|
The Release 18.0 Part Modeling User's Guide page 10-63 states: "The Accuracy option modifies the computational accuracy of geometry calculations." This means that Accuracy gives the user the ability to control the quality of Pro/ENGINEER calculations of model geometry with respect to an exact mathematical solution. Certain types of geometry require little computational effort to determine an exact solution, such as the geometry of a plane, cylinder, or cone. Other types of geometry, such as the intersection of two blended spline surfaces, require much more complex mathematical calculations to determine a solution. In these cases, Accuracy affects the calculations of geometry in order to achieve an acceptable level of approximation of the exact solution. Accuracy also affects the representations of the solid geometry to allow efficient memory usage, storage, and display. Typically, a default relative accuracy of 0.0012 allows geometry to be calculated with a reasonable amount of computation and within a reasonable amount of time. Sometimes, however, specific model geometry may require that geometry calculations be sensitive to fine features or complex geometric shapes. Modification of accuracy for a model with this higher "level of detail" may be used as a last resort to assist Pro/ENGINEER in solving the model geometry. Important Note - That...
2) accuracy does not affect angular calculations (eg. when edges are considered tangent or planes parallel). Relative Accuracy, the default method for specifying accuracy, relates the level of detail in a model to its overall size. This reflects the general trend that smaller models require tighter tolerances for manufacturing, and vice versa. Page 10-63 of the Release 18.0 Part Modeling User's Guide states "In general, you should set the accuracy to a value less than the ratio of the length of the smallest edge on the part to the length of the largest side of a box that would contain the part." Stated in equation form:
F = a factor based on part geometry s = smallest distance which the system will consider entities to be separate d = diagonal of box whose sides are parallel to default coordinate system axes and which just encloses the part The F factor adjusts this equation to more accurately describe how the Pro/ENGINEER application code describes model geometry. It is determined by part geometry and its value is always less than or equal to 10. In the simplest case of a part consisting of a rectangular protrusion and simple extruded cuts, the value is about 10. In general, however, this factor should be considered to have a value of about 3 or less. The diagonal of the part only increases in size. For example, if the model is cut in half, the diagonal value used does not become smaller. To illustrate the meaning of this equation, consider the following example: With a part in the shape of a sheetmetal plate which has a largest diagonal 10 inches long and has accuracy set to the default 0.0012, the smallest edge which is still discernible is about 1/10 * 0.0012 * 10.0 in. = .0012 inch. If the part accuracy is changed to 0.0001, the smallest edge can be about .0001 in. If the largest part diagonal is 1 inch and accuracy is default, then the smallest edge can be about 0.00012 in. With more complicated geometry, these values represent the lower bound of what distances can be discerned on the model. In general, the smallest distances would be about 3 times these values. Keep in mind that this smallest distance changes as features are added due to the change in the part diagonal. Please note that the relation above is an approximation used to represent Pro/ENGINEER's mathematical solution for describing model geometry. Also, be aware that decreasing the value of accuracy typically results in both an increase in regeneration time as well as increased file size and memory usage. Generally, as more computation is required to calculate geometry, more space is required to store the additional information. Absolute Accuracy Absolute accuracy can be specified for the model as an alternative to the relative accuracy method. Specifying absolute accuracy is roughly equivalent to directly setting the value for s (smallest distance) in the equation above, regardless of the size of the model. In general, relative accuracy should be used unless there are reasons to switch to absolute accuracy. Absolute accuracy should only be used when using imported features or when the absolute accuracy of two parts must be matched, such as during an assembly Merge operation. Setting the absolute accuracy of two parts to be the same will guarantee the geometry of both parts are looked at with the same accuracy. A typical example of when to use absolute accuracy is when a small part is being merged onto a relatively large part. The small part may have some tiny geometry on it which is acceptable at the small part's relative accuracy. When this tiny geometry is merged into the large part, the large part's accuracy will not be sufficient for the new relatively extremely small geometry. By setting the absolute accuracy of the large part to match the small part, the large part will evaluate the extremely small geometry with the same accuracy as the small part. Responses To Questions Now that you have some information on what Accuracy is, here is a response to your question: "Q1. Would you ever change Setup > Accuracy? If yes, when and why?" The Part Modeling User's Guide states the following examples which may require a change in accuracy:
2. intersecting (through merge or cutout) two parts of very different size 3. matching accuracy of imported geometry to its destination part 4. designing parts for manufacturing and mold design Absolute Accuracy Use in Pro/MOLDESIGN and Pro/CASTING One of the few areas of Pro/ENGINEER where absolute accuracy should be set as a rule is in Pro/MOLDESIGN and Pro/CASTING, where the mold assembly is typically on the order of 3 or more times the size of the design model/reference part. When small edges on the design model are transferred onto the much larger workpiece assembly, the same relative accuracy will in general not be sufficient. Absolute accuracy should be set for the mold assembly using Select Part and selecting the reference part. All components in the assembly, with the exception of the workpiece part, should have the same absolute accuracy setting. The workpiece is an exception because no features are created on the workpiece, the surfaces and edges of the workpiece do not intersect other parts, and since workpieces are intended to not be modified so they can be used in multiple mold models. Edge Display The current setting of part accuracy has a direct effect on the display of edges in a model. In order to create a graphical display of a model, Pro/ENGINEER samples a certain number of locations along the length of an edge and connects these points with straight line segments. Decreasing the value of part accuracy affects this display by first creating a more accurate representation of the part geometry, then increasing the number of sampling points along the edges. In a vast majority of cases, however, there is no need to increase accuracy for this purpose. The config.pro option EDGE_DISPLAY_QUALITY more efficiently affects the display of model geometry. By changing its value from Normal to High or Very High, the number of sampling points is effectively doubled or tripled, respectively. This config.pro option does not increase part size or regeneration time. Mass Properties When performing a mass properties calculation, the user is prompted to enter the relative accuracy. This mass properties relative accuracy should not be confused with model relative accuracy. mass properties accuracy affects the iterations of calculations used to determine mass properties. These iterations are based on the underlying model geometry which is affected by model accuracy. To improve mass properties results, an attempt should be made to first improve mass properties accuracy, then model accuracy only if necessary. Feature Creation In certain situations, adjusting accuracy will enable features to be successfully created. This typically occurs when the feature creates or intersects very detailed or tiny geometry. In general, however, it is better to save modifying accuracy as a last step after attempting other techniques for creating features, such as changing the regeneration order of the features on the part, changing the types of features on the part, or modifying/redefining features to avoid extremely detailed or tiny geometry. The decision to modify accuracy or not should be based on design intent. In some cases a small edge may be desired, so accuracy should be changed. If a small edge is not necessary, attempts should be made to avoid the edge from being created. Responses to Answers Received by Other Users Here are responses to the answers you received from other users:
Response:
Response:
For drawing cross sections, Pro/ENGINEER internally creates a cut on the model. If this cut intersects an edge so that only a tiny portion of the edge is left to one side, then this issue will occur, and may be corrected with accuracy. In part or assembly mode, Pro/ENGINEER internally determines the intersection of the cross section cutting plane with the model in the same manner as creating a Datum Curve using the Intr Surfs option. If this intersection does not form a closed loop of entities, the x-section may not be able to be created. This may occur, for example, when an edge almost, but does not quite, coincide with the cutting plane. These are two examples of the types of geometry issues which may occur from the intersection of the x-section with the model. In general, it is best to attempt to modify/redefine the model or the x-section to correct this issue before attempting to modify accuracy due to the increase in regeneration and file size which generally occurs when decreasing the value of accuracy.
a) adjust order of rounds at the corner. b) adjust accuracy. Response:
Response:
Response:
6. User: some people do use the smallest accuracy as their default. Response:
Response:
8. User: on R18, there is an option to set accuracy to an absolute number, such as .0001 inches (instead of a ratio of smallest feature to part diagonal of .0001, the kind of accuracy on R17). Response:
Response:
STL file generation is based on the need for refinement of the triangulation of the model to a point which is an balance between time for STL generation and an accurate representation of the model geometry. As a rule, parts with small details and/or high curvature surfaces are going to require smaller chord heights to accurately triangulate the geometry, but generally the minimum chord height is sufficient. 10. User: increasing accuracy can make a part change size, by .06" in 30", if there is no overall constraint on part size. Response:
Response:
Response:
13. User: parts with many boundary surfaces require an accuracy of .0001, gaps in surfaces appear if using default .0012 . Response:
Response:
Response:
Response:
17. User: use absolute accuracy option on merged parts Response:
"Q3. Is there any problem with letting individual designers select whatever values they prefer for Accuracy, default_dec_places, and sketcher_dec_places?"
Response:
2. User: no problem with decimal places varying, but accuracy should be the same for all parts for a master model merge. Parts to be merged must have the same accuracy. Response:
Best regards, John Kudzma Senior Customer Support Engineer
|