Continue to Site

Welcome to MCAD Central

Join our MCAD Central community forums, the largest resource for MCAD (Mechanical Computer-Aided Design) professionals, including files, forums, jobs, articles, calendar, and more.

Assemble by curve

geertdaenen

New member
Hello,


I'mtrying to assemble a cable chain in ProE. I'd like to do this by use of a pattern on a curve (because we want to change the curve for a few positions). But as you can see in the next picture, there is a small problem. The links of the cable chain will not stay on the curve during the bend of the curve.
Is there anybody who can help me with this problem?


View attachment 3867


If you want to, you can download my cable chain here: 2007-07-12_054618_cable_chain.zip


(The radius of the curve has to be 55mm and the distance bestween the links has to be 46mm)


Thanks in avance
 
As said by geertdaenen, go thru the tutorial. You must first define TWO points on the curve with distance = the link hinge points.


Pattern these points.


Assemble the link to the points.


Pattern the link using reference pattern.


The above is a broad idea. The tutorial gives a detailed step by step procedure.
 
TIP:


When you make the sketch, constraint on the perimeter. Your perimeter should be a product of the number of links* pitch *2.


I am making a movie file and will post the link when done.
 
Thanks for the help everybody. But there is still a problem:


the straight distance between the links during the bend, will never be the same as the distance between the links on the straight curve. (because the dimension you have to give by making a pattern, is always the dimension measured ON the curve).
Is there a solution for this problem.


In the movie on youtube, on is changing the dimension of the link. But this is not the solution for my problem, because this dimension has to be 46mm.
And changing the radius of my curve is neither a solution, because the radius has to be 55mm.
 
This placement using two datum points will fail, if the ratio between radius and part length cannot be neglected.
You could constraint your part to
- midplane
- axis through point
- mate/align to datum plane through axis and through second datum point.
ReinhardN

This placement will not show reality, because the joints will not meet, but for a chain it will do. If you use references of your top assembly for your curve (perimeter dimension !) your chain will move with your assembly.


Edited by: ReinhardN
 
geertdaenen said:
In the movie on youtube, on is changing the dimension of the link. But this is not the solution for my problem, because this dimension has to be 46mm.
And changing the radius of my curve is neither a solution, because the radius has to be 55mm.
You need not change the length of the link if your ratio is correct. The movie was made in haste, I used an earlier made link to assemble only to drive home the point. The fact remains.... The perimeter should be a multiple of the link length * 2.


Secondly I have taken 25 links of one type. Hence the dimension pattern increment is 1/25 = 0.04, with one point placed at ratio 0 and the other at 0.02.


In your case the perimeter P = PI*55+ 2*center distance between sprockets (C).


since the length of the link = 55; P=n*55 where "n" is a whole number.


Trust this helps.
 
Dear Srinivasaniyer1,


my link length is 46mm and my radius has to be 55mm. Will it be a problem?
And my cable chain is just a small part of a ellipse (so there will not be a perimeter).


Can you maybe take a look at the file i've put in my first post?
 
Srinivasaniyer1


In your model the points are patterned at a distance of 46 providingthe correctassembly reference for the first set of links. But when you measure the distance between points for the connecting links (not yet assembled) the distance is44.6710 or 44.7698 depending on which pair you choose. When I zoom in on yourpatterned points the second point on the link does not align through the curve. Have you found an easy way around this?


Bob
 
Dear Bob,


I am aware of the problem. It is because the links form a chord, whereas the pattern divides the curve. Hence the error. This error will reduce when the arc radius is large, nevertheless an error shall exist in the absolute terms. I believe the software uses the pattern to orient the part rather than position it exactly. I am yet to find a general solution to the problem. My method is as recommended by PTC's suggested techniques which I referred to some years ago while on WF1.


Dear Geertdaenen,


You must select the complete curve. I suggest you select the curve feature from the model tree and then invoke the datum point command. Secondly APNT0 should be as a RATIO and not CONSTRAINED to the END of curve.


For the perimeter dimension, in the sketch mode, window select all, Edit>convert>perimeter... and then select the dimension that you would like to vary. e.g In a rectangle, when you CONSTRAINT a perimeter and LENGTH, theWIDTH will be a calculated value and hence vary. I have given a perimeter dimension because you are going to wrap the links and should be a MULTIPLE of the link length which eases calculation. Otherwise not necessary.
 
Ok,


as you said, this was my problem: I only selected the first line of the curve, and not the whole curve. To put the second point on the whole curve, I've searched a while for the solution, but when you click right on the first point (in the "datum point tool") you can choose "duplicate".
There is still only the problem with the second datum point (as described byBob), but this will not be a big problem. Anyway, if you know how to solve this problem in the future, please let me know.


And now I also know how to make a perimeter! But this was not the solution for my problem.


Thanks a lot
 
Hi Guys,


I want to create a pattern of cones around a curve as described above (the cone has two Dat. points to simulate the two points for attachmentas the chain link, but the cones are not tangential to the curve in the round parts. How can I accomplish that?


Is there any other method


Thanks in beforehand


View attachment 3892
Edited by: ogama
 
Thanks srini......


But there was only one that could be relevant ("Cylindrical gear with helical teeth") and at the same time, not that relevant. because itpatterns around a circle (too easy) and not around an elipse.


In general there are many tutorials explaining the pattern around a curve, but the features to pattern are normally round and not features that needs to be tangent to the curve as mine.


If you have any other good idea, I'll be glad to hear it.
 

Sponsor

Articles From 3DCAD World

Back
Top