1 2 3 4 5 104 Replies Latest reply on Jun 9, 2018 4:48 PM by Matt Lombard Go to original post Branched to a new discussion.
      • 30. Re: modeling methods
        Matt Lombard

        Do you think any of this depends on what you're designing, such as for machine design, a skeleton approach would work best, but for plastic assemblies a 3D master model works best?

        • 31. Re: modeling methods
          Vladimir Urazhdin

          We use parametrically driven Master Assemblies linked to Master Drawings.

          Skeleton Sketch works fine with Global Variables so I don’t see a reason to consider SSP as alternative way to parametrically driven design.

          • 32. Re: modeling methods
            Dwight Livingston

            Matt Lombard wrote:

             

            Do you think any of this depends on what you're designing, such as for machine design, a skeleton approach would work best, but for plastic assemblies a 3D master model works best?

            For myself, I will only do master model if I have multiple parts that share compound curved surfaces that are too hard to duplicate reliably. Such things do tend to be molded or cast.

             

            I won't do skeleton in our design environment.

            • 33. Re: modeling methods
              Steve Calvert

              I think Anna just slapped a few of us up side the face...

               

              Thanks Anna

               

              Steve C

              • 34. Re: modeling methods
                Anna Wood

                Matt Lombard wrote:

                 

                Good to hear from you again. Knowing you, that point of view is fully from experience.

                Thanks Matt... Glad to see you hanging around here again.

                 

                Yes, hard won experience.  We regularly pull up and re-use SolidWorks designs from 15 plus years and counting ago.  They need to survive through time and keeping it simple with regards to the parametrics helps with that.   I have ancient cad (dxf) files from 30 years ago that we need to make changes to on occasion.   I expect the same shelf life and longer for our SolidWorks models.

                • 35. Re: modeling methods
                  Roland Schwarz

                  I never knew. Now I learned. THANKS.

                   

                  They could still do one better and put the orientation definition somewhere more obvious and intuitive.

                  • 36. Re: modeling methods
                    Dan Pihlaja

                    John Stoltzfus emailed me and wanted me to post this (he can't get online today):

                    Designing For Change is not about Master Modeling verses
                    SSP, or SSP verses whatever, Designing For Change, is not specifically about
                    technique it is about process.  Here I have used all of the above in
                    conjunction with the other.  Pitting one against the other is not about
                    Designing For Change, it only spurns the negative, Oh don’t ever use that
                    method” analogy.  There is a Master Model floating around on the Forum,
                    open the Master Part and change any dimension in Sketch1 and tell me what
                    happens.   I don’t care if my co-worker would design under water or
                    upside down, I just want to open the model, do a pack and go make the changes
                    required and bingo everything needs to move or he’s not doing his job.
                    The only reason for me wanting to respond this way is so that people, whoever
                    you are, get stuck in the rut of never trying anything else.  And please
                    believe me, I want to open doors, open horizons not close them. 

                    • 37. Re: modeling methods
                      S. Casale

                      yes yes and yes. I concur!

                       

                      The origin of the model statement is right on. Makes it easier when sectioning at NHA's.

                       

                      The issues we have with modeling 100% is the fact that we use the routing program for wires and cables, which messes with overall weight calculations. We also use CF step models at times.

                       

                      We don't break links as I know we don't keep links between parts and etc. This is due to user knowledge on how to handle files in PDM. I've tried .

                      • 38. Re: modeling methods
                        Matt Lombard

                        I'm not disagreeing with you about the origin, but I'd like to bring up just a couple of things, just for discussion, not an argument.

                         

                        First, what happens when you have a part that is not symmetrical in any axis? Where do you put the origin of the part in that situation?

                         

                        Second, if you had a good reason for creating a part in-context, and the part origin winds up at the assembly origin, what would you do in that situation?

                         

                        Third, I've had machinists ask me what kind of nincompoop I was for not putting the part 100% in positive XYZ space. I've heard the same argument for 3D print. What do you say in that situation?

                         

                        Isn't there a way to place a coordinate system in SolidWorks? Doesn't a decent CAM package have the same capability?

                         

                        The needs for design aren't always the same as the needs for CNC. With simple stuff like blocks it might be possible to always work that way, but not with more complex parts, especially in more complex assemblies.

                         

                        This is one of the reasons it seems pretty clear to me that you can't impose a single system on everything. Most of the plastic modeling I've done would absolutely not work this way, even if I knew what I was making when I started, which I usually don't. What kind of assemblies/products are you creating with this method?

                        • 39. Re: modeling methods
                          Matt Lombard

                          Dan,

                          This might sound harsh, but it's not meant that way. I've been around enough to see a lot of people do a lot of things, and I know everyone has their own reasons, and I'm not there to see everything.

                           

                          But, it kind of terrifies me every time I hear people not use PDM. Folders can go south so fast. There is no undo over the network. Keeping a history of what you've done in folders is almost impossible. What do you do with parts that are not library parts, but are still shared between two projects? It seems unnecessarily risky. And then there's all the extra reporting and statistic capability you get.

                           

                          I know some people don't really understand PDM, so they don't trust it, but the truth is if you make a mistake and need to go back, that's so much easier to correct in PDM. It's not that hard to use once you work through a project with it. And isn't it fairly cheap these days?

                          • 40. Re: modeling methods
                            David Matula

                            Dan Pihlaja wrote:

                             

                            John Stoltzfus emailed me and wanted me to post this (he can't get online today):
                            The only reason for me wanting to respond this way is so that people, whoever
                            you are, get stuck in the rut of never trying anything else. And please
                            believe me, I want to open doors, open horizons not close them.

                            I condensed the quote.  Cause I have been there seen that.  Structural member assemblies that fall apart when they get bigger or smaller.  Darn things are bolted together and holes would keep moving. I would look at the tree and it was a like they threw all the parts in the top level.  You would have expected something like that in the first few versions of SolidWorks, not the early 2010's.  Not since they came out with weldments.

                            another good one was that every hole that needed a bolt was a cut extrude.  If your paying for a tool box, why not use it and let the program give you a hand with the bolt count.

                                 it was a night mare, model like your using SolidWorks 2005 but get the job done in less time than before.  No need to attend a rollout session to see what is new, no desire to attend a user group meeting.  This is how it worked back then and this is how we are going to keep doing it.

                                  For those of you that automate parts of your design, be sure that you keep someone around that will be able to keep it up to date, and that will be willing to upgrade it to use some of the new features of the program.  

                            • 41. Re: modeling methods
                              Chris Saller

                              Matt,

                              I 100% agree with you. I think everyone should use some form od PDM, even if they don't like it. It will save time and headaches. PDM Standard is free.

                              • 42. Re: modeling methods
                                Matt Lombard

                                I'm not sure exactly what John's getting at, and maybe I didn't say something very well. I don't think we are in disagreement. I just think that the different methods, taken generally, like 2D skeletons and 3D master models, are best used in different situations.

                                 

                                "Design for Change" is a phrase I've used in place of "design intent" for the last decade and a half. I'm pretty sure we agree about what that means, although John may have something more specific in mind.

                                 

                                For example, how would you model a part like this below from a sketch? This assembly was very difficult to visualize. I didn't believe it was manufacturable until they sent me some parts (the orange part is hollow around the big thick section).

                                9781119300571 fg0108.png

                                 

                                There aren't that many parts, the individual parts can have a couple hundred surface features, and it has some 2D characteristics, but you couldn't drive it all that way. The only real reference plane that makes any sense is the one that goes down through the middle. Not all the parts are symmetrical. This one required a 3D master model.

                                Here's another one:

                                9781119300571 fg0104.png

                                9781119300571 fg0109.png

                                This one was designed as a single part and broken up at the end. The design and the modeling were happening at the same time.

                                 

                                Or this one:

                                9781119300571 fg0112.png

                                The main housing was done as a single part, and then broken up. This is what I call master model.

                                 

                                The coaster brake bike I did used a skeleton sketch to control the frame, wheels, fork, handle bar, seat. For this kind of thing, the skeleton works, but for the plastic assemblies, it's simpler to drive it from a 3D master.

                                coasterbrakebikebananaseat.png

                                • 43. Re: modeling methods
                                  Dan Pihlaja

                                  Matt Lombard wrote:

                                   

                                  Dan,

                                  This might sound harsh, but it's not meant that way. I've been around enough to see a lot of people do a lot of things, and I know everyone has their own reasons, and I'm not there to see everything.

                                   

                                  But, it kind of terrifies me every time I hear people not use PDM. Folders can go south so fast. There is no undo over the network. Keeping a history of what you've done in folders is almost impossible. What do you do with parts that are not library parts, but are still shared between two projects? It seems unnecessarily risky. And then there's all the extra reporting and statistic capability you get.

                                   

                                  I know some people don't really understand PDM, so they don't trust it, but the truth is if you make a mistake and need to go back, that's so much easier to correct in PDM. It's not that hard to use once you work through a project with it. And isn't it fairly cheap these days?

                                  I totally understand and totally agree.  I have pushed for PDM here since I started (6 years ago). I HATE the folder system.   Unfortunately, there are 2 guys here that absolutely refuse to even consider it.  And they are in positions of power.

                                   

                                  Their reasoning is:

                                  1) We are too far into the folder system.  It would take too long and cost too much money to switch over (we have a LOT of duplicate parts....*sigh*....on top of that....the folders are not clean....mine are!   But not everyone's).

                                  2) The last time one of them used it....some people at their company went around it and it caused more problem.

                                  • 44. Re: modeling methods
                                    Alex Lachance

                                    I don't use PDM but I use a 3rd party program that does some of the functions that I would require from the PDM. Our files are all stored on servers and the way we work is that there's a back-up made of the drives containing the drawings, parts and assemblies twice a day, every day of the week. Those back-ups are stored for 2 months before being purged.

                                     

                                    We didn't go the PDM way because the company I work for tought it was too expensive. There are some extremely interesting things in PDM that I've read on these forums and I'd love to see it in action on even work a week with one just to see how everything works. I believe it would be fairly hard to implement now that we're 5 years into SolidWorks.

                                     

                                    We don't really have a problem with back-ups and fixing mistakes because what we draw is generally produced within 2 months , but the company has been growing a lot the past few years and we might hit a wall eventually.

                                     

                                    I try to build using the planes too because I find that they are often better references to do mates with, especially if you need to switch between parts/assemblies a lot. At first, our goal was to build everything  accordingly to the planes the project was built on but that didn't go as planned. We left out the need to orientate the parts accordingly to the project but we try to build sub-assemblies accordingly to the project's orientation.

                                     

                                    Concerning your post about molding and machining, I've never had to deal with molding, I'm still rather young(About to turn 30) so I've only worked in the domain I'm in right now, which is Trailer manufacturing. Our machining is also very basic so we don't use any CAM software either.

                                     

                                    As you said, what works for me might not work for others, it all depends what domain you work in.

                                    1 2 3 4 5