Reprducing Sheet Metal Contour Bends from manufacturer's drawing

[wsimonton]

I am working on documenting a 1919 Climax Geared Steam Locomotive CN 1551 in Alibre.

I have obtained a copy of the original sheet metal drawings of the Cab from the company which manufactured the sheet metal parts for 1551.

The drawings show the length and radius of the bends which I have reproduced in the 2D drawing attached

The arcs in the drawings are not tangent which Alibre requires in a Contour Flange. When I use Alibre to reproduce the bends the sheet metal part opens up (under bends) after applying the tangent constraints required by Alibre. Is there a way to EASILY calculate (recalculate) the radius of the bends so that the arcs are tangent and Alibre can correctly reproduce the part. Note that the final bend follows the curve of the roof with the same Arc Radius of 150.47"

The 2D DWG has all the defined measurements required from the drawings but note again that the Arcs are not tangent. I have also attached a pdf for those of you who are reluctant to open live files.

 

[NateLiquidGravity]

Try adding a small almost imperceptible radius fillet at those points where it is not tangent.

 

[wsimonton]

Tried adding a small fillet per your suggestion. Some weird results. Small circles and partial circles offset from the point where the arcs meet and one arc shortened and one extended to meet different points on the partial circle.

 

[wsimonton]

I was was able to determine a "solution" which involved "fixing" the furthest point in the curve in each segment to correspond with its location determined from the manufacturer's drawing. Using the Tangent Tool to make each arc in turn tangent. Most of the arcs were affected only slightly. However, the third arc (6.35" with a 45.19 Radius) was affected and lengthened. See Sketch<2> and compare with the 2D. The end point of the panel ends up at the right place

 

[JST]

Should they not automatically be tangent if the center of each is on the radius line of the next one which is larger (as it appears those are)?

That seems as if it would be required by geometric definition of the tangent as being perpendicular to the radius........

 

[NateLiquidGravity]

Should they not automatically be tangent if the center of each is on the radius line of the next one which is larger (as it appears those are)?

That seems as if it would be required by geometric definition of the tangent as being perpendicular to the radius........

True but I don't think the centers all exist on the radius line.
There are 2 spots like this (one on each side) where I found they were not tangent.
However the 3 arcs in the center were made with the same radius 150.47" and I think that also screws it up.
Additionally there is a note on the side saying 45.79R but the measurement at in the drawing for that location says 23.21" Again perhaps unintentional and possibly the root cause problem.

 

[JST]

With a small error of tangent, I would happily make it tangent by moving the center over. The idea is a smooth curve with no "cusps", and I would just make it so. Almost surely the shop did that, they did not pay much attention to small things like that so long as it fit, and it would.

 

[Oldbelt]

Bill.
there is no simple way to make the ruf in sheetmetal.
One thing making it difficult is you can’t scale a SM part.

I will come back later with a solution

 

[sz0k30]

I am working on documenting a 1919 Climax Geared Steam Locomotive CN 1551 in Alibre.

I have obtained a copy of the original sheet metal drawings of the Cab from the company which manufactured the sheet metal parts for 1551.

One thing I might keep in mind is that there are very likely small differences in dimensions manually calculated prior to 1919 and those internally calculated by a 2019 3D CAD system. How many decimal places did they use, rounding errors, angles etc. I would say definitely enough to untangent some arcs.

 

[JST]

In 1919 and a metal fab shop, if it did not show up with a tape ruled in 16th of an inch (about 1.5mm) they did not worry about it at all.

Move the radius circles around to make it work.... that's about what was done back then.

 

[Oldbelt]

 

[Lew_Merrick]

Oldbelt -- What you posted appears to be a Die-Formed Part and not something made by what passes as "conventional Slip-Form" manufacturing today -- though it could be a "spin formed" Part. ???

 

[Oldbelt]

Lew it was think as rolled up, but the original ruf is in 3 sections ,2 side and a center part riveted together.
My former post was only to demo how sheet metal sketches shall be made to get a result.
How they produced in 1912 I have now idea about. You are right to day it wouldl be a die formed part.

 

[Lew_Merrick]

Oldbelt -- The design was almost certainly created based on "neutral axis pathways." Modern CAD systems assume that you only have one constant radius being formed in any Part. My "guess" is that the Part you are looking at is forned using three (potentially variable) radii values.

 

[JST]

Definitely overthinking being done on this.

I would bet on the shop folks back then forming it to a template that was probably cut out of wood. Bend, hold up template. bend a bit more, check and OK, now it "fits good enough".

They were building locomotives.... heavy, rough machinery that did not have reliably interchangeable parts. Very much like steel construction for buildings. Accuracy of 1/16 inch (about 1.5mm) would be considered very good.

Pretty much the same deal as with shipbuilding, only with shipbuilding, they often formed the template off the actual ship under assembly. With a locomotive, they would at least have a drawing for pretty much every part except maybe the general piping.

So, the drawing might indicate the shape to cut the template, and then the shop workers judged when the part was bent close enough to assemble.

Not something that was done precisely, it was just done "to fit". So "cheating" the centers to get it to work right is perfectly OK. I'd bet money that if you measured 5 of them, you'd have 5 different sets of dimensions. Not hugely different, but different enough to wash out all the errors that Alibre complains about.

 

[GIOV]

I suggest making a wood guide and bending the plate properly. In ADE you should only apply Contour Base Flange Feature.
I hope this help.

 

[Oldbelt]

Lev yes the ruf line consist of 4 radi.

JST you are right about how the worked in ancient time,
but I tried to help WSIMONTON who has as goal to document the
loko. Climax 1551.
To make a solid model in any CAD system you must have the exact dimensions
to make proper assemblies.
So overthinking or not is not aktuel in this case.

GIOV your respons shows an alternative method to copy the ruflines inner side.

 

[Lew_Merrick]

Lev yes the ruf line consist of 4 radi.

You do realize (I hope) that your are only allowed ONE IML Bend Radius on a Sheetmetal Part don't you? [That being an artifact of the Lockheed K-Factor usage.]

 

[JST]

I said the same thing a while back and was immediately corrected on that issue, with others mentioning that it can be changed per each bend, which makes much more sense. And when I looked at it, it appeared that they were right. I have not actually done a piece that way as of yet.

Oldbelt: The overthinking means trying to duplicate everything exactly from the drawing, instead of doing what it takes, such as slight repositioning of a radius center, to make it come out the way it is wanted.

Drawings from the past are often not perfect, and the shop would make what was actually needed, even if it was not exact as the drawing presented it. When making a new drawing, those things may need to be adjusted slightly to get what the shop actually made back then.

 

[Lew_Merrick]

I said the same thing a while back and was immediately corrected on that issue, with others mentioning that it can be changed per each bend, which makes much more sense.

Look at the definition of the Lockheed K-Factor.