ethical question

Well,

I'm very new to programing and am working on SLC 500's I've created a simple parts counting program that compares the number of flights on the conveyor to the number of parts going down the line

Every min. with a timer I write them to an interger then divide them to show line efficientcy.

Alls good so far, but I'm having a problem with double counting parts, Mostly due to extra material going down the conveyor.

A sensor change is possible and some more adjustment (some)can be done.

The qusetion is: Do I do some programing to filter out some of the extra counts and call it good, or continue to impale myself on the project? (this is in house and not a project for sale)

The good news is, that I am sampling every min. with excel and doing the final averages there, so on an 8 hour day at 480 samples the extra counts are'nt making a huge diffrence in the end result, except that it looks odd to make an extra part thats not there.

The results are for efficiency monitoring and not an inventory madder

Thanks Clint

I'm not one hundred percent sure I understand your post, but this might be the issue:

Every min. with a timer I write them to an interger then divide them to show line efficientcy.

Click to expand...

When doing integer division in a SLC500 the result is rounded, not truncated. Thus 3.5 becomes 4.

If I completely misunderstood you then perhaps you can post your program and I'm sure that the board can help identify the problem.

CJones said:

I'm having a problem with double counting parts,

Click to expand...

What are you using to sense the parts? If it's a photo eye is it possible you are getting more than one turn on per unit of product. Like you want to count a hand but the photo eye sees the fingers. Again if it's a photo eye can people walk infront of the eye?

Here is an interesting oddity you are getting 480 extra counts in an eight hour day and you are comparing every minute. How many minutes is that? Hmmmm!

I'm having trouble with what is meant by this comment?

CJones said:

Mostly due to extra material going down the conveyor.

Click to expand...

Interesting, about the rounding, I will try to figure out how to post the program, at times I am getting for example 30 flights and 33 parts, there could only be one part per flight, the bad part is that there is the possibility of extra material at any givin time, ususally when the line is running at say 80% the program works great, the other caveat is that I only have an inch to count the parts (the inch sticks up above the flight)

Let me try to post the ladder

Clint

The Counting Part is named "Debone" there are seven Conveyors using 10 rungs each.

The "staff sensor" is the flight and the "bird sensor" is the part.

I may need another sensor I'm using SMR31QD and an opposing SMR31QD with Y cables strait to an 1746-IB16, I'm also wondering If I shouldn't use a relay before the input to make sure I get a clean On/Off?

Well at 8 hrs excel is logging every min. (480 min) extra material= chicken parts and skin and grease and water and and and and and.

I am sure I have a small allignment problem

thanks clint

Try using one shots before the counter(s), this may help with clean on/off.

Not sure yet why you created LAD6, may see it after looking closer. I personally would use the inputs directly. If I was doing it this way I would want the inputs scanned at the beginning of the program, not at the end. I am not sure if it matters or not in this case, maybe someone will clarify that.

You seem to be assuming that you're overcounting parts. Don't overlook the possibility that you may be undercounting flights, particularly since you mentioned speed. If the problem is worse at speeds above 80%, the flight may not always be in front of the sensor for longer than a full PLC scan cycle.

Hey ummm, you guys help a pump guy out with the conveyor term "flight"?

Sorry

Stationmaster

Steve (Stationmaster),

Conveyor flights would be equivalent to the steps on an escalator -sections that are jointed or hinged on each side, with a protruding edge (if needed) to move the product.

Clint,

Steve Bailey hit on a valid point. Have you run it unloaded to see if your Flight Counter is accurate?

It seems you are assuming that your staff timer always times out on exact conveyor flights. What if the timer starts in between two flights and ends between two others? You could easily introduce an error of 1 or 2 here.

Why base your calculation on Time? You could do it the opposite and base it on Counts. Use the Staff Counter Done bit as the trigger to do the Moves, then Reset the Staff Counter and the Bird Counter. Count up to a certain number (say 100), and save to an F8 location. Now at same time, Start the Bird Counter, and when Staff Counter reaches Done, store Number of Birds in an F8 Floating Point memory location. Now Divide Number of Birds by Staff (100) to get Birds/Staff, then Multiply by 100 to get the % Efficiency.

Amazingly, if you set your Staff Counter at 100, you don't have to Divide by 100 and then Multiply by 100. The Number of Birds will directly be the same as the Efficency in Per Cent, so there will be no place for math rounding errors!

The advantage is that you are starting your count at the same relative flight location each time, so that flights are not missed due to the timing. You will not need any timers at all, unless you want to record the time for some other purpose.

OP said that he's not off by a significant amount during his 480 samples, not that he's off by 480.

milldrone said:

Here is an interesting oddity you are getting 480 extra counts in an eight hour day and you are comparing every minute. How many minutes is that? Hmmmm!

I'm having trouble with what is meant by this comment?

Click to expand...

The good news is, that I am sampling every min. with excel and doing the final averages there, so on an 8 hour day at 480 samples the extra counts are'nt making a huge diffrence in the end result, except that it looks odd to make an extra part thats not there.

Click to expand...

Thanks for filling me in Lancie. I had googled it and had an idea...... The only conveyors I work with (tomato packing house) are just continuous fabric belts with nothing to count. I can follow the thread a lot better now.

I agree with the observation: If the conveyor runs at a constant speed, why must TIME be a factor in the efficiency calculation?

Steve

These are pictures of a flight conveyor.
http://www.patchn.com/mroll001.jpg

http://www.patchn.com/mroll003.jpg

He did state that when the line ran 80% it worked fairly well; therefore speed may not be constant. Since speed is a product of time and distance it can be just as easy to use time with count.

May be easier to use the time clock and time stamp the count every hour, just depends. Using the clock can offer actual run time, count, etc. over an hour (I assume) shift.

Since a flight conveyor is involved you just need to know the gap is filled or not...basically.

There are numerous ways to do this.

If you can change speeds on the conveyor, then you need to change the preset of the de-bounce (part) timer. As the conveyor runs faster, then the part counter should run slower. If you set the timer for the fastest speed of the conveyor, then the short time may allow unwanted counts at slower speeds. If the trouble gets worse at higher speeds, this could be an indication of more trash on the conveyor due to inability to clear unwanted parts, (fat, skin and other stuff.)

Another consideration is that over time (months/years) things change, and so will timer presets. I don't think that fog would be trouble for you, but it can affect photocells. Another trouble for retroreflective photocells is target color. Again, I don't see this as trouble in your case.

I understand what you are up against, I pulled my hair out trying to count sheets of veneer on a conveyor that did not have flights and overlapping was common.

rsdoran said:

Not sure yet why you created LAD6, may see it after looking closer. I personally would use the inputs directly. If I was doing it this way I would want the inputs scanned at the beginning of the program, not at the end. I am not sure if it matters or not in this case, maybe someone will clarify that.

Click to expand...

I noticed that too. Move the JSR 6 in Ladder 2 rung 6, to Ladder 2 rung 0 and shift all the others down one. (Note, you can drag and drop rungs by click-draging on the rung number to reorder them instead of re-entering the rung.)