Heat tunnel jam detection using FIFO

monkeyhead · May 19, 2005

Attatched is the logic for a dumb little project i'm working on.

I tested it today and it woks fine other than needing the photo eyes debounced and a some logic to compensate for the clock rolling around to zero. It just seems like there should be an easier way.

Anyway basic setup:

products take 9 secs to travel through the tunnel.2 photo eyes, one before the heat tunnel and one after. packages arrive at completely random intervals.

1. first eye copies the absolute value of the free running clock + 10 secs to a fifo stack.

2. second eye unloads the fifo stack.

3. the absolute value of the free running clock is compared to the first value on the fifo stack continuosly. if it exceeds the value the package is assumed to be lost in the tunnel and an alarm is set.

Just curious to know how other people would have handled this.

Lancie1 · May 19, 2005

Monkeyhead,

It looks efficient. An encoder or position indicator on the conveyor is the only thing I could see that would have made it any simpler.

monkeyhead · May 20, 2005

Thanks Lancie. I just didn't like the way I had to handle the clock. I have this neurotic aversion to output instructions without input instructions.

monkeyhead · May 20, 2005

dang, no takers?

bernie_carlton · May 20, 2005

I have one small question. If a part arrives just before clock rollover its "plus ten seconds" point may end up being much smaller than the current clock. It would seem that this part would be incorrectly tagged as "lost". I can't tell if this scenario is being addressed.

monkeyhead · May 20, 2005

that's exactly what i didn't like about this setup. i'm gonna have to add some rungs to just to handle the clock rollover. which isn't a big deal, but it annoys me having to spend so many rungs messing with setting up the clock.

my other ideas for this project sucked though, so i ended up sticking with this one. the second best i came up with involved a counter and ten seperate timers. that's why i got curious to see how other folks would have dealt with this.

Terry Woods · May 20, 2005



           +-INC-----+

  IN-EYE   |  Count  |

---| |-----+    +1   +--------(    )  Reset Timer

           |  -----  |

           |  Count  |

           +---------+





           +-DEC-----+

 OUT-EYE   |  Count  |

---| |-----+    -1   +--------(    )  Dummy Bit

           |  -----  |

           |  Count  |

           +---------+



                   

 Count     0    +-------+  

------| > |-----+ TIMER +-----(    )  Timeout

                +-------+

(Timer set to 19 Seconds)





 Timeout    Count     0     

---| |-----------| > |---+----(    )  ALARM

                         |

 Count     "X"           |

------| > |--------------+



"X" equals the max number of parts that can fit in the tunnel.

.
Add the debouncers etc to finish the job...

monkeyhead · May 20, 2005

Thanks Terry, that's retty slick. My rube goldberg solution is just about all finished up now or i'd have to steal it.

So to absolutey worst case scenario this (not trying to pick it apart, it's just my first step when looking at logic):

Product 1 enters and jams. Just before the timer expires, product 2 enters reseting the timer, this trend continues (it's a slow day on the ol' packaging line) until count reaches X, leaving the first product in the heat tunnel for slightly < X * 19secs?

is that right or am i missing something here?

Terry Woods · May 20, 2005

monkeyhead · May 20, 2005

and the final version of my glorious mess... thanks for humoring me everybody.

rsdoran · May 20, 2005

I am a simple person so all that FIFO stuff seems elaborate to me. Why couldnt you use an up down counter and reset a timer on exit when count is greater than 1? The following picture is an example, it may also need a zero move to the counter accumulator on startup and after clearing jams. If necessary you could set a range..count below 3 use 10 second timer, count above 3 use 5 second timer...not sure how many packages could be in the tunnel. When troubleshooting this is alot easier to understand.

monkeyhead · May 20, 2005

you're right... it's totally elaborate. i was just having trouble coming up with a good simple idea, so this is where i ended up.

one of the reasons i like coming here is seeing some of the simple yet so effective techniques people come up with since i have a tendency to overcomplicate everything i do because i'm too busy obcessing over doomsday scenarios.

thanks though, i appreciate your insight. simple dense code like you guys are showing me is where it's at.

GregPLC · May 20, 2005

No offense, but Terry's and Ron's would be so much easier to troubleshoot.....I do give you an A for thinking outside the box a bit with the FIFO load and unload for this application. Makes a headache for an average maintenance guy though...

Greg

monkeyhead · May 20, 2005

No offense taken and exactly the reason I started this thread. Thanks for the A, but the real question is, is it worth a raise?

Now it's on to the more enlightened part of the night...

rsdoran · May 21, 2005

No offense, but Terry's and Ron's would be so much easier to troubleshoot.

I can not take any credit for this. When it was first posted I opened RSLogix and started a program but mindlocked. I read Terrys post and conciously it didnt sink in but undoubtedly subconsciously it did because right after that I created the counter/timer code.

Basically all I did was convert Terrys ascii into RSLogix but my code didnt offer all the benefits that Terrys did. Attached is a pdf of the code using RSLogix with more features.

Feel free to tell me if you see any problems with it.

Heat tunnel jam detection using FIFO

monkeyhead

Member

Lancie1

Lifetime Supporting Member

monkeyhead

Member

monkeyhead

Member

bernie_carlton

Lifetime Supporting Member + Moderator

monkeyhead

Member

Terry Woods

Member

monkeyhead

Member

Terry Woods

Member

monkeyhead

Member

rsdoran

Lifetime Supporting Member

monkeyhead

Member

GregPLC

Member

monkeyhead

Member

rsdoran

Lifetime Supporting Member

Similar Topics