Guidelines/Best practices for 3D Modelling of Complex drawings 6

[sluzzer]

Greetings Mechanical Engineers!
I request experienced mechanical design engineers to suggest guidelines/best practices for doing 3D CAD modelling of complicated enginnering drawing.
I received a complicated casting drawing for 3d modelling, which consists of 12 A1 sheets with 100s of views and details! I am comfortable with doing drawing from the 3d CAD model, but the reverse when it is too complicated, is too difficult.
One of my colleague suggested to import, scale and arrange all sheets in AutoCAD and then to use construction lines to easily map between the views! It helps but I wonder is there any small tricks / tips / guidelines to do 3d modelling from a complex drawing?
Any suggestion is much appreciated!
Thank you!

 

[3DDave]

I usually contact whoever made the drawings and ask for a copy of the original model. Otherwise that sounds like 5-10 days of work.

 

[CWB1]

IMHO these are usually an exercise in organization more than anything. I usually start by modeling a few major datums relative to one another from the print, then start modeling major features and work down to lesser details. Labeling model datums the same as the print helps prevent confusion while you’re modeling, and if you need to add datums not on the print, use a second labeling system so it’s clear what datums are yours vs the prints. I also keep a paper copy for complex prints and literally check off features as they go into the model.

An important one IMHO is to keep facts and assumptions separate via layers or other means. Quite often you’ll get deep into a model and discover you’re missing dims bc they’re on a print you don’t have (ie. You have a machined part orint and not a raw casting print), so you might need to assume a relationship. If you need to troubleshoot a potential mistake you want to easily know what’s fact and what is an assumption.

 

[drawoh]

sluzzer,

Twelve A1 (594mm[×]841mm) drawings? Is this a complex casting, or bad drafting?

One of my favourite tests of any messy, complicated CAD drawing I have done, is to make a new 3D[ ]model from the drawing. With SolidWorks, I do not care that there already is a model. If I can 3D[ ]model the part quickly and efficiently, the drawing is good, and the fabricator should have no problems.

Will your AutoCAD drawing be better organised and more readable? If it is hard to make sense of your drawing for 3D[ ]modelling purposes, won't it be hard to make sense of it for 2D[ ]drafting purposes?

Perhaps you need to request a better drawing.

--
JHG

 

[Jboggs]

I feel your pain.
1) Spend some time trying to identify the major shape features without getting distracted by the smaller details. Do not even begin until you are confident of the rough three dimensional image you have in your head.
2) Identify as many symmetrical features as you can. Try to place your origin at the center of symmetry.
3) When you start, be very careful about which planes you select as front, right, top. Your initial selections here will have major impacts later when you try to manipulate the piece. Learn to think several steps ahead to plan your work.
4) Your final indicator of success will be when you can create A1 drawings and views from your model that are identical to your source drawing.
5) Try to ignore the small details at first so you can concentrate on forming the major shape features.
6) When I learned that you could link solidworks sketches and features to previous sketches, I began to start my more complex part models with one or two "geometry sketches". I don't use them as absorbed sketches but as definitions of overall part geometry. I name them GEOMETRY FRONT and GEOMETRY TOP. Then I build the following sketches and features based on elements of those sketches. That way I can control multiple features and overall part geometry from those initial sketches. It also helps to prevent errors as you refine your design.
7) Learn to use the less commonly used tools available to you. For example offset extrudes, in which the extrude doesn't have to start at the plane of the sketch. Another one is extrude to vertex, in which the length of your extrude can be determined by the position of a vertex in a previous sketch or feature. All this leads to more reliable complex models with fewer errors in development.

Good luck! Go for it.

 

[sluzzer]

Thank you everyone for inputs!
@3DDave & Drawoh
The part I received is a old (1970s) aircraft engine casting part! They are drawn without CAD! I really wonder how they managed to draft and manufacture without CAD in those old days! Huge respect to those guys for their imagination! But for me now, there is no model availability & no one to clarify any doubts! 3D scanning is one option which remains open if customer sends the part. But they are like "drawing is having all the details, plz model from it"!
@CWB1 & @Jboggs
Excellent inputs! Thank you! Once I complete, I will share my experience. Thanks once again.

 

[3DDave]

Well then, be aware that inconsistencies can abound in hand-drawn drawing, ones that the pattern makers just sandpapered and filed to cover. They depended heavily on the pattern makers to smooth transitions between features or simply missed the need and the pattern makers did whatever they felt best. They should send the part to see what the pattern makers actually did.

 

[JohnRBaker]

Yes, I worked with a company in Australia, Bradken, that built the 'trucks' for railroad cars, the carriage with the four wheels. Anyway, I was given one of the their '20-year drawings' and was asked to built a solid model of it (this was part of a sales effort and we agreed to take one of their more complicated parts and model it for them). Well, I think the designer I was working with came away with a better appreciation of the work that the pattern makers did. It was a large casting and in several places I noticed some strange dimensions, mostly the blends/rounds or fillets. Like in one view a blend/round would be labeled 25mm but in the side view where the same blend/round simply extended around a corner and down another side, it would be labeled 26mm. I asked the designer why the two different dimensions and he said well that's was probably what the engineer told him was needed. I then asked, do you really think the pattern maker ran a 25mm blend/round along one edge and when he got to the corner, transitioned it to a 26mm blend/round down the other edge? Back when patterns were made of wood and clay/putty, for fillets, pattern makers used their thumb and fingers to form the radius. They knew how close a thumb or forefinger was to a certain radius and it worked for them for years. In reality when they started to machine patterns instead of carving them by hand, they learned a lot about how pattern makers really did their jobs to say nothing of how poorly their drawings were dimensioned. CAD actually helped a lot, because if you could model something with consistent and appropriate dimension, then there was a good chance that it could be manufactured the same way. Of course, you now had to worry about things like taper and parting lines, again something that the pattern maker used to worry about.

But some companies actually took advantage of this sort of thing. I remember working with some people from Toyota in Japan and I was looking at some of their older drawings and I noticed that on many castings they never even drew in the blends/rounds or fillets, they drew the part with sharp corners and just labeled them with the size of the blend/round/fillet. The designer knew enough about the size of the part that he knew what was an appropriate size but why waste the time drawing them in when the shop was going to do what the drawing said anyway, irrespective of what the drawing actually looked like. Of course, for these people the concept of creating a 3D model that looked like the final part was very foreign to them and was the hardest thing that we had to try and teach them in terms of how to approach getting the model to represent what they actually were expecting to be manufactured.

John R. Baker, P.E. (ret)
EX-'Product Evangelist'
Irvine, CA
Siemens PLM:
UG/NX Museum:

The secret of life is not finding someone to live with
It's finding someone you can't live without

 

[3DDave]

Ha - that reminds me - I worked a job where that old "It's symmetric, just use mirror" once again blew up in people's faces. They drew up the casting for one side and then said the other side was "OPPOSITE."

Great job.

So the first article casting comes in. They used the wrong radius which was tied to a wall thickness. The inspector fails the part. Engineering is given the info, Program manglement decides to accept as is, so a drawing change was initiated to change the radius value on the drawing. Then QC inspects the OPPOSITE part. Before the change is approved. It meets the original drawing requirements.

Hilarity ensues, because not only didn't the two sides match, they changed the radius that was correct for one side to match the error on the other side, but left the "REF" radius that was hanging around in a separate view where the wall thickness dimensions was placed; two different dimension values for the same feature. I know "REF" doesn't count. But when I looked at the actual part, the REF was right on the one side.

Which leads to another caution - if you only got the drawing, you may not have gotten the waivers, so if they want the engine casting to be like the one they actually used, you need to get an engine casting inspected to the current drawing.

 

[sluzzer]

Thank you very much 3DDave & JohnRBaker!
I designed some complex aero engine parts. From that experience, I can say that we should understand the design before we starting the model.

For example, I designed a 'Fuel metering unit' for an aircraft engine. It is a part with several interconnecting ports and provision for valves seating. To design and model this part, I first positioned the valves and then used Tubes to model the connections. From the tubes, I offset the outer surfaces and merged/blended those surfaces to form the external surfaces. Finally, I subtracted the tubes to make ports and made the drawings.

Now, if someone tries to model the external surfaces first and then to internal port, I am sure it will be very difficult.

@JohnRBaker I bookmarked many of your NX threads and using it most of the time. Can you plz some more on the topic with respect to NX? As Jboggs mentioned in point no.7, is there any less commonly used / known tool in NX, which are very much useful in modelling complex parts?

Thank you.

 

[JohnRBaker]

sluzzer, I've been retired for better than six years now and haven't touched NX since. I had the option to get a courtesy license (for years I signed for a courtesy license for an acquaintance in Australia) but I decided that that I was going to cut the cord completely. This NX forum is the only thing that I've continued to follow.

John R. Baker, P.E. (ret)
EX-'Product Evangelist'
Irvine, CA
Siemens PLM:
UG/NX Museum:

The secret of life is not finding someone to live with
It's finding someone you can't live without

 

[Jboggs]

sluzzer: "I really wonder how they managed to draft and manufacture without CAD in those old days!"
Agree 100%! Want to blow your mind? Go to a hanger where they're working on some old warbirds and have some of the skin removed. I remember watching in amazement as they tried out the wing-folding mechanism on Corsair. They figured it all out by hand. Or once when I was able to see inside the engine cowling of a P-38. Nothing but tubes and fittings and brackets and connections. And absolutely no space to spare. A dropped penny would not have made it through. And they did it ALL with sliderules and manual drafting boards! This is beyond engineering. This is inspired art. My respect and admiration for those guys is overwhelming. I can only hope to do as good a job with the tools I'm given.

And speaking of art - see if you can find some of those original detail and assembly drawings!

 

[moon161]

Dave, yes, I deal with lots of drawings where non-critical dimensions of a casing or something were just left up to Gepetto in the pattern shop.