Surface Normals

Issues When Programming From CAD When Using A CMM

By Mark Boucher, CMM Quarterly http://www.cmmquarterly.com/



There are several issues that arise when bringing a CAD model into your CAD based coordinate measuring machine (CMM) programming software. One of these issues has to do with surface normals (surface vectors). You can bring in a model and the entire model or a portion of that model is visible, but dark (Figure1), or certain sections are not visible at all. This problem arises from the surface vectors pointing in the opposite direction than your CAD system views them. You are looking at the back side of the surface. You must reverse the surface normal (Figure 2). If your software has this capability, you are looking for something similar to ‘reverse surface normals’. This will flip the surface so the front faces in the correct orientation for your software to view the surface.

All surfaces have a front and a back side; a CAD program must know which is which. How is this done? The model must somehow include information to specify the front of a surface. This is done by surface normals.

This is a line perpendicular to the front surface and beginning on that surface pointing away from the surface. Meaning it exists only on the side of the surface that is its front. The CAD system must have this information to shade the model properly. Those that
use a CAD system need this to drive the probe normal to the surface.

Direction vectors have been covered extensively by Richard Clark, his three part series was featured in CMM Quarterly (http://www.cmmquarterly.com/ ). Suffice it to say that these surface normals are what give you the direction vectors from CAD models when programming. If you do not use a CAD model to program then you must calculate the normal vector. Contact rcmetrology@yahoo.com for a Direction Vector Calculator.

When picking a feature off a CAD model the software will extract the normal vector from the CAD surface. As mentioned above you may have the ability to flip surface normals or you may have the ‘view surfaces from both sides’ option. Care must be given to this selection because the surface will be visible but the vector may point in the opposite direction you need to probe the part, your probing direction vector. Just know when viewing the vector after feature selection, that it is correct.

This article is copyrighted. Please contact Mark Boucher at info@cmmquarterly.com for permission to reprint.

Runout vs. True Position When Using a Coordinate Measuring Machine

When runout is checked on a surface plate the part is rotated under an indicator and the deviation between the highest point and the lowest point is the called runout. Typically this is related to a circular part with a shaft or a cylinder as the reference datum. As the part is rotated any deviation in the datum is taken into effect.

Runout on a CMM is static. The part is not turned but is measured by points or scanned and these points make up the best-fit circle that will be use to check to the datum. The datum is also made up of measured features that are fitted into an element.

This brings up a dilemma. Which tolerance strategy is better, runout or position? This is something you will need to clear up with your customer so that both parties clearly understand what the data off the coordinate measuring machine (CMM) is really giving you. In most cases, position is the best strategy to use. Using a positional tolerance will give you more accurate numbers when using a CMM. Using the runout tolerance is not incorrect but when factoring in the best-fitting functions of CMM calculations then the positional tolerance is best approach.


Mark Boucher, CMM Quarterly http://www.cmmquarterly.com/

CMM Calibration And These Tight Financial Times

By Mark Boucher, CMM Quarterly http://www.cmmquarterly.com/

With these pressing financial times may companies will be looking to cut costs. The Quality Control department will certainly be one of the areas that will be looked at. There have been some discussions about how often do we to have the coordinate measuring machine (CMM) calibrated. The standard for the industry has been yearly. Can this be pushed out to two years? Am I violating any customer specifications?

Yearly Calibrations

It has always been standard procedure to have the CMM calibrated yearly. This has been the accepted practice for good reason. One must look at how often your individual CMM is being used, the wear and tear, and the history of your CMM to properly determine if foregoing a calibration cycle is wise.

Before a decision is made about moving out the calibration cycle, it is highly recommended to check with your OEM to how this will affect the OEM warranty and how this will affect their standing behind the manufacturing specification of the CMM. Of course, this will be like checking with the oil company to see if the oil in a car really needs to be changed every 3,000 miles. Yearly calibration is a continuing revenue source for the OEMs but if you have already moved away from the OEM to a contracted 3rd party for calibrations then you have already addressed any concerns about the OEM warranty.

When looking at the CMM as a valued piece of your Quality department and your commitment to the customer to ensure you are producing a good product care must be given to your decision to push out the calibration. My personal recommendation is not to skip any calibration cycle unless the CMM is truly not being used on any regular basis. Even with some of the costs being demanded by the OEMs for calibration these days it is still important to have the preventative maintenance done yearly. Think about the cost associated with repairing a CMM that goes down during a production run. Trying to schedule a service repairman to get your CMM up and running in a timely manner can result in extra costs that may have been avoided if yearly maintenance had been done.

Important things to consider about the consequences of foregoing a yearly calibration:

How often do I use the CMM?
How will this affect my warranty?
How will this affect the OEMs standing behind the published accuracy specs?
Can I afford the cost and time associated with a ‘down’ machine?
Does my 3rd party calibration service offer preventative maintenance as part of their service?
Do I fully understand what my OEM offers when it comes to calibration?
How much money will I really be saving by pushing out the calibration cycle?
Will my customer be affected? Do they have any requirements that will prevent me from changing my calibration schedule?

3rd Party Calibration Services

When contracting a 3rd party calibration service it is important to establish trust in the contractor you choose. As with any contracting service I would recommend calling a few of their customers to get a good understanding how services that are offered where performed. Many of the contracted calibration services are former OEM service repairmen and are well qualified to perform calibrations and repair but it will be important to establish and understand exactly what services will be performed and what type of warranty comes with the service.

You can contact me at http://www.info@cmmquarterly.com/ if you would like a list of 3rd party calibration contractors.

This article is copyrighted. Please contact Mark Boucher at info@cmmquarterly.com for permission to reprint.

CAD Modeling The Basics

By Mark Boucher, CMM Quarterly

I want to cover some basics about CAD models that might help us understand what is happening with some model features when you program from a CAD model using your coordinate measuring machine. By understanding surfaces we can better evaluate any anomalies we may encounter when we import a model into our CAD base CMM software.
There are several model types and we will cover two of the most common ones you would come across today, solid models and surface models. To be more accurate, they are parametric models and freeform surface models.

Parametric Models

Parametric models are created from features that are defined by parameters, or dimensions. These dimensions can be changed and the feature moves with the change. Prior to parametric modeling if a change was made then the feature was recreated, extruded, trimmed, etc…, in the new position and the old feature was deleted. Parametric modeling maintains the relationship of part creation, assembly, to output of the blueprint and a change anywhere along that process will update the model at every level.

Parametric models are referred to as a solid model, as opposed to a wireframe model. A wireframe model is made up of lines that represent the part but have no surfaces on them and makes 3d viewing somewhat tedious.

Parametric modeling revolutionized the CAD industry and allowed more affordable CAD software to become available to anyone. You can now pick up parametric CAD programming software up at your local Best Buy right off the shelf.

Freeform Surface Models

The second model type I want to cover is the surface model. With surface models curves are used to define the surface area and surfaces are applied between the curves then they are trimmed and merged, to make a solid. The problem with this method is to make sure all the voids between the surfaces are filled in. The surface definition changes as the need requires. Let’s say, you have a plane that requires basically four lines to define the boundary of the plane. A chamfer merging into a radius requires a greater amount of defining to create this type of feature. While creating the surfaces you may end up with a small void as you try to fill in the feature. Point placement from your CMM program will be dependant on where it sees the plane boundaries and a void will not be inclusive in this plane so the boundaries are redefined not giving you a true representation of the surface. In parametric modeling these types of transitions are automatically resolved.Some CMM CAD based software have a ‘healing’ or ‘repair’ functionality that will mend some of these errors. It is always advisable to use healing when using this type of model. If your software does not include this functionality there are third party softwares that do the job for you.SurfacesCreation of surfaces begins with a spline, aka curve. Splines are single lines that make up the shape of the surface. Imagine points that make up the shape of your surface and spline will fit through these points. These splines are then used to create the surface through a method known as ‘swept’ (using the curves as a guide rail) or meshed (lofted) through. A ‘swept’ surface follows the shape of the curve line. If you had a helical curve the swept surface will follow that helical shape as it extrudes the surface.

If your engineering department does any sort of reverse engineering they will ask for a series of curve files that they can import into their CAD system. The curve files are then used to ‘mesh’ or ‘loft’ the surfaces. The density or frequency of the curve lines along the surface will depend on the complexity of the surface being scanned. For flat planes only several are needed but scanning the chamfer to radius transition we discussed before would require a greater amount of curves to define the feature.

Another method is direct creation of the surface with manipulation of the surface control points. Points are created along the curves that can be grabbed and drawn in any direction to create a new surface shape. This inherently will create new surfaces to fit the new configuration.

It is important to note that the majority of CMM software in the market today do not have true CAD functionality and thus do not have the ability to manipulate surfaces as described above but it is important to know what is happening during model creation.


This article is copyrighted. For permission to reprint this article please contact Mark Boucher at info@cmmquarterly.com