A hard part of a drawing

[lrm]

Another “old schooler” here.  Let me take a different approach to this discussion and assume the roll of the designer of the part and it’s the 1960’s (pre-CAD). I set the WABAC (Wayback Machine) to 1965 and I’m a summer intern at American Can where I work on the design of parts for packaging systems. I’m working on a drafting board, pencil and erasing shield in hand, along with a brush to wipe away the residue from my multiple erasures. I’m at the stage of designing the handle. The 12.5 cylindrical surface has been defined and I have chosen the 40 and 16.5 dimensions for the slope of the handle. I’ve additionally decided that the handle will end with a R4 spherical surface. My next task is to create an appropriate bulge and neck for the handle. I’ve decided that 12mm is a good maximum width and the neck should be 4mm.  These two dimensions will be easy to measure at inspection. I draw parallel lines to the centerline of the handle offset by 2 and 6. To be consistent with the rest of the part I decide to give a R10 radius to blend the handle in with the main part. Now, what radius should I use to blend the R10 and R4 radii and still provide a 6mm offset from the centerline?

 

My initial guess is that a R20 arc will do.   I create a circle of radius 30 (20 + 10) centered at the R10 circle which is tangent to the R12.5 circle and the offset 2 line. I draw a circle of radius 16 (20-4) concentric to the R4 circle.  The results are show below (drawn with CAD but imagine this drawing done manually).

I note that the R20 arc (yellow) goes outside the 6 offset (green) and that I should use a larger radius to flatten the curve a bit.  I get out my high-tech electric eraser to “undo” the construction and try a radius of R22 and get the following.

Gee, that looks pretty close to meeting the 6 offset line and decide to go with it.  Its deviation is less than the implied precision of +/- 0.5.    The R22 radius is more difficult to measure and it’s not a critical dimension for the operation of the part.  I find it a bit inconsistent that the neck dimension included a DIA note but the 12 does not.  Should this be interpreted to mean that the 12 represents a flat section to the profile? If so, then I think two R22 dimensions should be included.  Not clear.

 

Yes, I’ve made a few assumptions in this story (or “narrative” as they like to say today) that some may take exception to but I think they are representative of the design process.  The bottom line is that to make a continuous profile of the handle an additional segment is needed or one of the dimensions needs to be adjusted slightly.

Merry Christmas to all!

[eldon]

Again, I appreciate the delving into history, but this is a modern problem, and as such I would expect the solution to be contiguous lines satisfying all of the information provided.

 

I think that the position of the 12 dimension is crucial. It is not wrong where it is, but it does not add to the solution.

 

My solution incorporates all of the information. Your solution ignores some information, because you know better.

 

I wonder if the tutor who set the original problem will honour us by providing the solution. One can but hope!!

[lrm]

@eldon I think the 12 is a requirement and should be met (if possible). We both agree that something has to be added to make the precise shape.  You added a straight segment by breaking the R10 arc in two.  I added it between the R10 and R22 arcs. 

[eldon]

My latest thought was not to split any arc, and not to use the 12 dimension to draw the outline. It is where it is and valid, but is a distraction in drawing the contiguous geometry.

 

Sorry, I have changed my first thoughts since the beginning due to reading contributions on this and other forums.

[paulmcz]

I believe that to have 100% tangency and to satisfy all the dimensions, there has to be very short cylinder between two radii of 10 units.

(drawing attached)

Hard part.dwg

[vuxvix]

17 hours ago, paulmcz said:

Tôi tin rằng để có 100% tiếp tuyến và thỏa mãn tất cả các kích thước, phải có một hình trụ rất ngắn giữa hai bán kính 10 đơn vị.

(bản vẽ đính kèm)

Phần cứng.dwg 60,1 kB · 1 lần tải xuống

I believe you are like me when you first looked at the problem and placed the apex of the arc R10 2cm away; arc 22 is 6cm to the centerline. I don't think that's correct.