5 Replies Latest reply on Jan 8, 2009 1:37 PM by John Matrishon

    Best practice for modeling in regards to tolerances?

    Joe Dunfee
      I am fairly new to high precision manufacturing, so haven't really had to deal with this issue before.

      When modeling a part, do you model it at the average size for a part, or do you model it at maximum material so that any worst-case interferences are apparent?

      E.g. I have a part 2" +.001" / -.005" Diameter. Should I model it at 2" Dia, or perhaps 2.001" Dia?

      I know I can indicate the tolerances on the sketch, and bring them into the drawing. But, I am thinking about the model itself at the moment.

      Joe Dunfee
        • Best practice for modeling in regards to tolerances?
          Josh Brady
          It really depends on what you are using the model for and what is most critical. If you are doing the initial design and interference is critical, go ahead and model for MMC. However, if you are going to create a drawing and you want the drawing to show 2" +0.001"/-005" you are going to have to do some fanagaling to get it to show up the way you want. If you are going to be sending the model for CNC, a lot of places want it modeled to the middle of the tolerance zone, although some may want it cheated towards MMC to give room for error. It's always easier to cut material off after inspection than to add it back on.
          • Best practice for modeling in regards to tolerances?
            Harold Brunt
            A lot of Engineering Departments generate a design guide so that modeling this kind of part is consistant from project to project and designer to designer. If you don't have a guide you might consider generating one.

            Our SOP is to model mechanical parts at the nominal and then generate configurations for tolerance stacks as needed. I use the configurations to run ray traces at extremes so I need the actual models and I also take advantage of the inteference detection fuction in SW. If the model is fairly simple and will not effct the beam path, I'll use one of several online GD&T calculators. I think if you were to try to model every possible tolerance error (such as angular, positional, etc) your model will become difficult to manage.
              • Best practice for modeling in regards to tolerances?
                John Matrishon
                Joe, I was just thinking about this issue. We have battles with MFG about how the model is created. Engineers need to convey the design intent, and MFG wants to shoot for the middle of the tolerance so they have the best chance to have acceptable parts, and they want to use our models to put into CAM Works. It is very similiar to the "Engineering drawing" or "Manufacturing drawings" that many companies deals with. I've been at companies that have done either one or the other, and both.

                As Harold said, it's a good idea to develop a standard, that way when other use the model, they know what they are looking at.

                As far as using configurations to show MMC, it's really difficult to do as the parts get more complicated. When you start having relationships between features, you tend to not be able to make an accurate MMC model.

                I'm not a big fan of making models at MMC, I like to make them at nominal to relay design intent, and if we have to use a bilateral tolerance, then we use it. Let MFG alter the model to their liking, and be responsible for it. That way, the burden is not onto the Engineers who already have enough to worry about.

                Bottom line is, what does your company want to see, and how easy is it accomplished? Does it add value?





              • Best practice for modeling in regards to tolerances?
                Joe Dunfee
                Thanks all for your comments.

                John, Regarding what the MFG wants. I imagine they would love to have a piece of paper that shows them how the model is created, so they can determine if the model is appropriate for their purposes, or needs to be modified. Perhaps if the main dimension given is how it is modeled.

                In my case, I just finished modeling a part with a minimum and maximum dimension... that would be ambiguous about the model if I tried to apply the rule I just gave. E.g. 2.00" Max/2.01" Min. Perhaps I will just stop using that type of dimension to make it more apparent how the CAD model is done. E.g use instead 2" +0/-.01 .

                Joe Dunfee
                  • Best practice for modeling in regards to tolerances?
                    John Matrishon
                    It's not an easy decision. You want to make them happy, but we can only do so much. The "main" dimension you talk about is the nominal dimension, so people do believe the nominal is the middle number of the tolerance zone, but that's not what nominal means.

                    like you said if you only use symmetric tolerances, then your models will be fine for MFG.

                    If you need to do a tolerance study, then you either get out the paper & pencil or buy premium......

                    real tolerance studies seem to floating away at the companies I've been at due to lack of time mostly. The big companies still probably do them, but it's an art form that is dwindling in this fast paced world of "get it out the door"