Timing,timing,timing

jaimeng6

Member
Join Date
Aug 2007
Location
alabama
Posts
52
Hello,fellas!
I'm looking for alittle help on this project. I have a 60ft. 1"x1" steel rod moving towards a rotating shear to be cut into 9 inch pieces every 3/4 of a second, the rod is driven by a freq. Drive untill the shear takes the first bite after which the shear grabs and feeds itself. My problem is i can get the bar just past to the shear point and stop the drive as it waits for the shears to come around.the boss doesnt like the scrap waste and would like me to synchronize
the flywheel on the shear with the drive so they meet at the same time.would i be correct by just plugging in frequencies on the drive according to the degree of the revolution on the flywheel compared to the distance of travel needed? I could just say deal with it. Would appreciate any input,thanks.
 
Synchronizing genarally involves the ability to speed up or slow down one of the axes to match a position on the other (master) axis.

I would think you would need a servo to do this reliabily.
 
Agreed. That is a typical flying shear system, which requires precise coordination between the infeed and the shear, including positional tracking of the leading edge of the stock, and taking into account the inertia of the stock and feed elements.

My guess is that just fiddling with the speed reference to a frequency drive won't accomplish much, or be repeatable.

You would probably have better control, if the drive can operate in vector mode WITH and encoder, and you close the loop completely with a servo controller at least.
 
Yes, i see where your going with this. I have a drive controlling the stock, the shear connot be adjusted. I appreciate the input ,i totally forgot about using a servo.the shear is belt driven with 40hp induction mtr. The servo would control the stock reading signals from the shears axis position. I will give this a shot and your right i need better control. Thank you!
 
This is not an easy project, at least not the first time.

However, I have applications like these worked out. What is needed is an reference encoder that measures the speed of the bar. From the data the bar must be moving about 12 inches per second but what about during the time the bar is accelerating or decelerating. If you simply use time then you will have scrap.

The cure is to gear the shear to the reference encoder. This way the shear will cut 9 inch pieces no matter how fast the bar moves.

There is another trick to cutting. The shear MUST match the speed of the bar while in the cut. The motion profile is not simple. If the rotary motion causes the bar to speed up or slow down there will be a feed back effect through the reference encoder.

The motion profile will need to accelerate over the top and then when coming down it the angular velocity will need to be just a little bit faster than the bar. This is because the shear will hit about 30-40 degrees before the bottom. The horizontal component of the angular velocity of the shear must match the speed of the bar so the shear doesn't pull or stall the bar. After the shear is out of the cut the shear must race around the top. I have the math figured out. One can even change the cut length on-the-fly.

This is not a PLC problem. A PLC by itself cannot do this.
This is not a simple motor control problem. The math and technique is most important.

So what is the circumference of the shear? Is it really a shear where the knife is always pointing down or is it a knife or clipper? Veneer mill clippers have blades sticking out 180 degrees.

A freq drive is ok for the bar feed but the rotary shear should be a servo. In the past we have had to not only use a full PID with feed forwards but also an extra term that generates torque as a function of the angle of the shear to offset gravity. This is a sine function.

Is the shear counter balanced?

The shear will be massive to cut 1 inch bar.

I am home now. When I get to work I will see if I can find some motion profile plots
 
My initial estimate

http://www.deltamotion.com/peter/Pictures/RotaryShearBar.png
The picture shows the motion profile required to cut a bar every 9 inches when the bar is moving 12 inches per second. What isn't known is the radius of the shear. I chose 5 inches. This means the shear must rotate 2*π*r or 31.4 inches for every 9 inches of travel of the bar. This results in very high speeds, almost a 175 ips angular velocity, for the shear as it whips across the top to make another cut.

An explanation of the plot.
The lines are color coded and the legend in the lower left. This graph is generated using a previously written program. I changes the cut length and lines speed for this application.

A new cycle beings at .49 seconds. At this point a cut has been made and the shear is at the bottom. The bar position (yellow) and the shear position ( red) are reset to 0. At this point the shear angular velocity ( blue ) matches the bar velocity ( black ). The bar position keeps moving at a constant velocity of 12 inches per second but the shear must start speeding up so that the horizontal component of its angular velocity matches the bar velocity. This must happen until the shear blade rises above the 1 inch stock. I computed the shear most rotate 38 degrees before the blade rises out of the bar. At this time the rotary shear executes one 5th order polynomial until the shear is about 38 degrees from the bottom. At this point the shear must again match the horizontal component of the angular velocity to the bar speed while the shear is in contact with the bar.

The thing to note is the peak angular velocity at the cursor. It is at almost 175 inches per second or 175/5 or 35 radians per second. I should have plotted the accelerations because then one can get an idea of the torque required once the mass and inertia is known. I think the real problem will be the accelerations.

This motion profile is generated using 3 5th order polynomials. One for entering the metal, one for leaving the metal and one for moving across the top.

I plotted the accelerations and the peak accel is about 395 radians/sec^2.
That is asking for a lot if the shear mass is high.

One can see that this is not a simple application. There is some math and physics involved. There are no points where the shear is moving at a constant angular velocity.
 
peter
i see what you say it not being a simple profile.please allow me to explain in more detail the type of shear.it consists of two disks about one foot in diameter with 4 (protruding about one inch) shear pieces every 90 degrees,with one disk vertical directly above the other.the bar gets fed until it reachs the pinch point then the cutters while rotating one quarter turn every 0.71 sec. Cuts off a nine inch length.i have an encoder to give me the position of the bars leading edge, so i could stop it about half an inch past the pinch point, but they would like no scrap.so i placed a proximity on the flywheel that is a constant speed.i know all the parameter yet not experienced yet to know where to plug them in a formula.i am excited about this path i havent traveled,time wont allow me to walk it alone so i called an automation specialist for next week. I thank you all and will be listening.what about a resolver?
 
peter
i see what you say it not being a simple profile.please allow me to explain in more detail the type of shear.it consists of two disks about one foot in diameter with 4 (protruding about one inch) shear pieces every 90 degrees,with one disk vertical directly above the other.the bar gets fed until it reachs the pinch point then the cutters while rotating one quarter turn every 0.71 sec.
OK, it would be good to know the effect radius of the shear disks. I assumed only one shear so you can see the it requires high speeds to go over the top.
With 4 shears the shear needs to travel one revolution as the bar travels 36 inches. Hopefully the effective radius of the shear motion is about 5.75 to 6 inches so the shear circumference is about 36 inches too. THIS IS IMPORTANT because it will reduce the variance in the shears angular speed.

The key thing you must realize is that the shear will need to move a little faster than the bar feed velocity because the horizontal component of the shear angular velocity must match the bar feed velocity. Now there is little distance to correct between the distance the shear must move and the bar must move to be synchronized at the next 9 inch interval.

I have an encoder to give me the position of the bars leading edge, so i could stop it about half an inch past the pinch point, but they would like no scrap.
I have customers going at 150-180 fpm with accuracies of 0.010 inches. There is no stopping.

so i placed a proximity on the flywheel that is a constant speed.
I don't see how you can cut at a constant speed. The shear will stall or push the bar or the other way around. An absolute encoder is a must. I recommend 17 bit SSI absolute encoders.

i know all the parameter yet not experienced yet to know where to plug them in a formula.
Like I said. I have the program written. It will take a person some time to work this out from scratch.

i am excited about this path i havent traveled,time wont allow me to walk it alone so i called an automation specialist for next week.
The automation specialist better be good at math.

I thank you all and will be listening.what about a resolver?
I prefer a 17-18 bit SSI absolute encoder
 
Peter, I think jaimeng6 is looking to go the other way with the control concept. He leaves me wit hthe impression that the shear has a huge amount of inertial and he wants to run that at a constant speed. It is the bar position he want to control. It doesn't sound like he has a real tight grip on the bar since, in his first post, he states he expects the bar to travel with the shear once the shear contacts the bar.

So unless he changes his mechanical system to get better control of the bar he needs to determine the leading edge of the bar after a cut is made and synchronise that position to the shear cut.

Keith
 
jaimeng6,

Could you please post a picture of this device. I think there are still so many variables we don't understand that it is not easy to help you.

I am envisioning something very low tech where, perhaps the bar gets pulled through the shearing section once the leading edge gets into the rolls. Yes there would be rubbing but I have seen systems like this where tolerances are wide open, end cuts don't have to be straight and other marks on the product from sliding tooling are tolerable. If that is the case, then it truly is a matter of putting the leading edge of the bar in the right place at the right time (and making sure the bars are the correct length).

BUT

We still don't really know what this thing looks like or how it really operates.

Words are not the best way to describe systems like this, at least not by themsleves. Please post a picture or drawing of this section of your line so we have a better idea of what is going on.

Thanks,
 
So what is being controlled.

Peter, I think jaimeng6 is looking to go the other way with the control concept. He leaves me wit hthe impression that the shear has a huge amount of inertial and he wants to run that at a constant speed.
That possibility exist. The mathematical relationship will still exist. Now the bar must change speed. The bar and the shear can't run at a constant speed. It is physically impossible while the shear is in the metal and there still must be an adjustment in velocity so the next cut is at 9 inches.

It is the bar position he want to control. It doesn't sound like he has a real tight grip on the bar since, in his first post, he states he expects the bar to travel with the shear once the shear contacts the bar.
jaimeng6 must be able to vary an control either the shear or bar velocity.

So unless he changes his mechanical system to get better control of the bar he needs to determine the leading edge of the bar after a cut is made and synchronise that position to the shear cut. Keith
Yes, I think think jaimeng6 needs to have an encoder on the feed line and the bar cannot be permitted to slide on the feed line. Having four shears will make a big difference if the radius is in the range of 5.75 to 6 inches. This would better match the shear speeds to the bar speeds and the extreme accelerations can be avoided. Now one may be able to control the shear and vary its speed only a small amount.

Some what OT. Have you ever thought of how the defects are cut out of french fries?
 
Norm said it best. We really don't know what he has to deal with so it is hard to determine what he needs to do. However, reading between the lines:

Originally posted by Peter Nachtwey:

jaimeng6 must be able to vary an control either the shear or bar velocity.

Technically I think he has control over both. However, I think he also has a significant power mismatch. The bar will do what the shear is doing once the shear engages the bar. Worst case the bar is dragged past any bar driving element on the machine, meaning the control system loses track of the bar location. Better case is the bar drive system is overcome by the shear force so the bar advances but the control system still knows where it is at. Best case is the bar feed system is profiled to match what the shear does to the bar during the cut and corrects for any bar position error to get the cut right. The impression I am left with from previous posts is that the shear is designed as a constant speed shear. So it has a huige amount of inertial built into it to make sure the speed stays constant. Much like the mechanically profiled cutoff knives in the paper industry.


Originally posted by Peter Nachtwey:

Some what OT. Have you ever thought of how the defects are cut out of french fries?

No. I get enough ugly looking french fries that I didn't think they did that at all. Now that you bring it up I would have guessed they cut them as normal, inspect after cutting and divert the bad ones. But since you bring it up in this context I and guessing that isn't right.

Keith
 
Some what OT. Have you ever thought of how the defects are cut out of french fries?

I'm guessing it has something to do with a color sensor from a company based in Battle Ground, WA.

You mention that the defects are cut out of the fries where I would have thought the offensive julienned stick would just be pushed off to be made into mashed potatoes or filler for dog food or some other process.

This video: http://www.youtube.com/watch?v=dkZXwyr3Mvo is for frozen fries for consumer and mentions a camera and blow off at around 3:30 in the video...
 
I'm guessing it has something to do with a color sensor from a company based in Battle Ground, WA.
Yes, but there is much more than a color sensor involved. The motion control is pretty extreme.

You mention that the defects are cut out of the fries where I would have thought the offensive julienned stick would just be pushed off to be made into mashed potatoes or filler for dog food or some other process.
It is turned into animal feed. One doesn't use the long potatoes to make mash potatoes, hash browns or tatertots. The length control is important. The length of each strip logged as it enters and leaves and histograms are made so the length distribution can be monitored.

This video: http://www.youtube.com/watch?v=dkZXwyr3Mvo is for frozen fries for consumer and mentions a camera and blow off at around 3:30 in the video...
Yes, after the camera sorter sorts the good fries from the bad fries the bad fries go to a automatic defect removal machine which wasn't shown. The defect removal machine must cut out the defects where ever they are. They don't come equally spaced like in jaimeng6's problem. Changing the cut position on-the-fly is essential.

The whole point of this it to get jaimeng6's supervisors to realize that this is not a simple problem that can be solved with simple solutions. OK, and to show off a bit. It is hard for me to get excited about latches too.

It looks like they are expecting big speed increases. I think the speed increases can be hard but it will take some effort/money.

If all the pieces are going to be 9 inches long then I think it would be much simpler to have many saws or shears come down at once so 6 ft of the bar is cut into 8 pieces in one stroke. Then the bar is index down another 6ft against a metal stop that pops up and the bar is cut again. This is a brute force approach it can be done by almost any PLC and there is little math involved. I can see using the rotary shear and all the math if the cut length is going to change.
 
Hi Peter.
There was one of those machines just down the road. Was used to sort corn kernels. It used to blow any miscoloured kernels into a reject bin
Regards Alan Case
 

Similar Topics

Hello, I am struggling to figure out some logic. I have an analog pressure sensor coming into the PLC and would like to have multiple timing...
Replies
4
Views
129
the plant I am supporting uses a siemens graph function to do some of the step procedures. in several steps it waits for the step to execute for...
Replies
1
Views
755
All Using a MicroLogix 1400. Not a huge program doing an auto stop function based on a calculated roll diameter Using encoder feedback for...
Replies
12
Views
3,192
Hi all, I'm trying to fully understand the different effects of the Loop Update Time parameter within the Logix PID instruction. I was always...
Replies
18
Views
4,551
Logix 1756-L82E processor. I am using the .ST bit to time how long a MSG takes to complete. This works fine except for every 10th MSG cycle takes...
Replies
5
Views
1,611
Back
Top Bottom