[Me...] Posted June 20, 2019 Share Posted June 20, 2019 In my 15 year Zeiss Programming history, it has always been a small to huge issue in regard to diameter readings compared to physical gaging. I understand the 5 evaluation settings and what they're used for, the problem is that most others in manufacturing do not. I also have not seen one drawing that required any of those 5 evaluation methods, but I have heard of them being required. The company I currently work for has not made any issue about it but there are operators that like seeing the numbers and try to hold roundness as best as they can. My question to you fine ladies & gentlemen: What do any of you do in these questionable situations? Is there anything in a standard about this? Link to comment Share on other sites More sharing options...
[Da...] Posted June 20, 2019 Share Posted June 20, 2019 I deal with close tolerance IDs almost exclusively. Diameter tolerances below .0005" as low as .00005". Of course we are talking mirco below 10 ra. It is important to check the roundness of each and every circle that you check . If roundness exceeds 1/2 of the diameter tolerance, you probably won't get a correlation. If the roundness is 3 lobed or 2 lobed that will also make for a difficult time syncing the size with a 2 point check. Remember a 2 lobe will produce almost double the diameter variation as you move around the circle. A 3 lobe bore will act tighter than the 2 point check will show. I get best correlation using LSQ, probably because of the tight tolerance and very low roundness. Link to comment Share on other sites More sharing options...
[Br...] Posted June 20, 2019 Share Posted June 20, 2019 Please sign in to view this quote. The Y14.5 standard doesn't tell you how to inspect or manufacture a part. Its focus is on design intent. That being said, you as the programmer/inspector have to make the decision as to what methods best represent the standard you are working to. For instance, A position tolerance on a cylindrical feature of size at RFS locates the Axis of the cylindrical feature. The way the feature axis is determined per Y14.5 is by the actual mating envelope. Y14.5 defines the Actual Mating envelope as "A similar perfect feature(s) counterpart of smallest size that can be contracted about an external feature(s) or largest size that can be expanded within an internal feature(s) so that it coincides with the surface(s) at the highest points." Minimum Circumscribed/Maximum Inscribed best represent the definition of an actual mating envelope in this case. Rule #1 governing limits of size states that there is a boundary of perfect form at MMC that no surface of the feature may violate. Once again, Maximum Inscribed/Minimum Circumscribed on a circular or cylindrical feature being evaluated for a diameter will verify if the boundary(envelope) has been violated. Same goes for any features. Having a good understanding how datum reference frames are established, and how features fall within their respective tolerance zones will help you in your discernment of which fitting algorithms to use. Link to comment Share on other sites More sharing options...
[Me...] Posted June 20, 2019 Author Share Posted June 20, 2019 Thanks guys. To play devil’s advocate here... Say I measured a bore right at nominal using least squares that has a tolerance of + / - .0005 inches. If my roundness is a little above where I want it like say .0006, even though there is no roundness dimension on the print, is that considered ok to you guys? Link to comment Share on other sites More sharing options...
[Br...] Posted June 20, 2019 Share Posted June 20, 2019 Please sign in to view this quote. The form of an individual regular feature of size is controlled by its limits of size. If the feature is produced at its MMC limit of size, it must have perfect form. So as your feature departs from its MMC size toward its LMC size, a local variation of form is allowed equal to that departure. So to answer your question: to put it in simple terms, if a cylindrical feature is produced at its "split limit" with a tolerance of ±.0005" , then the form(cylindricity/roundess) cannot exceed .0005". If produced at its LMC size, the form can be up to .001". Link to comment Share on other sites More sharing options...
[Me...] Posted June 20, 2019 Author Share Posted June 20, 2019 Ok so said diameter happens to be a bore, Datum -B- on print with a .000M T.P. callout to -A-, which is a plane. The diameter is controlled by MMC, the evaluation for the diameter characteristic is Outer Tangential. Let's say that said diameter isn't a Datum and has no GD&T reference. The evaluation setting for the diameter characteristic will always be LSQ right? Link to comment Share on other sites More sharing options...
[Br...] Posted June 20, 2019 Share Posted June 20, 2019 No I would evaluate it as Outer Tangential or Maximum inscribed. This will verify it doesn't violate the boundary of perfect form at MMC. Make sure you're using the proper filtering and outlier elimination as those evaluations are susceptible to outliers. Calypso is going to force you to use the same evauation settings for your Position Characteristic as the Diameter characteristic used to evaluate the MMC bonus tolerance on the Position. Link to comment Share on other sites More sharing options...
[An...] Posted June 21, 2019 Share Posted June 21, 2019 Evaluation LEAST SQUARES: Give data to the machinist. Evaluation of capability (cp/cpk) Evaluation OTE: Cause of failure analysis Link to comment Share on other sites More sharing options...
[To...] Posted June 21, 2019 Share Posted June 21, 2019 Please sign in to view this quote. And when the customer's mating part doesn't fit, how do I explain that we used LSQ during the run and had excellent process control but your parts don't fit? Sorry! Link to comment Share on other sites More sharing options...
[Me...] Posted June 21, 2019 Author Share Posted June 21, 2019 LOL exactly Tom. Link to comment Share on other sites More sharing options...
[To...] Posted June 21, 2019 Share Posted June 21, 2019 Not to sound negative but this concept always bothered me about the cookbook. I realize the cookbook is not the original source this concept. I understand using LSQ is more stable when you're running SPC but at the end of the day, the part has to fit. Maybe someone could shed some light on how I'm suppose to satisfy both sides. Link to comment Share on other sites More sharing options...
[Br...] Posted June 22, 2019 Share Posted June 22, 2019 The Cookbook I think explains their "Process Control" type measurements as only for in process measurements to detect tool wear, do Gauge R&R studies, etc. Not to verify finished goods. At least that's how I understand it. Link to comment Share on other sites More sharing options...
[An...] Posted June 24, 2019 Share Posted June 24, 2019 Please sign in to view this quote. Form tolerance avoids OTE. It's up to the designer to put it in. See attached.Contribution_24_06_2019.pdf Link to comment Share on other sites More sharing options...
[Me...] Posted June 24, 2019 Author Share Posted June 24, 2019 Please sign in to view this quote. I try to work with machinists so that the results we create make sense. I do not agree or disagree with anyone on this thread & I use this forum for advice and as a reference, but I have to know Andreas... How would you explain your results to a customer if they couldn't properly assemble their parts and it was because a hole maybe too small for the mating part to fit in? Link to comment Share on other sites More sharing options...
[Me...] Posted June 24, 2019 Author Share Posted June 24, 2019 Please sign in to view this quote. You can't. The envelope rule, by definition, won't allow you to use LSQ to verify assembly. It's a moot point anyway. All portions of a feature need to be in tolerance. You can't average a diameter and technically call it good if any portion of it is not in tolerance. Think 2 point diameter measurements. Robert Link to comment Share on other sites More sharing options...
[An...] Posted June 25, 2019 Share Posted June 25, 2019 The distance of opposing points on the surface is also limited by the cylinder form tolerance. (if using LSQ Ø) About envelopes and least squares. See attached.Contribution_25_06_2019_a.pdf Link to comment Share on other sites More sharing options...
[Ch...] Posted June 25, 2019 Share Posted June 25, 2019 It sounds like we are just arguing about ASME vs ISO. The Y14.5 standard has always considered the control of all points on the surface of a feature. By extension, the ‘mating size’ concept is used to describe the size of a feature, meaning every point on the surface of a feature must fit within a boundary having the specified size. For example, if the diameter specification for a cylindrical pin is 10.0 ± 0.1, the entire pin must fit inside a theoretically perfect cylindrical shell of diameter 10.1. Alternatively, the ISO standards consider the least squares size of a feature as the default. This means that some of the points might lie outside of a boundary that is defined by the size specification, while still conforming to the size specification. For the same size specification as above (10.0 ± 0.1), a pin whose least-squares size is 10.1 would conform to the specification although points on the surface could exceed the perfect boundary of diameter 10.1. -- https://www.qualitymag.com/articles/934 ... comparison Link to comment Share on other sites More sharing options...
[Br...] Posted June 25, 2019 Share Posted June 25, 2019 I'm a bit surprised that the quality mag got the Rule #1 a bit confused here. It isn't quite right to say the points need to fit into a "cylindrical shell", and that cylindrical shell is the boundary that is talked about in Rule #1. The boundary is ONLY at MMC. The actual local size controls the size at LMC. That isn't really an outer boundary though it does still control form. The boundary is always referred to in Y14.5 as only existing as perfect form at MMC. You can override Rule#1 with an indepenency symbol for example which drops the MMC boundary, in which case your feature is only controlled by its actual local size at LMC and MMC. This would be ISO's default, wheres in ISO you apply the "envelope" symbol (an E with a circle around it), to the size dimension, or a note in the title block that says "ISO 2768 -E" to get the MMC boundary. To me, neither of these examples vindicate the use LSQ. Even when only evaluating the "Actual local size" of a feature of size, LSQ is not going to bring up those maximum or minimum point-to-point measurements. It finds the average. Calypso actually has a specific characteristic tolerance to handle this called "two point diameter". This will align perfectly with all of these concepts. It's based on the ISO envelope requirement, but it is the exact same thing as Rule #1 in Y14.5. You can also drop the envelope with this characteristic and only evaluate the actual local size. Maybe ISO actually says you can use LSQ in their standard, but if that's the case I'd like to know more about how they look at the limits of size to truly understand why that would be allowable. Link to comment Share on other sites More sharing options...
[Me...] Posted June 25, 2019 Author Share Posted June 25, 2019 Link to comment Share on other sites More sharing options...
[Ba...] Posted January 23, 2020 Share Posted January 23, 2020 Please sign in to view this quote. If I have evaluated my diameters(s) involved in a true position as Outer Tangential, should I keep that evaluation specified by calypso, never change? I noticed a considerable difference when changing from outer tangential to lsq. Link to comment Share on other sites More sharing options...
[Br...] Posted January 24, 2020 Share Posted January 24, 2020 Please sign in to view this quote. It depends on the application. If you have a cylinder feature being evaluated for position, then you should use something like Outer Tangential because it reflects what the GD&T standards describe--namely, it evaluates the axis of the unrelated actual mating envelope. If you also have a circularity tolerance applied to the same feature, you can use Minimum feature because you want to optimize the fitting of all the surface points into a tolerance zone. If it is a diameter tolerance I would use Outer tangential to simulate the Actual mating envelope to satisfy Rule #1. (There's even a characteristic where you can do both the AME, and the actual local size but most people don't want to see that much information on a drawing.) When you are dealing with MMC, Calypso will force you to use the same evaluation methods for both the size characteristic and the position characteristic. Link to comment Share on other sites More sharing options...
[Ba...] Posted January 24, 2020 Share Posted January 24, 2020 That makes sense. In my case we have a (TP ø .0035 to B) on a bore (cylinder) perpendicular to B which is also a cylinder but an OD. There is no MMC/LMC given. I evaluated the size of each as Outer Tangential and applied filters/outliers. When I created the true position of them together, calypso defaulted to outer tangential but the very large .015 result did not seem correct. I then changed the features in the true position to LSQ and my result was a more believable .0045 A manual table, indicator check was close to the .0045 so I was leaning to the LSQ method but want to be sure I am not missing something. I will be presenting data to engineer and vendor so I need to be certain this is correct!! Link to comment Share on other sites More sharing options...
[Br...] Posted January 24, 2020 Share Posted January 24, 2020 Please sign in to view this quote. It sounds like you might have some outliers affecting your data. the Outer Tangential evaluation is very susceptible to outliers. You might even want to take a look at the actual data points and see if something weird is going on. Link to comment Share on other sites More sharing options...
[Cl...] Posted January 24, 2020 Share Posted January 24, 2020 Barb, your name sounds very familiar. Have we ever worked together? Link to comment Share on other sites More sharing options...
[Ba...] Posted January 24, 2020 Share Posted January 24, 2020 Thanks Brett! Clarke, not that I am aware.? I have been doing calypso training in NC this past year and met some from your area. Other than that most work in VT/NH area Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in