Constraints/Dimensions or No Constraints/Dimensions in 3D Modeling

[GreenTwin]

The 20 hp Stanley was not very difficult to model or sketch.
The Stanley drawings from the Stanley museum are quite good and accurate.

 

[GreenTwin]

The Mississippi has probably been the most difficult engine to model.
The original drawings from 1840 are terrible, and they have to be comletely reverse-engineered because they have errors in them.

I have create each part exactly as shown on the 1840's drawings, unless there was an error I had to correct.

The gothic columns were not fun.
And everything is modeled with the intent to make 3D patterns and cast this engine, so it is also a 3D model with pattern intent.

Many hobby engine builders will simplify engine designs in order to make the model building process easier.
If I have drawings or good prints/photos for an engine, I like to make a model exactly like the original, with no simplifications.

The Mississippi engine has what seems like a million tiny details.

 

[GreenTwin]

The shell feature works pretty well with simple shapes, but does not work very well with complex shapes that may have draft angle, or non-linear features.

The shell feature is one thing that causes me a lot of problems, especially with pattern making.

.

 

[GreenTwin]

I think I touched on this my other post but just to put pictures to words, here is a series of screen grabs that hopefully clarifies what you mentioned. The dimensional values in the properties box and the dimensional values of the classic arrowhead dimension (using smart button or whatever) are one & the same. If you adjust either one, the other will update automatically. The reason you see values in the properties box and not as dimensions is only because you didn't bother to use the dimension tool or subsequently deleted it from view. That is equivalent to being turned off or hidden. But hiding the dimension is not the same as deleting the numerical dimensional value. Its still there, in the database to use your words.

My screen grabs

1) from within SW, sketch a random line going off in space, click to begin, click to end. Note how SW is actually conveying the current length in the little box as it rubber bands? Now if I prefer, I can enter a keyboard value before terminating the line it will make it that exact length & auto-terminate. Anyway once its sketched, it is blue indicating 'not fully defined' (a SW term).

2) now look at properties box. The rough sketch reflects its current length & angle values however it was drawn, hence all decimal places. My point is the numerical values match identically to however the sketched line terminated. they are one & the same.

3) now I go back to sketch & define the length and angle to what I really want. The line segment now turns black confirming SW agrees its fully defined. It tells me that visually & numerically on one screen view

4) now look at the Properties box, it has reflected these changed value & once again they match

What you are doing is importing ACAD geometry that +/- matches step 3) but without the benefit of co-displayed dimensions when it would be advantageous & useful to see. You did 1&2 in ACAD which you are more proficient at, but dimensionally its the same data. Negating dimensions in SW sketch is not saving any computation time, or improving accuracy or any advantage I can think of. Its just how you prefer your current workflow which SW facilitates. IMO its equivalent to generating a Drawing of a part but neglecting to display critical dimensions. The actual dimensions are indeed numerically known within the model. But it's just not being conveyed at the appropriate stage where its most useful. In the case of a drawing, for the machinist. In the case of a sketch, for the designer. Even if you are the sole designer & CAD file never to be shared, maybe you will open that file & wonder - how did I cook up that sketch? what file version is it in? does that version match what's now in SW ? Did I tweak that dimension in ACAD or SW after the fact? etc. etc.

Sorry, you seem very disciplined in your approach but for me, all I see is a minefield of potential things to go wrong. But that's just my own personal opinion. Often I take screen grabs of different design mods or decision points or whatever & save them as jpegs or in a OneNote file for personal sanity. Even that pseudo documentation process is hard to connect the thought process a year later when it looks less familiar & especially if parts/assemblies start getting more complex.

I can see the power of learning how to sketch and utilize all of the features of 2D sketching in a 3D program.
One day I will get there.
It won't be this year.

.

 

[awake]

I think you are using the wrong terminology.

The values in the database define the line, ie: start point, end point, length, angle, but they do not constrain the line to vertical, horizontal, parallel, concentric, etc.
The contraints typically are set manually, unless your program does some of it like Solidworks where it can set a vertical or horizontal constrain on its own.

You guys are mixing and matching terms, which is what makes it so confusing.

.

I can't speak for others, but I would say the fault lies not with the participants in this conversation, but rather in the fact that different software has (apparently) adopted different language for different things. If you have a line with "start point, end point, length, angle" defined, then you have indeed constrained the line precisely. You may not have said, "make this line vertical," but if the angle is 90°, well ... there you go.

Based on what you have said, I am guessing that SolidWorks does not call any of that "constraints." However, any and all of that is considered a constraint in other CAD programs (at least in FreeCAD, and it sounds like perhaps also in some of the others). There is no separation of "definition" of the line from "constraining" the line; they are the same thing in FreeCAD et al.

So yes, confusing indeed ... we are speaking different dialects, and getting frustrated that the other does not understand the obvious things we are saying ... which mean something different in their dialect. Certainly it would be nice if all the CAD programs everywhere got together and agreed on a common set of terms, but I'm not holding my breath for that to happen ... !

No doubt the SW folks are confident that their terms are the correct usage, while the Fusion360 folks are equally confident that theirs are correct, and so on ... but I would humbly suggest there is no absolute standard by which to judge which usage is better than another. I can fuss all I want to at the way the folks across the pond "wrongly" insist on calling the covering over the engine a "bonnet" instead of a "hood," but a better use of my time is to learn the dialect so that I can communicate clearly.

 

[petertha]

Yes I agree. In any CAD discussion, qualifying sentence #1 should be 'I am using software X' because things may well be different in software Y

I cant speak for other packages, but SW constraints are very well documented. Most everything we have discussed has pertained to constraints related to sketching.

https://blog.epectec.com/what-does-...imensional shape,angle corner in your design.
But constraints extend well beyond that to other facets of 3D modelling, particularly interaction of Parts within Assemblies, motion studies, simulation...
https://help.solidworks.com/2021/English/SolidWorks/cworks/t_Defining_Constraints.htm

 

[GreenTwin]

This thread has been surprising to me in that it revealed just how many different terms and definitions there are in the 3D world, and how different and confusing they are.

I sort of assumed most 3D programs were on the same page as far as definitions of terms, but apparently not, so that does really confuse things.

Better to know about this though so we can clarify if offering some thoughts about 3D.

.

 

[petertha]

I can see the power of learning how to sketch and utilize all of the features of 2D sketching in a 3D program. One day I will get there. It won't be this year.

And I have yet to melt any metal... unless we can include what drips off my soldering iron LOL. Yes, only so many hobby hours in week.
OTOH, your castings, related infrastructure & experience is pretty awesome. I think I could eventually get to the pattern/mold stage base on somewhat similar projects. But the rest of the bubbly brew process is both mysterious & intimidating, therefore casting is on the 'maybe one day' list.

 

[GreenTwin]

This thread has also made me deep dive into the various terms and functions and better learn what can be done in 3D, and that is basically why I get involved in these conversations, because I learn a lot every time.

.

 

[GreenTwin]

I have mentioned a little history of 3D modeling usage since I discovered this forum in 2011, but at the risk of repeating some things, I will summarize it again.

In 2011, the only folks I recall doing 3D modeling were Brian Rupbnow (the Engergizer Bunny of 3D modeling), and a guy in England named Rob Wilson.
I watched what Brian was doing closely, and asked him some questions about 3D modeling.

I was more interested in Rob Wilson's work because not only was he doing stellar 3D modeling work for model engines, but he also had a foundry, and was making some spectacular engine castings.

Rob Wilson was instrumental in me being able to use Solidworks, and he shared a lot of training material with me (thanks Rob !).

I don't recall any 3D modeling training videos on youtube in 2011, but I am sure there must have been some.
I certainly don't recall any 3D training videos on youtube that showed how to design model engines.

There were some on this and other model engine forums that were openly dismissive of the potential benefits of 3D modeling as applied to model making.
Most folks did not use 3D modeling, and did not know how to start using 3D modeling.

It is very exciting to see 3D modeling for model engine building become somewhat mainstream.
It wasn't always that way, and there was not the support or encouragement then that there is now.

And when lower cost 3D printers came along, I realized potential of 3D modeling as far as making patterns for foundry work.

So while 3D modeling is a very potent design tool for the model engine builder, it can also be a very good way to make patterns by using a 3D printer.

On the casting side of things, the 3D printed patterns can be extremely accurate, and with the right sand, the castings can also be extremely accurate.

Now it is possible for hobby folks to make a better Stuart than Stuart, and a better Cretors than Cretors.
Saying this in 2011 would earn you the label "crazy".
Thanks to the Rob Wilsons and Brian Rupnows of the world, today I can show you exactly how to do it.

When I first started using a computer in school, it was punch cards fed into an IBM mainframe, with a giant line printer.
The programming in FORTRAN, which is a fantastic scientific programming language.
No screens, no mice, no keyboards (other than with the card puncher), etc.
Very primitive stuff, and yet we went to the moon with it.

I find it phenomenal how far things have come in what seems like just a few years.

I learned 3D modeling somewhat late in life.
I have never let age slow me down as far as learning new technology.
You can actually teach old dogs new tricks, if the dog is willing to learn new tricks.

I am going to add "Solidworks User" to my signature, so that others will immediately know which 3D program I use.
I would hope others would include the 3D program they are most proficient with in their signature.

.

 

[GreenTwin]

Rob Wilson is a member here.
Last post in 2016.
Anybody know what became of him?

His foundry and 3D work was inspirational to me, and he helped me learn 3D, and provided learning materials.

Here are a few of Rob Wilson's foundry/pattern/casting photos.

 

[GreenTwin]

Here are a few of Rob Wilson's foundry/pattern/casting/3D modeling photos.

 

[GreenTwin]

Here are a few of Rob Wilson's foundry/pattern/casting/3D modeling photos.

 

[GreenTwin]

Here are a few of Rob Wilson's foundry/pattern/casting/3D modeling photos.

 

[GreenTwin]

Here are a few of Rob Wilson's foundry/pattern/casting/3D modeling photos.

After seeing Rob's pattern and foundry work, I decided "I have to do this".
It was a pivotal moment for me.

.

 

[Richard Hed]

C'mon. All you 3D CAD junkies. You all have so many opinions, but I'm seeing little examples of your work, besides Pat & Jason.

Would like to see some of the examples expressed in your posts.

OK, here is one I did for a new cross slide for a Grizz G4003G. I want a better cross slide so I can attach either a milling attachment or the other accoutrements like a back side tool post, attach parts, etc. like lathes USED to be made or like small lathes are done today.

 

[johnmcc69]

Yup. I didn't even want to bring up surfacing, shelling, draft angles... that's where the rubber hits the road. The underlying sketches are very rarely 'standalone-able'. They are more likely to be based on more complex geometry of each subsequent feature step and/or have interconnected references many levels deep in the tree.

My induction tubes are generated from an assembly level. Its a simple 2D circle swept on a 3D curve. But, the curve is defined by multiple references - start & end planes, bend radius, included angle, tangent straight segment... This is an example where making a part in isolation & mating to assembly will not work.

Nice work Peter! You get it.

OK, here is one I did for a new cross slide for a Grizz G4003G. I want a better cross slide so I can attach either a milling attachment or the other accoutrements like a back side tool post, attach parts, etc. like lathes USED to be made or like small lathes are done today.

Nice little project & model!

Sometimes things do go wrong, but generally it is due to using a very complex starting sketch, with perhaps a lot of splines in it, not due to a problem with sketching in AutoCad and importing into Solidworks.

The more complex the sketch, the higher the chances of something going wrong, or the higher the chance that the compound surfaces generated will not fillet correctly.

I will try and dig out some complex sketches.
I am sure someone well versed in the 2D aspects of a 3D program could work at a good speed.

This bottle engine frame was a bit odd to sketch and model, and I had to try several different things to find something that would work reasonably well.

You can see the beading at the top of the cutout is not working like it should.


View attachment 141555View attachment 141556View attachment 141557View attachment 141558

Like I think I mentioned, I haven't ran SW in years & use Pro-E, so I can only speak with that softwares "terminology". To create that kind of beading, I would use a "swept trajectory" along that curve (the cut out). I would choose the cut out curve as the trajectory to follow & within the command I would then sketch the bead profile & the sketch would follow that curve.

How did you create the beading in your model?

Thank all you fellas for sharing your examples. Seeing what you're trying to accomplish makes understanding all these things much easier.

John

 

[GreenTwin]

How did you create the beading in your model?

As I recall, I sketched out a spline, probably in 3D, and then lofted a circle along that line.
These may not be the exact terms that SW uses, but you get the idea.
It is possible that I just picked the line along the edge of the opening.

This model was created before I started saving the screencaps of most of the steps, otherwise I would glance back at the screencaps.
One can still figure out how something was created by picking the shape and or lines associated with the object, and opening the sketch and/or the extrusion.

Solidworks is a pretty robust program, but definitely not flawless.
If you create something complex enough, you can throw off SW, and totally confuse it.
The loft along a spline command should work flawlessly, but does not always work, especially with compound curves.

Most SW commands work as intended.
Some SW commands do not work always as intended, but there are always work-arounds in any 3D program.

If a function is suppose to work correctly, but does not, I consider that a bug in the program.

Solidworks does have a polished look and feel to it, and is pretty well known to be very stable.
I can't recall Solidworks ever crashing while I was building a model.

I could probably create sections of beading, and not try to bead the entire window perimeter all at once.

.

 

[GreenTwin]

These discussions of some of the finer details of 3D modeling and features pushes the envelope of understanding for me and probably for others too.

While I am sure every 3D modeling user here will never agree on exactly which is the best approach, methodology, or 3D program, we do all reach a better understanding of 3D modeling by comparing notes and discussing methods.

I learned a lot just by observing what Brian Rupnow and Rob Wilson were doing with 3D modeling for model engines.

I have learned a lot just with this discussion of constraints and relations.
I am anxious to try programming some relations, if I can find a good application for an engine to use that function.

Sometimes you have to learn 3D modeling as you go; it is beyond me to understand all of it immediately, if ever.

.

 

[Bazzer]

GreenTwin

I have read quite a bit of the recent threads and taken in your unorthodox approach to importing 2D sketches from AutoCad, all I can say is that it clearly works for you.

I would add that AutoCad had a 3D modeller called Mechanical Desktop, that programme was so rigid that it would not allow a sketch to be used unless it was fully defined and dimensioned.

Anyone else remember the awful Mechanical Desktop?