OSR, me, and my photoeyes

Okay, so I have to admit something.. I never use OSR's, OSF's, or ONS's. Simply because the logic in using them eludes me. In my current small but extremely useful project, I am sure that a OSR could serve to fix my current dilema. Keep in mind this project is put together with spare parts. It wasn't done with the luxury of fast L/D and built in timer photo eyes. These are just basic retro-reflective photo eyes.

---PE1 [/]--------------------------TON1----

---PE2 [/]--------------------------TON2----

---TON1 [ ] ---TON2 [/]------------TON3----
----------------------------------|--------|
---------------------------TON3 [DN]--(OTL1)-

---OTL1 [ ]------------------------TON4----
----------------------------------|--------|
-----------------------------------TON5----

---TON4 [ ]---TON5 [/]-------------(OTE)---

---TON4 [DN]---TON5 [DN]---------(OTU1)---


I tried real hard If I have to screen shot it I will. At any rate, this is a part kick-off. If the bottom PE is made but the top PE isn't, it extends the cylinder.

The problem is the trailing edge of the part is not the same as the leading edge of the part, creating problems with my program.

Any suggestions?

Not completely sure of your problem but my first thought is if Time delay PE can fix your problem then the PLC Can.
Just Add Blocked and Clear Timers.

Time delay PE wouldn't fix my problem. L->D PE would. That way the only edge triggering the input is the leading edge of the part.

Ok, So if you set The Blocked Timer to 1/100 of a sec it will be done on leading edge. Then Set The Clear Timer to Time Value That the PE needs To Be Clear before you Decide you are done.

Not sure what L- >D PE Does. Please define.

What is it you are trying to "remember"?

I see some logic with 5 timers, but I don't know what your goals are in real world terms.

It looks like you are trying to latch a fault bit when PE1 is blocked and PE2 is not with TON3 allowing for tolerance, is this a fault detection type of thing? Or, is it just a timed sequence?




Describe what you are doing please.

At any rate, this is a part kick-off. If the bottom PE is made but the top PE isn't, it extends the cylinder. . .The problem is the trailing edge of the part is not the same as the leading edge of the part, creating problems with my program.

Click to expand...

Apparently you tried to detect the TRAILING edge, but would rather use the LEADING edge. The program attached runs on LogixPro to detect the TRAILING edge. Adjust the timers as needed for time between detection and kick, and the time needed to extend the cylinder. This all assumes that your kick can be done before the next part arrrives. If not, then you have to set up some type of part tracking or bit shift to track each part from detection to the kick cylinder.

PS: If you had rather detect the LEADING edge, switch Rung 000 PE1 to a XIC and PE2 to a XIO. I may have AssUMed the wrong set-up. If you could draw a sketch (side view) of your PEs and kicker, that might be more useful to determine what logic you really need.

Here is the logic I've already been using. It works great for the leading edge detection. I need it to ignore the trailing edge detection.



The timer presets are preset based on some math done against current drive frequency of the conveyor, so that if the conveyor speed gets changed by an operator, the KO will speed up and trigger earlier with it.

CWAL61: If I can avoid doing it that way, I am going to. Part size\contour\shape varies, as does conveyor belt speed, and I would have to change those numbers for each parts and\or incorporate more logic for adjusting it based on conveyor speed and part type.

Picture of the two KO's using this logic. Ignore the cylinder KO, that's going to be changed to an air burst.

Resized for Lancie's 1024x768 view

The timer presets are preset based on some math done against current drive frequency of the conveyor, so that if the conveyor speed gets changed by an operator, the KO will speed up and trigger earlier with it.

Click to expand...

That, and the first two debounce timers, sort of explains why you have 5 timers instead of 2. It seems that T4:41 and T4:51 could be combined into one timer.

If you are looking for leading edge detection, do you really need the debounce timers T4:45 and T4:46? It seems that if a photoswitch output goes on at all, it is probably going to stay on at least .02 seconds. Are these timer presets also changed depending on speed?

Lancie1 said:

That, and the first two debounce timers, sort of explains why you have 5 timers instead of 2. It seems that T4:41 and T4:51 could be combined into one timer.

If you are looking for leading edge detection, do you really need the debounce timers T4:45 and T4:46? It seems that if a photoswitch output goes on at all, it is probably going to stay on at least .02 seconds. Are these timer presets also changed depending on speed?

Click to expand...

Those timers are there because the bracket that the PE's were initially mounted on was flimsy and I was trying to find a way to deal with the PE's not being made at exactly the same time.

I was trying to find a way to deal with the PE's not being made at exactly the same time.

Click to expand...

Effectively, your Rung 016 is only going to look at the status of I:4/14, the bottom PE. The other PE could be on or off, it would not matter and you would get the same result if you removed Rung 015. If you want to always consider the status of BOTH PEs within some time period, that will need something different. If PE2 trails PE1, then it seems to consider BOTH, you might need 1 timer wired like this (using ony T4:45 to trigger your Knock-Off Latch):

 

|     BOTTOM PE               +-----------+    |

|---+---|/|--------------+----| TON T4:45 |----|

|   |                    |    |           |    |

|   |  TOP PE  T4:45/TT  |    |           |    |

|   +---|/|----| |-------+    +-----------+    |

Question:

If Top and Bottom PE's are both True anytime then part is good ?
True or False

Maybe something like this. This Logic requires kickoff to be done and Reset before next part detection.

Looking at Addohm's Photoeyw pictures, there really is no deliberate leading or trailing PE. Both appear to be vertically in-line. Maybe what you need is to look for either Normally Closed PE within some time window, then see if BOTH go on for some minimum time during that time window. Whichever PE is slightly ahead starts the time window, then you look to see if both go on sometime during that window. If so, latch your Knock-Off relay. Maybe something like this would work:

I think Lancies last example most concisely accounts for the photo proof and overlap, but you need to stick another timer in there to account for the physical offset distance (calculated from speed command or feedback) to the blow off nozzle.

Actually, just insert his code in front of your rung 17.

I normally have debounce timers set up to phase shift a signal for debounce of false inputs (set to zero for a latch) and then a delay off time to trigger the first action in the program, then cascade a second timer from the done bit of the first. As your rung 17 is written, both timers start together, so the duration of the output actually being turned on is the difference between the two. I would cascade the timers, start the succeeding timer with the DN bit from the preceding, and seal aroudn that DN bit with the TT bit for itself to ensure it times fully (for things like contactors and solenoid valves, you want a guaranteed minimum firing time). By doing so, you can adjust the delay without affecting the duration and vice versa.