TheSwamp

Code Red => AutoLISP (Vanilla / Visual) => Topic started by: T.Willey on March 24, 2009, 12:53:54 PM

Title: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 24, 2009, 12:53:54 PM
I know this isn't a lisp question, or even a programming question, but I couldn't figure out where to put it, but I thought that someone here might know it.  If this topic gets moved, it's all good.

How can I figure out the distortion between a rotary die and a flat surface?  I have searched, and can't quite seem to find the answer.  My co-worker has done the same, and came up with the same non-answers I have.  It seems like the die dimension are smaller than the flat dimensions along the paper, but we can't figure out a constant formula for it, as all the numbers don't seem to add up the same.

All help is appreciated.  Hope that made sense.  Thanks.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 24, 2009, 02:09:24 PM
I talked to some of the people here, and it seems like there is no equation.  Sucks.  But we are still searching incase we find some obscure web posting somewhere.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: Strucmad on March 24, 2009, 05:15:57 PM
Do you have a pic, or sketch of the distortion...just to get juices flowin :-)
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 24, 2009, 05:24:49 PM
Not one I can share, but let me see if I can come up with something that is similar.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: CAB on March 24, 2009, 05:31:38 PM
Doing a quick search, it looks like a complicated situation involving the material being punched, the size of the hole, and the speed of the punch. Maybe I'm not understanding the problem presented though. Happens all too often. :-o
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 24, 2009, 05:43:26 PM
Doing a quick search, it looks like a complicated situation involving the material being punched, the size of the hole, and the speed of the punch. Maybe I'm not understanding the problem presented though. Happens all too often. :-o

That is what I was thinking also, but they say there is no equation.  I think there should be; it might be complicated, but there should still be on.

Do you have a pic, or sketch of the distortion...just to get juices flowin :-)

Here is one I think that is similar to what we were trying to figure out.  With the layout given, what would be the equation to layout the die above?  The pdf is the same as the png.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: Strucmad on March 24, 2009, 05:58:18 PM
Bare with me , I have not worked in your industry and I may not be on the same track :ugly:,

Could you explain what I'm looking at ? does the circle need to be punched in that one rectangle or all.

looks like a cool math exercise  :-)

maybe a matrix solution (n,m)  :|
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 24, 2009, 06:02:25 PM
No problem.

Lets make it simple ( as simple as it can be ), and say the first row needs to be punched as shown, and then repeat at the distance shown.

I thought it might need to be something like that, but my math is so rusty its sad for me to admit that I love math.   :cry:
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: James Cannon on March 24, 2009, 06:04:43 PM
What 'fuzz factor' are you trying to account for?

Does the circumference of the die not equal the actual linear travel distance across the sheet/part ?
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: Strucmad on March 24, 2009, 06:13:36 PM
what would be the origin of the equation be..
why does the circumference come into it? is something rotating? I'm lost
tell me to bugger off if like  :-D
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 24, 2009, 06:15:01 PM
What 'fuzz factor' are you trying to account for?

Does the circumference of the die not equal the actual linear travel distance across the sheet/part ?

No.  We have a continuous roll that will travel through many rollers before it is cut down to fit within a pouch.  Each set of rollers does something different, and some seem to be in spec at some point, and the loosing spec spacing at the next; until a full cycle comes around, and it is in spec again.

I'm not sure if it's the speed issue that is causing the problem ( or need for differing dimensions between sheet and roller ), or if it is something else.  That is why I was trying to see if there was an equation that would take that into account.  Or even a principle that shows what is supposed to happen and why.  Then maybe I / we / someone could come up with something.  But I haven't even been able to find a true principle.  It could also be my search terms, as I keep coming up with distortion for sound waves.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 24, 2009, 06:21:22 PM
what would be the origin of the equation be..
why does the circumference come into it? is something rotating? I'm lost
tell me to bugger off if like  :-D

So in the pic shown.  The sheet will feed that way into the roller ( left to right on the png ).  The roller should then make the appropriate cuts to the sheet, so that they pattern shown will be true to dimensions shown.  The roller will rotate about it's center axis, and continue to cut the pattern at the spacing shown.  The origin is not as important as the repeatability of the pattern on the sheet, as we could cut off the beginning of the sheet that doesn't meet spec.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 24, 2009, 06:28:55 PM
Sorry I might have been clear in the drawing.  The roller is the 5 diameter circle, and the 20 unit long rectangle ( below it in the png ).
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: James Cannon on March 24, 2009, 06:45:00 PM
I understand the process you illustrate, but having no actual experience, I can't think of any reason there would be variance due to principle or an "ideal situation" and any problems that occur seem, to me, that they would be coming from natural deviance in machine operation and the general chaos of happenstance.

How good are your operators?  That makes all the difference in the world that no engineer or designer can plan for.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: CAB on March 24, 2009, 06:48:26 PM
My thought was that maybe your tolerance is too tight for the equipment.

What is the tolerance?
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: Strucmad on March 24, 2009, 07:09:45 PM
How can I figure out the distortion between a rotary die and a flat surface? 

Ahhh. I think I get it... yes the die will be distorted (only in one direction) to compensate for the for the arc of the roller, guessing it should be elliptical..

Just to make sure we are on the same page,  are we talking about the difference of A & B
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 24, 2009, 07:14:29 PM
I understand the process you illustrate, but having no actual experience, I can't think of any reason there would be variance due to principle or an "ideal situation" and any problems that occur seem, to me, that they would be coming from natural deviance in machine operation and the general chaos of happenstance.

How good are your operators?  That makes all the difference in the world that no engineer or designer can plan for.
Ummmmm.... operators.... ( don't want to get fired here )... Good!!

I just know what people that clam to know more tell me.  I can see why certain circumstances would affect the outcome of the pattern; mainly speed comes to mind.  If they are moving at the same speed, then I don't see why there should be distortion ( in a perfect environment, when no metals expand / contract when heated / cooled ), but those who are in the know, say there is.  I have no data to defunct their statement.  This is a big learning idea for me, so anything someone can find, stating one way or the other, I'm willing to read.

My thought was that maybe your tolerance is too tight for the equipment.

What is the tolerance?


Not sure, but I don't think they are too tight right now.  What I've heard is that, say for example the circle, will be out of spec by as much as half the circle at one point, and then at others, totally out of the rectangle pattern, until it comes all the way around, and will be back in spec for a little while.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 24, 2009, 07:15:33 PM
How can I figure out the distortion between a rotary die and a flat surface? 

Ahhh. I think I get it... yes the die will be distorted (only in one direction) to compensate for the for the arc of the roller, guessing it should be elliptical..

Just to make sure we are on the same page,  are we talking about the difference of A & B

Yes.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: Strucmad on March 24, 2009, 07:41:47 PM
O.k - Circle maths.

As per the pic. inprevious post.

The relationship between A & B..

You know 'A'

so using the cosine rule (A2 = 2.52+2.52 - 2x2.5x2.5 cos0) you can calculate 0 in terms of `A'

Then simply (B=0 x 2.5) will give you an equation of B in terms of A. - note 0 must be in Radians.

So you can plug in any `A' and get the `B' dimension
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 24, 2009, 07:44:01 PM
O.k - Circle maths.

As per the pic. inprevious post.

The relationship between A & B..

You know 'A'

so using the cosine rule (A2 = 2.52+2.52 - 2x2.5x2.5 cos0) you can calculate 0 in terms of `A'

Then simply (B=0 x 2.5) will give you an equation of B in terms of A. - note 0 must be in Radians.

So you can plug in any `A' and get the `B' dimension

Time to go, but I will try this out on some real life prints, and see what comes up.  Thanks for taking the time.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: gile on March 25, 2009, 06:26:37 AM
Hi,

I'm not sure having understood the request, but for calculating B (arc length) knowing A (chord length) and R (arc radius), you can do this (A have to be smaller or equal to 2*R):

(setq B (* R 2 (asin (/ A 2 R))))

With asin define as:
Code: [Select]
;;; ASIN Returns the arcsinus of a number in radians
(defun ASIN (num)
  (cond
    ((equal num 1 1e-9) (/ pi 2))
    ((equal num -1 1e-9) (/ pi -2))
    ((< -1 num 1)
     (atan num (sqrt (- 1 (expt num 2))))
    )
  )
)
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: ElpanovEvgeniy on March 25, 2009, 06:58:22 AM
 Here you can find all necessary mathematical functions, for work with arc, arc segment polyline, in autocad.  (http://elpanov.com/index.php?id=35)
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: Strucmad on March 25, 2009, 07:07:45 AM
I was having fun with the math..Well sad but true..

Tim,

here is the formula for your situation (5 diameter)

cos-1 is the inverse cosine, probably called arccosine in your world..

equation craps out at A=5 anything under is cool
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: Strucmad on March 25, 2009, 07:09:07 AM
You other guys had to go and ruin it with Lisp and Autocad... :lmao: :lmao: just muckin guys  :-)
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: gile on March 25, 2009, 09:05:30 AM
PS_Port,

converting degrees to radians isn't necessary. LISP deals with radians.


Tim,

To use PS_Port formula (which returns the same as the one I posted), you can use this acos sub:

Code: [Select]
(defun ACOS (num)
  (cond
    ((equal num 1 1e-9) 0.0)
    ((equal num -1 1e-9) pi)
    ((< -1 num 1)
     (atan (sqrt (- 1 (expt num 2))) num)
    )
  )
)

(setq B (* R (acos (/ (- (* 2 R R) (* A A)) 2 R R))))


Evgeniy,

Here're 2 more routines to complete your arc stuff.
They return the arc angle knowing tha arc length and respectivly the chord length and the sagitta length.
These routines use the Newton's method to solve equations as, respectivly:
chord / arc = sin(ang/2) / (ang/2)
2*sagitta / arc = (1 - cos(ang/2)) / ang/2)

Code: [Select]
;;; ArcChord->Angle (Newton's method)
;;; Return the arc angle knowing the arc length and chord

(defun ArcChord->Angle (arc chord / k x)
  (setq k (/ chord arc)
x (sqrt (- 6 (* 6 k)))
  )
  (repeat 6
    (setq x (- x (/ (- (sin x) (* k x)) (- (cos x) k))))
  )
  (* 2 x)
)

;;; ArcSagitta->Angle (Newton's method)
;;; Return the arc angle knowing the arc length and sagitta

(defun ArcSagitta->Angle (arc sagitta / k x)
  (setq k (/ (* 2 sagitta ) arc)
x (* 2 k)
  )
  (repeat 6
    (setq x (- x (/ (+ (cos x) (* k x) (- 1)) (- k (sin x)))))
  )
  (* 2 x)
)
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: Strucmad on March 25, 2009, 09:26:19 AM
PS_Port,
converting degrees to radians isn't necessary. LISP deals with radians.

Thanks gile, I understand that Lisp deals with Radians..

I am not very fluent in LISP  :oops:, and would not have attempted to answer this question if it was in LISP. I am always amazed at the programming talent here, and someone (like yourself) would easily answer this.

Tim's title "[Non-Lisp] -- equation] just gave me enough incentive to give it a go..  :-)

Hope we helped Tim  :-)
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: James Cannon on March 25, 2009, 09:32:49 AM
I love seeing manufacturing application topics like this come up.  Even though I'm out of the manufacturing environment, I still find the machinery and design concepts so interesting.  I love being able to observe and soak up what I can about it.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 25, 2009, 11:00:40 AM
Thanks for all the help guys.  I'm not sure if I will get to it today, but I will get to it this week, and then I will let you know what I find out.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: SEANT on March 26, 2009, 05:37:43 AM
It is likely just a problem with this example but: Would the 16.25+ pattern actually fit on the 5 dia. roller?
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 26, 2009, 11:02:49 AM
It is likely just a problem with this example but: Would the 16.25+ pattern actually fit on the 5 dia. roller?

Don't know, as I just wanted to show an example.  How would one find out?  I guess if we find the needed space for one pattern, then we could see if it would fit on said roll.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: James Cannon on March 26, 2009, 11:09:04 AM
well on a 5"%%c roller, you have ~15.7" circumference, meaning that your 'punch' on the roll will hit the 'part/sheet' every 15.7" o.c. linearly along the part/sheet.

So the spacing between sheets (if there's no automatic adjustment that starts the sheet at a certain point in the rolls spin) and the spacing of parts ON the sheet.

Unless I'm just way off or not thinking of something else that affects it or SEANT was talking about something else..
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: Strucmad on March 26, 2009, 05:43:04 PM
Good points..

thinking out loud (can you hear the gears)
so the pattern is on the die, and it is rotating, the sheet is flat.....slaps forehead 'doh'
man I was way off with my mental picture of this one, I had it all ass about

what josh said sounds fair..
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 26, 2009, 06:02:00 PM
What he says would make sense if it was sheets, but it is one continuous roll, so what doesn't matter where it hits really.  Only the first or last would need to be trimmed down, so that they are the correct sizes.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: James Cannon on March 26, 2009, 06:08:21 PM
What he says would make sense if it was sheets, but it is one continuous roll, so what doesn't matter where it hits really.  Only the first or last would need to be trimmed down, so that they are the correct sizes.

oh, ha... yea that pretty much solves that issue.  Forgot about that.  I come from a manufacturing background centered around smaller runs that didn't facilitate such bulk quantities.

I really have trouble wrapping my head around how your die could be off on one spin, then correct on the next.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 26, 2009, 06:09:52 PM
What he says would make sense if it was sheets, but it is one continuous roll, so what doesn't matter where it hits really.  Only the first or last would need to be trimmed down, so that they are the correct sizes.

oh, ha... yea that pretty much solves that issue.  Forgot about that.  I come from a manufacturing background centered around smaller runs that didn't facilitate such bulk quantities.

I really have trouble wrapping my head around how your die could be off on one spin, then correct on the next.

That there is the question.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: SEANT on March 26, 2009, 06:21:00 PM
I was picturing everything on one roller, as in example drawing with 5.5 dia. roller, but suppose that isn’t necessarily the case.  And even if there were multiple rollers, it would be odd that one could slip but then catch back up some revolutions later.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: SEANT on March 26, 2009, 06:22:56 PM
Unless – with multiple rollers – a loose timing belt and oscillating tensioner.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 26, 2009, 06:31:42 PM
One roller with the pattern repeating as much as possible on that roller, at least that is my understanding of it right now.  I'll check out the dwg Seant when I get a minute.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: CAB on March 26, 2009, 09:16:20 PM
Man I'd love to see a picture of this machine, or is that top secret? 8-)
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: T.Willey on March 27, 2009, 02:03:39 AM
Man I'd love to see a picture of this machine, or is that top secret? 8-)


That is a problem, as the machine is so old, and most of the drawings for the machine don't exist as we would like them to be.  Lets just say it is a machine that makes drug patches.  It is pretty big, and pretty complicated.
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: Spike Wilbury on March 27, 2009, 11:02:32 AM
Alan,

I am not familiar with this stuff, don't know but maybe if you do a search on google images for "rotary die cutting machines"
Title: Re: [ non-lisp ] Rotary die distortion equation
Post by: CAB on March 27, 2009, 11:24:38 AM
Thanks Luis. Looking through some of the links that I turned up I see many types &
many applications. Really cool stuff. When I was a plumbing contractor I toured the
Price Pfister faucet factory in California and the State Water Heater factory in Tennessee.
I found that very fascinating. Remarkable how things are made.