instance.

LoganB: As for the application I basically taking a the parts we make here and the robot is loading them into a heat plate (30-part tray) where the parts have to be heated at a specific temp for 5 minutes. The parts have a round shape and they are placed into this tray where the opening will be enough to allow the robot to place them into the hole on the tray...place them, not drop them. There will be a lead in into the hole so even a small offset will still allow the part to be placed into the tray. The top portion of the part is much larger than the bottom portion, which is how the robot grabs the part and loads it into the tray. As such the spacing between parts is very minimal, which is one reason why we cannot have part detection sensors for every part.

bbishop108 said:

LoganB: As for the application I basically taking a the parts we make here and the robot is loading them into a heat plate (30-part tray) where the parts have to be heated at a specific temp for 5 minutes. The parts have a round shape and they are placed into this tray where the opening will be enough to allow the robot to place them into the hole on the tray...place them, not drop them. There will be a lead in into the hole so even a small offset will still allow the part to be placed into the tray. The top portion of the part is much larger than the bottom portion, which is how the robot grabs the part and loads it into the tray. As such the spacing between parts is very minimal, which is one reason why we cannot have part detection sensors for every part.

Click to expand...

So long as there can be no movement of the parts in the tray, either by mechanical handling or human intervention, you should be good to go with the single part sensor on the robot gripper.... In effect, it becomes a single-pixel "vision" system..