Math question?

I am not sure if this would be a math related problem or not.

I have a flow wrapping machine that runs x amount per minute.
The speeds can go up and down depending on factors such as rejects upstream startup or what not.
At times there are flags on this machine that say a certain product is bad for one reason or another.

There is a reject belt attached to the machine at the exit but it is based on the speed of the machine/geared with it.

The problem I have is in kicking out the reject at the exact right time.

Sometimes it is too early to open and close others it is too late.

I have no idea where to start other than setting timers based on run speed at the time of reject. But this seems too easy and I would think there is a function or something I can base this off of.

Has anyone encountered this and what did youdo?

Thanks in advance for the help

If setting the timers preset based on machine speed will work, then why not do that, even if it seems too easy? Without knowing more about the machine, it is hard to go farther.

What is a "flow wrapping machine" anyway? Is this a machine that wraps plastic sheet around a product, and runs at a speed determined by flowrate of the product? If so, it seems you would need an internal trigger from the machine (or from the program that controls the wrapping machine) to know the exact time that it is ready to reject.

Without an internal trigger, then you are left with trying to calculate your own based on how the machine works. Only you know how it works at this point.

There is a virtual axis I can use based on 360 of course. And I can count the number of product it puts out based on this, And I do have external triggers to tell the machine bad product. But my problems still comes in that the reject belt is cammed? into the machine so the speed changes of the reject belt also.



The way I get around it right now is to kick out say 2 before and after the product. which the powers that be would rather just kick out the one bad product. Which I think is harder to do and nearly impossible to get right every time.

any other thought or info needed, Ill try to provide it.

But my problems still comes in that the reject belt is cammed? into the machine so the speed changes of the reject belt also.

Click to expand...

The key trick is to use the speed change of the reject belt to change the time when you kick out. You need to know the speed change as soon as you can get it, not after it has happened.

...kick out say 2 before and after the product

Click to expand...

Well, What is the product? When does the product happen? How do you know when the product arrives? In short, could you make a one-line schematic of the machine and post it here as a JPG picture, or PDF file?

Another method is to add a detector of some type (photoswitch, proximity switch) at a point upstream so that you have enough time to activate the kicker before the reject product arrives.

If you know the speed of the belt and when the part is placed on the belt then all you need is the distance to the reject. When the part is placed on the belt 0 the position. Increase the position at the rate the belt is moving. When the position is >= reject position then turn on output.

As far as the math goes, this is straightforward:

The speed of the belt = max speed x actual rpm / max rpm

Measure the distance between the sensor that detects the reject and the kickout point

Since speed = distance / time

time = distance / speed

If the velocity is consistent between the detection and arrival at the kickout point you could use a timer with the preset a variable equal to the time calculated above.

Or if you cannot accurately measure/calculate speed; fit a sensor that counts gear teeth then use a counter to activate the reject mechanism.

Plcnewbie
All of the above are correct but my experience with similar systems is that reject problems are almost always caused by the reject mechanism rather than a simple timing issue. How does the system actually remove a rejected pack from the conveyor?
Most of the systems I have worked on use either a blow off air jet or a pneumatic cylinder to reject the pack. If this is the case problems are almost always down to changes in air pressure.
Andybr

plcnewbie said:

There is a virtual axis I can use based on 360 of course. And I can count the number of product it puts out based on this, And I do have external triggers to tell the machine bad product. But my problems still comes in that the reject belt is cammed? into the machine so the speed changes of the reject belt also.

Click to expand...

If I am understanding this correctly, you have an encoder on the conveyor and the conveyors are tied together?

In this case your problem is very easy. You would count 360 degree cycles then reject based upon position rather than time.

IO_Rack said:

If I am understanding this correctly, you have an encoder on the conveyor and the conveyors are tied together?

In this case your problem is very easy. You would count 360 degree cycles then reject based upon postion rather than time.

Click to expand...

His problem is the rejection delay. His problem will not be solved with cam-triggering. With his variable speeds he needs to implement either variable timers or another trigger for his rejection.
He needs to have a constant time between triggering the rejection and the actual rejection. At variable speeds, this means the point of this triggering is also variable.

Jeebs is correct. You forst need to determine the amount of time it takes your reject assembly to move between the reject and non-reject position. And don't necessarily assume the move times are the same in both directions, though they may be.

A distance based reject is the best bet. What you need is an offset distance based on the actuation delay. This distance is the delay time (in seconds) times the produst speed (in distance units per second).

Keith

As I understand it the reject belt is controlled by the downstream machine and you are programming the Flow wrapper? What is your control system?
Having programmed many flow wrappers, in either case you will never be accurate using timers, unless you want to also reject good product.
The machine that has the position of the reject belt needs to manage the reject. You would use the position of the reject belt to track the product to the reject. Based on the operation of the machine there are a number of ways to do that.

This looks very similar to a system that I have running.
Detect a bad product 30 products before the reject. Must get only that product at 10 products per second without disturbing the products either side. Oh can also stop and restart the machine within 10 products.

Method I use is (Logix5000)


Tracking array with a UDT for each product I also have a simulated axis - 360 Deg per product
Combining the tracking array and 360 deg position. The reject output is turned on when tracking[5].reject is true AND axis deg between some values (eg 90 and 210 deg)


This is Position dependant system (ie it does not care about time or speed). It relies on the product Not slipping)
Your Reject CAM should also be position dependant on the simulated 360 Deg position so can be turned on as needed.

More thought required when the output is required to be on as the tracking array advances
eg (Track[6] and Deg > 300) OR (Track[5] AND deg < 20)

Keep us in the loop if you go this way as I remember some coding traps to do with position rollover (near 360 deg) and the tracking array