Docstoc

CMM Quarterly Spring 2008

Document Sample
CMM Quarterly Spring 2008 Powered By Docstoc
					cmm Quarterly

Spring 2008
Volume 2 Issue 2

#1 CMM & VISION PROGRAMMERS E-ZINE

Laser Scanning Brings In-Process Inspection to Moldmaking p. 24
In This Issue
Correct Part Rotation for ‘Perfect’ Vectors Torq Smooth vs Aluminum Parts Using Charts to Define Data Application of Scanning Technologies DMIS Corner Laser Scanning Brings In-Process Inspection to Moldmaking

From the Editor
Well, here it is the spring of 2008 and this is the second issue of four for the year 2008. We’ve gone through some changes. We switched publishing software just before the release of the first issue this year and this caused some problems. In the rush some things were missed. We will attempt to do better next time. Some changes have included getting a web site up and archiving the past articles. You will also be sent a link that will take you to the website as opposed to a .pdf file if you wish. PDF attachments sometimes don’t make it through firewalls and can make delivery difficult. The entire magazine is dedicated to CMM / Vision programmers. My intent here is to provide indepth articles to aid other programmers; I would like to receive articles from other programmers who have had different experiences. There have already been some great articles from Richard Clark, Stanley Schnuerer, and Paul Jackson. Sometimes putting a sentence or two in an open forum doesn’t get the entire message across. Here you can delve into the topic more with an article. Every month I am sent press releases from various OEMs and I have decided to include some in the magazine. They will be clearly marked with an Orange banner with the company name to indicate their source. This will not take over for any of the informative content in articles provided by programmers and operators of CMMs and Vision equipment but will bring some useful information on what’s going on in the industry. I am adding a “Meet People in the Industry” section. This will allow you to meet different people who make our little world turn. If you have anyone you would like others to meet drop me an email with their email address and I will attempt to contact them for an interview. I am excited to announce a new section entitled DMIS corner. This is brought to you by Bailey Squier and his associates to better explain and bring to us the working of DMIS. Look for great articles on exploring the workings of the DMIS code. This issue is on Tips and Tricks.

I am still working out how best to handle the employment section. I’m a little concerned that companies wouldn’t want this readily available to their employees when this is being sent into their facility. Look for this on the website in the future. I welcome any comments or ideas that you have that would make CMM Quarterly more informative. Please send any correspondence to info@cmmquarterly.com Mark Boucher CMM Quarterly

What they are saying about CMM Quarterly
“This is the first issue of CMM Quarterly that I have seen. I thought it was excellent … very close to magnificent… “ Bailey Squier “I was impressed with the content. You had quite a bit of info packed in there.” David Rossow “I received this CMM Quarterly issue today. It was good to hear from you again. The CMM Quarterly looks to be a great idea.” Jim Wessling “Thanks for putting this out, this is a great publication. I really appreciate it.” Matt Haverstock CMM Quarterly is reaching the world. With distribution throughout the United States, China, Poland, Pakistan, Australia, New Zealand, Sweden, Germany, U.K. and expanding everyday.

Correct Part Rotation Can Allow For “Perfect” Vectors
By Richard Clark
“Think of the drive of the axis. If the probe is moving with the drive (positive direction), the vector is one and if the probe is moving opposite the drive (minus direction) the vector is minus one” In the previous 2 documents concerning Direction Vectors, techniques were discussed to calculate vectors for circular measurements and/or features neither parallel nor perpendicular to an axis. Understanding these concepts is vital to creating stable measurements from the programs used with your CMM, however, a more elementary approach Step #1 – We have a known angle of 210° that may be the preferred one. rotates from the X-axis. In our first example we can review by using the “90 minus rule” to calculate the direction vector when probing a point located on a feature that is neither parallel nor perpendicular to the X or Y-axes. We can determine from our print specifications that a 210° approach angle would be perpendicular to the surface being measured. Step #2 – Using the “90 minus rule” we can determine the unknown angle to be -120° and since we are in the XY plane the angle rotation is from the Y-axis. Step #3 – Calculate the I-J-K vectors using the Cosine of the angles. I = -0.86603 (The Cosine of our 210° X angle) J = -0.50000 (The Cosine of our 120° Y angle) K= 0 (Our probe is not moving along the Z axis)

In my short history of writing CMM part programs I have found that vector points are usually needed to construct a plane or line feature on a part. Most DCC CMM software has autoThis is clearly illustrated when the Part Coordinate measure plane and auto-measure line macros built System is viewed using Polar coordinates. within the software, which various geometric data

about the feature can be entered and the software and CMM take off and measure the feature. If your software has this you should do 2 things: First, get very comfortable using it, because it will be very beneficial. Second, get comfortable not using it because if you haven’t ran across an application, part, or fixture where this won’t work, you are very lucky. The only sure thing I know about luck is that it always runs out. To illustrate how easy this process can become we will use the earlier example part, rotate about an axis and create the “perfect” vectors needed to probe points for a line measurement. The first step is to rotate the part coordinate system in such a way that our line feature is “true” with an axis line. Since the part angle rotation from the X-axis is -60° and the Y-axis is 30°, we can rotate about the Z-axis (-60°), which “clocks” our part and aligns the feature with the X-axis.

(positive direction), the vector is 1 and if the probe is moving opposite the drive (minus direction) the vector is minus 1.”

For our probe to contact our line at a correct vector we use: I = 0.0000 J = -1.0000 K = 0.0000 We do not want the probe to move along the X or Z-axes but we do want it to move in a Y minus direction. J is the Y vector, so it equals -1. So here’s how it works (at combat speed). We program a movement or series of movements to place our probe on the Y positive side of our line at the desired clearance from the part and the desired Z-axis elevation. We used the comp point, go meas, meas direction feature, which basically tells the probe to move until contact.

Now our part is “true” to the X axis. Here is where it gets really cool. This is where the angle rotation from the axis, and 90 minus rule, need MEAS/CPOINT,F(CPT_1),1,AXDIR MEAS_DIR/I-J-K,0.000,-1.000,0.000 not be used. A direction vector that is needed to probe a feature aligned with an axis will always be After the point is taken we can use a CMM goto calculated by the cosine of 90°, which equals 1. movement (relative) to move the probe a certain distance only along the X-axis. Mr. Jerry Guffy, CMM software trainer from Mitutoyo, told me a rule of thumb that I’ll always GOTO/INCR,CART,3.00000,0.00000,0.00000 use and never forget. “Think of the drive of the axis. If the probe is moving with the drive Now we can copy and paste the 3 lines and change

Advertisement

the designation of the point label (red text) to probe the desired number of points along the line. We’ll use 3 for simplicity only.
MEAS/CPOINT,F(CPT_1),1,AXDIR MEAS_DIR/I-J-K,0.000,-1.000,0.000 GOTO/INCR,CART,3.00000,0.00000,0.00000 MEAS/CPOINT,F(CPT_2),1,AXDIR MEAS_DIR/I-J-K,0.000,-1.000,0.000 GOTO/INCR,CART,3.00000,0.00000,0.00000 MEAS/CPOINT,F(CPT_3),1,AXDIR MEAS_DIR/I-J-K,0.000,-1.000,0.000 GOTO/INCR,CART,3.00000,0.00000,0.00000

Now we can construct the line from the 3 data points.
MEAS/LINE,F(LINE2),3 CONSTPT/FA(CPT_1) CONSTPT/FA(CPT_2) CONSTPT/FA(CPT_3) ENDMES

The goal of today’s industrial facility is to exceed their customers’expectations for quality in the products they produce, and the foundation for meeting that goal is the measurements (or control of the measurements) system used to determine the status of the quality. “Exposing the Myths...” examines what national and international standards actually say, and don’t say, about the application and control of inspection, measurement, and test equipment. In easy-to-follow detail, it discusses how to build a strong system that meets both the letter (and the intent) of these standards by designing, developing, implementing, and maintaining an inspection, measurement, and test equipment control system as required to achieve and sustain ISO, QS, or TS certification. To obtain more information or to own a copy (at a discounted cost) today contact Richard Clark via e-mail or cell phone: rcmetrology@yahoo.com 260-251-2557

And that’s all there is to it. Read the print carefully and rotate your part correctly. All of your vectors will be as easy as counting to one. Richard Clark works as a Metrology Consultant and CMM operator in Portland Indiana, to receive a freeware version of his “Vector Direction Calculator 4.02” e-mail feedback to rcmetrology@yahoo.com Learn More about Richard Clark in the “Meet People in the Industry’ section

Training Department
Coordinate Systems
What is a Coordinate System?
One of the first things that you must do in creating a part program is to create a coordinate system on the part. This locates all subsequent measurements back to this location. How do I know where to create this coordinate system on my part? The location of the coordinate system needs to be the same as on the blueprint. The blueprint will indicate where the engineer wants the measurements to be taken from. This is usually determined by how the part is assembled or its relationship to the other parts in the assembly. A coordinate system is the ‘you are here’ indicator on a map and how far it is to the next feature is the linear dimension. Linear dimensioning is the method of locating a feature. In order to know the exact number to this feature you must have a start point and that is the coordinate system.

the ‘6 degrees of Freedom’. This was covered in detail in a previous CMM Quarterly; however, in order to cover the coordinate system thoroughly we must discuss some of the aspects of constrains in this article. Some CMM software use different terminology but the basics are the same. First, you must create a spatial orientation, spatial rotation, or a base plane. You must use a plane or any feature type that has a 3d axial line through it to create this base plane. This will constrain the first 3 degrees of freedom. This limits the part in its movement on the CMM and makes the part normal to CMM. You would use a plane, cylinder, cone, or a sphere. These 3d features have a theoretical line that runs through the feature and is normal to that feature; it is the axis line of the feature. Your CMM software will align that 3d line normal to a CMM axis.

How Do I Establish Create A Coordinate System?
In order to create a coordinate system you must constrain

be used but we will not cover that in this article), a plane, a cylinder, etc…. You may use any feature that is a line or has a line running through it. Here we probed a line on the Y axis and used this line for our axial alignment. This squares the part to the CMM Y axis. We also set the X origin on this line. This constrains 2 degrees of freedom. We can no longer rotate around the Z axis and can not move back and forth in the X axis. We now have only one degree of freedom left. We can still move the part back and forth in the Y axis.

In this picture a plane was probed on the top surface and used to create the spatial orientation. This forces the part orientation normal to the Z axis of the CMM. We also set the Z origin to this plane. As you can see it prohibits the part from moving up and down in Z but it can still move side to side in X and Y and rotate around the Z axis. This has constrained 3 degrees of freedom. 1) Movement up and down in Z. 2) Rotation around X. 3) Rotation around Y. The next step is to use an axis alignment. This will make one edge of the part normal to a CMM axis. You may use either the X or the Y axis to align the part. The terminology in your system might be Axis Alignment, Planar Rotation, or some other term but the process is the same. You may now use a point or a line on the remaining side of the part and set the Y origin to that feature. This fully constrains the part. We now have a coordinate system on the part. The coordinate system indicator shows the location of the origin of the part. This should match the starting location using the same features on your blueprint. What if I use a CAD based system? The alignment routine is exactly the same. You would just use the model system to generate the features for your program. This type of alignment is known as a 3, 2, 1 alignment, indicating the method of constraining the 6 degrees of freedom.

Myths And Truths About Establishing Coordinate Systems
I’ve been told I can set the part on the CMM anyway I want to. True with some exceptions. If you set your part at some skewed angle to the CMM axis’ the software will still align it mathematically to be normal to its own X, Y, Z axis’. As you have seen above the software will square the part to be normal to the CMM. So

What features can I use to create the axial alignment? You can use a line, 2 circles that are aligned (offset circles can

yes, this statement has some truth to it but some care must be given to how the part is to be probed. If you set your part on a compound angle and you are probing manually or with a joystick you may encounter some probing error. Each probe point should be probed normal to the feature. If a point ‘slides’ or glances off the surface at an angle that is not normal to the surface then you will have some cosine error. I have to use this 3, 2, 1 alignment every time. False. This is by far the most common alignment you will come across but is not the only alignment. For parts that have free form surfaces and no defined prismatic features you will need to use an iterative alignment. This is accomplished by probing several points on the surfaces of the part and letting the software mathematically align the part to the nominal values of the points. This may take several iterations, each time getting closer and closer to the nominal value. I can measure my alignment manually and run the program in DCC. True. Well, you can certainly do this but the best course of action would be adding a DCC alignment. After you measure your part manually you should include a repeat of the alignment in DCC making sure you are driving normal to the feature. This will better align the part and measure the alignment

in the same measurement parameters that the remaining features will be measured with. If I want to quickly check a feature I don’t need an alignment. False. Let’s take a gage ring as an example. If you wanted to check the diameter you could lay it on the CMM and probe the diameter and probably get a good reading because the top and bottom surfaces of a gage ring are typically ground. At a minimum you should probe the top surface and use this as your spatial orientation before checking the diameter. This will square the top to the CMM axis and allow you to check the diameter normal to the top plane.

Let CMM Quarterly’s Graphic Arts Dept. design a graphic ad for your product.
contact info@cmmquarterly.com

Using Charts to Define Data
The Meaning of Data Part II
By Mark Boucher CMM Quarterly

In the last issue of CMM Quarterly we discussed the meaning of data. I have data but what does it

mean and how can I convey this data to the customer or machinist to give them the information they need. This article will explore some basic concepts of data and data collection. At Chiron America, the machine tool builder in Charlotte, NC, they use a system that works very well to bridge the gap between CMM inspection data and the machinist. Along with the main division in Germany, ChironWerke, they have developed a system on how to best use the data coming off the CMM. SPC Statistical Process Control, or SPC, is a valuable tool when used correctly. SPC in some companies means Sort Parts Continually, until you make the data read what you what. It has had its abuses but when used properly it can be a great asset. In this article we will look at a process and show the basics of SPC and CMM data. This is only a glimpse and will be very basic in its scope but you will get the idea. Process Whether you realize it or not each part coming off a machine center has a process. The things that go into making that part are the building blocks of that process. If one of these things breaks down or needs attention then the whole process is affected and it will directly affect the quality of that part or a series of parts coming off that same machine. Chiron America uses a Zeiss CMM with Calypso software. Each inspection result is written directly to the company’s internal server both as a PDF, for permanent record storage but also a .chr and .tab file which can be pulled directly into Excel. This allows Chiron to be totally paperless. A SPC program has been developed in Excel to give each machinist the ability to analyze the data for their machine based on the CMM results.

One aspect of the SPC program is the ability to graph run charts which give a telling picture of the actual machine process. This is what we will examine in this article. Terms To Be Familiar With CP = How much of the specification limits are being used. CPK = How close your process is to the target value. You might a good CP but have a low CPK meaning the process meets the specified limits but is far from the targeted value. True Position Charts Let’s look at some charts and see what data is revealed in them. In figure 1 we see a true position run chart. At first glance we see the data is at the top of the chart above the UTOL, or upper tolerance limit. Look at the bottom left corner of the chart and you will see the CPK is a -3.26. These parts are out of tolerance but this chart reveals something else. Look at the range between the parts, it is 0.045 microns.

This is acceptable and within the tolerance of this process so while these individual parts will be scrapped the process can be targeted. However this one chart can’t reveal what exactly needs to be done other than the obvious that one of the axis that make up the true position needs to be better targeted.

Figure 2 shows the X axis that make up the above true position. The data is distributed along the median of the tolerance. In this chart that is the MW line. Both the CP and CPK look very good for this individual axis. This feature is fine and we can move to the next axis.

This is the Z axis and it shows the axis is out of tolerance but capable. Look at the CP this shows we have a good process but we need to target Z better toward the mean and it will bring in the true position. These charts were created with a machinist in mind so the KOR, or correction value, is shown on the far lower right of the graph. If the machinist moves Z in the negative direction 0.311 he will be at the mean value for this feature. This will correct the CPK value also. Out of Process Figure 4 shows a process that is out of control. Obviously we would want to re-clean and recheck parts 21, 22 and 24 to make sure we have the proper data.

If we get the same results then this process needs serious attention. With a CP of 0.87 we can not simply make a targeting move but look at the last four parts. They seem to have an acceptable range. This creates some questions. Did the process settle down? Was the tooling inserts replaced prior to machine and now have settled in? Is the part moving in the fixture? Is there sufficient clamping pressure in the fixture? Do we need to run a larger sample to get the proper overall view of this process? With so many variables involved in the machining process, such as fixtures, tooling, coolant consistency, hydraulic pressures, etc.. it will take time to get the answers to these questions. As a Quality Inspector you may not be involved in the individual tasks of finding the solution but you need to be equipped with the tools to give a reasonable answer as to what data is coming off your CMM. Conclusion Starting out with the right SPC program is key in giving you an overall view of your process. There are plenty of programs out there and you need to look at reporting functions of all of them to rightly discern which one will work for your situation. This is just one tool you can use to improve the quality of the parts coming across your CMM. Mark Boucher is the publisher of CMM Quarterly and can be reached at info@cmmquarterly.com

Advertisement

NewCastle Measurement Your Offline Solution

DMIS 5.1 Complient XP, Vista Ready Model Transparency

www.newcastlemeasurement.com

This is the easiest off line software to use on the market today. All the features are available to use, such as, Model Transparency, Smart probe, Clearance Plane programming and much more. For more information contact sales@newcastlemeasurement.com

Announcing 2 new websites:
CMM Quarterly announces the creation of two new websites that will be CMM and Vision focused.

www.cmmquarterly.com

CMMQuarterly.com will bring you the current CMM Quarterly in a digital format and will also contain past articles along with a Training section. New material will be added between the release of the quarterly magazine from time to time. So check back often to get the latest in CMM and Vision articles. www.cmmsite.com

CMMSite.com is a site dedicated to the CMM and Vision Community at large. Here you will be able to look for Employment, search the Classifieds, view Events and Tradeshows, and view a CMM / Vision Directory. The CMM / Vision Directory is a one stop “yellow pages” of the CMM and Vision industry. If you are a business, or a contractor that would like to get the word out about your business use our industry specific website to market your business.

These sites will be available on May 1st.

Meet People In The Industry
This is a chance to meet people who like us work in the metrology industry. In each issue we will try and introduce you to different people that you may encounter. If you know someone we would be interested in meeting send their information to: info@cmmquarterly.com

See Richard Clark’s article on Direction Vectors.

What mentors, if any, did you have? That would be a very long list....Dave Schwab and Mike Dukehart from Mitutoyo are at the top of my list along with Scott Beavers You’ve written one book, are there any more in the future? I have an outline for a “What all CMM operators with their bosses knew about CMMs...” type of book, but so far I have not found a publisher who wants to sign on for it. But no matter what... I’ll always be a writer....

How long have you been in the CMM field? A total of 12 years How did you get your start working with CMMs? When I was put in charge of gage calibration at my full time facility What CMM or Vision products do you currently use? Mitutoyo BRT 707 CMM with GEOmeasure 6000 also some vision with Metronics QC-4000 What is your educational background? High School Graduate and some Leadership Schools from the 82nd Airborne Division Who do you currently work for? A Tier 1 automotive manufacturer in Indiana What made you decide to start RC Metrology? I felt others could benefit from my experience. I think before you learn how to do something right, you learn how not to do it many times over. What services does RC Metrology provide? Consulting and Training: Calibration MSA (Gage R&R) CMM concepts Shop-floor measurement

Torq~Smooth vs. Aluminum Parts Inc.
TorqSmooth Transmission Co. and Aluminum Parts Inc. are disputing a quality rejection from the first production run of transmission AWD transfer case housing parts made for TorqSmooth’s new medium duty CVT transmission. Receiving inspection at TorqSmooth rejected 3000 housings claiming that 24% of the parts are out-of-specification for the position of a fastener clearance hole. They spot-checked a 30 piece sample discovered the discrepancy then verified it checking a larger sample. TorqSmooth’s quality control analyst predicted process capability for position of the mounting flange hole at 0.24 Ppk far below the minimum levels of Ppk required for submission 1.67 initial and 1.33 continuous. The PPAP submission from Aluminum Parts stated that the position specification was verified on every 10th piece with “functional gaging” and records showed that there were zero non-conformances in the sampled pieces from the production run. TorqSmooth rejected the parts and informed Aluminum Parts Inc. to retrieve and promptly replace the rejected parts. Aluminum Parts, confident of their quality measures, decided to re-check the same 300 piece sample on-site at TorqSmooth with their production attribute gaging. To the amazement of TorqSmooth all pieces passed the gauge. Does this story sound familiar? Have you ever wondered why the go-position-gages that you purchase for your processes seem to consistently pass the product that variables data and SPC analysis predict is defective? The difference is not in the integrity of the measurement techniques, attribute vs. variables data, rather it results from comparing product variation to different limits. Geometric position tolerance go-gages generate discreet data (passfail) by verifying feature positions to their worst case physical boundary considering the limits for size and position simultaneously while variables gages segregate the size and position data and compare the statistical distributions of each to their respective specification limits separately. The specification for the hole, n8.9-9.4 [j|n0.36m|A|B|C] includes the tolerance modifier m which stands for “Maximum Material Condition.” It means that when the hole size is at its maximum material condition or smallest hole size Ø 8.9, its geometric tolerance is the minimum specified amount Ø 0.36. When the hole size increases the geometric position tolerance increases respectively because a larger hole can be a little further off location than a smaller hole and still pass the fastener through to its threaded hole location. When the hole is at its largest size Ø 9.4 its geometric tolerance increases to the maximum Ø 0.86 which equals Ø9.4 USLSize - Ø 8.9 LSLSize + Ø 0.36 USLPosition.

Aluminum Parts’ attribute gage has an Ø 8.54 pin which is equal to the “virtual condition” or inner boundary of the Ø 8.9 hole (Ø 8.9 MMC, LSLSize minus Ø0.36 USLPosition.). It is designed so that an Ø 8.9 hole that is off location by 0.36 will still fit on the gage. The average measured diameter of the holes was approximately 9.13mm so the average variable position tolerance of the sample verified on the attribute gage was approximately Ø 0.59 (USLP Ø0.36 + ¯XS Ø9.13 – MMCS Ø8.9). Of course each hole had a unique value for position tolerance according to its size but every one of the parts fit the gage. The receiving inspector at TorqSmooth checked the size and position of the sample parts with a coordinate measuring machine, presented the data in separate histograms for size and position, and calculated the process capability ratios from the data. So why are process capability predictions from TorqSmooth so different from Aluminum Products attribute gage results?

production sample respectively with continuous data. Once the process is predictable, due solely to common cause variation, and is deemed ‘in-control’ an estimation of the process performance can be predicted from the period sample using the following equations:

These formulas apply to normal distributions that have constant specification limits. For bilateral tolerances the sample process capability (Ppk) equals its potential (Pp) when the distribution’s mean is centered relative to the USL and LSL. The process capability of a unilateral geometric tolerance on the other hand is derived from figuring the encroachment of the distribution on only one of the specified limits (either the USL or LSL). The process potential Pp has typically been considered irrelevant for unilateral tolerances because predicting the encroachment of the distribution on the boundary representing perfection, zero deviation, or infinity-the limit opposite the MAX or MIN limit, does not examine the probabilities of a defect.

The equation that TorqSmooth used to predict capability of the hole position treats the Upper Specification Limit as a constant value and it disregarded the variable “bonus” tolerance. Aluminum Products took full advantage of the variable tolerance by applying the variable “bonus” tolerance physically with the attribute gage. In order for the process capability predictions to be comparable both methods have to address the variable portion of tolerance in the prediction. Statistical indices such as Pp-Ppk and Cp-Cpk are commonly used to compare process variation to specification limits and predict the process potential and capability of a process period sample or a sequential

The relationship between feature size and variable tolerance Ø8.9 MMC = Ø0.36 MIN and Ø9.4 LMC = Ø0.86 MAX can be shown graphically by overlaying distributions for position deviation and feature size on the same histogram and aligning their associative limits. We can see that the “Virtual Condition” aligns with a 0 position tolerance, the MMC or LSL size 8.9 aligns with the MIN variable position tolerance 0.36 and and the LMC or USL size aligns with the MAX variable position tolerance 0.86 respectively.

The classic reliability distribution model for stress vs. strength Figure #1, mirrors that of the distribution parameters of a variable tolerance and exemplifies the method to include the variable tolerance in a process capability equation. The probability of failure is predicted by figuring the area of interference of the two distributions. The area of interference of two normal distributions for stress and strength is figured from the equation: where μI and σI are the mean and standard deviation of the stress, and μs and σs are the mean and standard deviation of the strength.

Typical process capability predictions are derived by comparing the portion of the area under the distribution’s normal curve that is beyond the USLP to that of the total area. The attribute gage, however, compares each parts total tolerance (specified + bonus) to its position error. There is a big difference in the predicted conformance to specification with TorqSmooth claiming 24% defective vs. Aluminum Parts claim of 0% defective. This analysis shows that the actual probability of a defect is probably closer to Aluminum Parts claim than that of TorqSmooth but it is certainly not 0%. The ZUpper value 0.713 of the tolerance regarded “as constant” translates to a defect rate of 23.79% therefore one would expect 0.2379*300 = 71.37 defects in the 300 piece sample. With the tolerance analyzed “as variable” the ZIntersection of 2.5 translates to a defect rate of 0.62% or 1.86 defects in 300. The fact that no defects were discovered in the 300 piece sample could be more a testament to the unreliability of attribute measurement predictions than variables measurement especially when there are extreme differences in the ratios of conforming to non-conforming product. Aluminum Parts may have felt exonerated of the claim that 24% of their parts were defective but the generally agreed acceptable quality level of 1.33 Cpk still appeared to be a remote possibility if continuous data was to be used to predict the process performance capability. TorqSmooth consequently could demand attribute inspection levels near 100% to insure that the minimum levels of capability are maintained.

Similarly, the probability of a defect for a variable geometric tolerance can be predicted by figuring the area of interference of two distributions, the geometric deviation and its related feature size. The Z values are easily converted to process capability indicator Ppku by dividing it by three. According to the continuous data collected by TorqSmooth the difference in predicted capability is:

You can see from the graph above that even with a confidence level of 50% the attribute sample size for a capability of 1.33 must be >10,000 so what could Aluminum Parts do to demonstrate that their

process is better than the minimum level of capability 1.33 Ppk without inspecting all pieces with attribute gages? They improved the process and used continuous data to predict process performance! Aluminum Parts decided that rather than using go gages to check every piece they would gear for up variables inspection, institute a control plan for process monitoring, and use the variable tolerance capability formulas to predict process performance. They also decided to monitor the X & Y coordinates of the position separately so that control variables would match the machine adjustable machine parameters. They observed the individual distributions for X & Y and saw that they were both normal and “in-control” but slightly “off target”. Scatter plots of the X & Y coordinate data helped visualize the position deviations. They adjusted each coordinate distribution by its mean deviation X (-.09) & Y (-0.06) and recomputed the position deviations to see the potential process capability with X&Y means on target.

either side of the target will produce the same positive position deviation. Since a centered deviation cluster has more coordinates closer to and on either side to the target it will appear skewed toward the target “zero.”

The histogram showed that the interference of the two distributions dropped dramatically just by centering the coordinate distributions on target and the variable tolerance capability formula now shows a predicted capability improvement from 0.89 to 1.39 Ppk but they wondered if it could be trusted since only the size distribution was normal the other, position, was nonnormal (skewed).

Aluminum Products immediately observed something interesting about the recomputed positions. What was interesting was that even though the shape and variation of the coordinate scatter plot remained unchanged the shape of the computed position deviation histogram changed from more normal to more skewed and its standard deviation dropped slightly. They realized that the more the coordinates are off target the more the computed position deviation distribution appears normal and the more on target the coordinates are the more skewed it appears. They discovered that it happens because the computed position deviation is always a positive variable that has a boundary value of zero. An equivalent deviation on

To estimate the error of the variable tolerance formula Aluminum Parts decided to replicate a very large sample of the skewed position distribution and variable tolerance and compare each instance in a MonteCarlo simulation to produce a pass/fail attibute just like the production gage. Then they would compare the predicted PPM defective from the Monte-Carlo simulation to the “assumed normal” continuous data prediction. From the Attribute Sample Size chart they determined for a capability of 1.39 and a confidence level of 95% that the sample size should be roughly 100K. So 100K rows of random normally distributed data were generated for X and Y each at its observed standard deviation with a mean value of zero. Like-

wise random data was generated for the feature size at its observed values for its mean and stadard deviation. Then each row’s position deviation was computed 2*(X2+Y2)1/2 creating the skewed position distribution and it was compared to the variable position tolerance to produce an attribute pass/fail statistic. In the simulation 73 of 100K failed (had position deviations greater than the variable tolerance). From the following Ppk vs PPM Defective graph we can see that a normally distributed unilateral distribution with a capability ratio of 1.39 Ppk should yield a defect ratio less than 32 PPM but the Monte-Carlo simulation yielded an estimated 730 PPM defective which is approximately 1.06 Ppk.

Aluminum Parts still had a problem however, even if they adjusted their coordinate distributions to the target and used the new formula to predict their capability the monte-carlo simulation showed that they could only achieve an estimated 1.06 Ppk where TorqSmooth requires 1.67 at initial production authorization “PPAP” and 1.33 there-after for continuous production. Taking a fresh look at the entire problem Aluminum Parts realized that the location of feature size distribution within its limits directly controls the amount of variable tolerance in the position specification. As the size approaches its MMC limit the variable tolerance is reduced and as it approaches the opposite limit the variable tolerance increases.

The underestimate of defects may seem substantial 32 PPM vs. 730 PPM (32-730)/1000000 = -0.07% but when compared to the current practice of ignoring the variable portion of tolerance altogether it is significantly better. It compares to a 4.86% overestimate of defects that would occur if the variable tolerance was ignored. If this same improved “coordinate centered skewed position deviation distribution” was analyzed according to current robust data analysis practices for a non-normal distribution, a Box-Cox transformation would conclude a Ppk of 0.55 and the error relative to the simulated attribute gage results would be (49350730)/1000000 = 4.86%.

Hole size is the one parameter that Aluminum Parts has little concern over. The Process Potential is greater than 3.0 Pp and the process is well centered in compliance with process improvement guidelines. If Aluminum Parts wanted to increase the variable tolerance for position all they would need to do is increase the average size of the hole but how much? Too little and the position capability would be more vulnerable to USL defects and too much and the size would be more vulnerable to USL defects. Aluminum Parts decided to target the feature size so that the predicted defects would be minimized for both size and position simul-

taneously. They figured that if they set the equations for ZUpper Pos and Zupper Size equal to each other and solved for the feature size that it would give them equal probabilities for a defect.

Applications of Scanning Technologies
By Ray Xing
Fifteen years ago when we talked about “scanning” or “digitizing” on cmm, it sounds very fancy to most of cmm people. And not too many cmm guys had a chance to work on a cmm equipped with scanning probes even though the software that time has the scanning module already, for example, most Mitutoyo cmms were running Geopak with Scanpak as scanning module. But technology moves fast, today most cmm guys are familiar with terms like “reverse engineering”, which most likely involves scanning with an SP600 or SP25M probe, or even REVO probe. There’re so many cool things about scanning that I even see articles years ago talking about the total replacement of conventional triggering probes being a final solution. I myself am a scanning fan too. And I worked with SP600, SP25M and Vast gold xxt on cmms running pc-dmis, camio studio, calypso and umess with kum. Even though the software interface looks quite different, the way they define scanning path is very similar: start point, end point, direction point and increment, etc… What I like to share is some examples in which, if scanning method is properly employed, you can make your inspection process 3 times faster. Take a look at this part:

To figure the optimum feature size you have to convert the mean position deviation to its equivalent feature size reference (0.18 +8.54) = 8.72

Convert the optimum feature size back to its equivalent mean variable tolerance reference (9.254 - 8.54) = 0.714 and then plug the values into the capabiliy formulas

By monitoring and controlling the individual X&Yposition coordinates in production and by targeting the feature size to its optimum level Aluminum parts was able to increase the capability ratio to 1.81 and decrease the predicted defects to 27 parts/per/billion. Now Aluminum Parts is faced with another challenge, that is, convincing TorqSmooth that even though the feature size is targeted above its nominal size sacrificing its potential capability and the hole positions, with a typical capability analysis, appear to be out-ofspec far too frequently, they can trust that the probability of finding a defect in this process is extremely remote. Paul F. Jackson Retired Product Development Engineer, Ford Motor Company spcandgdtman@yahoo.com

collect points so you can capture the true profile of the whole parts and benefits are: 1, Minimize cmm occupancy and it’s so headache free when scanning, you don’t have to worry about any alignment, rotation, constructions, etc… 2, Data is saved permanently. You can go back to double check anytime. 3, When challenged, you have full confidence to defend with pictures which are worth thousands of words. 4, It’s so easy to compare two parts or compare same part a year later when you make 2 scans different color and overlay them in cad. This list can still go on and so is the scanning technology. Ray Xing
Dimensional Measurement Specialist STAVELEY Services Canada Ray.Xing@staveleycan.com

It’s a flat gauge and of course there’re lot of dimensions like diameter, radius, angle, distance, etc… I believe it’s not a difficult job to most cmm guys at all: create an alignment according to the drawing and then call up circles, lines, constructing intersection points… Now take a look at the two dimensions in the enlarged view: R1.264 and 0.713. How will you check them? Let’s try, for the R1.264 maybe you can just probe a few points to get a circle, but because you really have very little tiny area to probe how can you trust the result? I guess many people know a better method: using polar points from theoretical center, not too bad (actually the topic of “small arc, large radius” is my another one). But to get 0.713 is very difficult because the top is not something you can probe but you have to construct by intersecting a line with the circle you just did. You only have +/- .001” tolerance so if your circle is not accurate enough then you won’t get a good result. With scanning method, this job becomes really easy: because it’s flat, you don’t even need create your alignment, just make sure you scan (it doesn’t matter with analog or triggering probe) your –B- and –C- (the scanning path set up for this part should be really easy in any software) then output your scan to iges file. This should take no more than 30 minutes (you can tell it’s a relatively small part). The rest of work are all in your cad station, all you need is something like autocad (MDT), cadkey or mastercam. If you’re pretty good in cad then creating circles, lines, intersections off this spline is not a big job at all. Especially for those 2 dimensions, you can zoom in 100 times and construct a circle that fits your scan really good then you’ll get a very accurate reading for 0.713. And if someone challenges you on this dimension just print a nice graphical report to show him. In this particular case, you use your cmm as a tool to

Have your ad seen here.

CMM Quarterly is an industry specific magazine reaching the CMM and Vision markets throughout the U.S., Asia, and Europe.
For more info contact: info@cmmquarterly.com

DMIS Corner
DMIS Tips & Tricks -- by Bailey H. Squier

As the first Tip & Trick, we need to understand how to fix a program if its broken. To any programmer, one of the most frustrating problems is sorting out something that is ambiguous in the language being used, or discovering something that is “wrong” with the language with no method to get it fixed. The usual procedure is to contact the company that developed the software, report the problem, and request a fix. This process can take a very long time, and may never get done if it is not high on that company’s priority list. Here we run head-on into the issue of standards vs. proprietary solutions, which is a crucial topic, but one we must leave for another day. Because the Dimensional Measurement Interface Standard (DMIS) is a standard available through both the American National Standards Institute (ANSI) and the International Organization for Standards (ISO), one can be assured of the opportunity to participate in the maintenance and enhancement of the standard, and to suggest “fixes” that will resolve individual company and general industry problems. The DMSC, Inc. is an ANSI accredited Standards developing organization, and is also recognized by ISO as an A-Liaison to TC184/SC1, which is an ISO committee that develops standards for Physical Device Control. The DMIS Standards Committee (DSC), a standards sub-committee of the DMSC, Inc., does the actual work of maintaining DMIS. To provide the widest possible opportunity for input from those directly and materially affected by the standard, the DMSC, Inc. provides a method for the public to interact through a process called a Standards Improvement Request (SIR). Anyone who can access the world-wide-web can view current SIRs and can submit a new SIR via the internet at http://www.dmisstandard.com/sirs/ sirlist.cgi Most corporations rely on standards as the basis for their quality systems infrastructure; thus, when they develop robust processes today, these processes will still work well in the future, using software that is compliant with the standard. Additionally, many corporations see the value of direct involvement in the development of these standards by participation or

membership in the standards developing organization, providing them a voice on the direction that these standards should move. The DMSC, Inc. is one such organization that provides the opportunity to participate in broad discussions with others in the metrology industry concerning the path forward for today’s technology and the metrology industry in general. In future articles, actual “how to” techniques to solve real-world measurement problems will be discussed. But, for now: The TIP: Participate in the continued development of the DMIS standard by submitting SIRs for technology enhancements. The TRICK: Attend DMSC meetings and give support to your industry needs, and help direct other emerging metrology standards. (Information about the DMSC, Inc. can be found at www.dmsc-inc.org ) Advetisement

DMIS Programming Aids Step by Step DMIS Programming was written by a professional DMIS programmer and is a guide to creating Dimensional Measuring Interface Standard (DMIS) programs for Co-ordinate Measuring Machines (CMM). The book covers the most important aspects of DMIS in a task oriented fashion while providing examples and illustrations throughout. This book is the perfect companion to the DMIS standard for anyone interested in mastering this programming language. CMMTS DMIS Editor is a windows based text editor designed around the DMIS programming language. This editor will help users create and edit code while away from the measurement system quickly and easily. Features include: code auto completion, menu and icon based code creation, auto refresh and save, etc. The editor also includes a specialized calculator with trigonometry and vector function areas. These products can be purchased separately (Step by Step DMIS Programming $169.00*, CMMTS DMIS Editor $495.00*) or together as a DMIS Programmers Kit ($499.00*). Please visit www.dmsc-inc.org Online Store to purchase or for further information.

Laser Scanning Brings In-Process Inspection to Moldmaking
By Jim Clark, Metris USA
Until recently,in-processinspectionhasn’tbeenverypracticalin mold production, according to Bill Berry, president and chief executive officer, Die-Tech and Engineering Inc. (Wyoming, MI). “You’re not mass producing molds, so you can’t use the measurements as feedback for the next product,” he explains. Like any toolmaker, the best he and his staff could do was to measure the product after the fact and rework any deviations that might fall outside the specified tolerances. Quality control remained a post-production cost, rather than a value enhancement. Berry didn’t give up, though. He remained committed to finding a way to implement in-process inspection. It was a principle of good manufacturing—one of principles that he learned while working for one of the Big Three automakers and that had made his 40-employee company successful at producing tools for plastic injection molding and aluminum and zinc die-casting. He has introduced as many of these techniques as he could, as well as the advanced technology to support them, since establishing the business back in 1984.

His steadfastness on in-process measurement was rewarded when Berry came across the LC50 laser scanner and the supporting Focus suite of software from Metris North America Inc. (Rochester Hills, MI). The technology gave his quality-control experts the ability to scan the contours, sharp corners, and deep, narrow pockets on the molds made by his company in a fraction of the time that it would take a coordinate measuring machine (CMM) fitted with a conventional scanning touch-probe. Finally! He had found the cost-effective method that he needed to measure these features and elicit the appropriate feedback for correcting the process and avoiding costly and time-consuming rework. CMMs for Inspecting Cavities? Despite his willingness to invest in high technology, Berry had never considered CMMs for process control and resisted buying one much longer than others in his position did. “It would have been purely an expense with no real pay off— just a quality control cost with no benefit other than record keeping or score keeping,” he says. “Without laser scanning, there was really no cost effective way to apply CMMs.” Because he could not justify the capital outlay, he sent his molds to measurement service whenever a customer required him to validate molds on a CMM. The problem with conventional CMM technology is that the touch probes and supporting software were simply too slow and cumbersome for getting regular process feedback from molds. Even if the probes could reach into narrow recesses and sharp corners with radii as small as 1 mm, gathering the large amount of datum points with conventional touch probes, even scanning ones, would require a significant programming effort. Then, interpreting the mountain of data collected by the CMM would be time-consuming and tedious. After all that, scans compiled by probe would contain blind spots between touches, allowing anomalies there to go unnoticed. Moreover, measuring the molds after machining contributed nothing to in-process measurement. It was still inspect after the fact. Measuring the 10 to 50 electrodes typically used by the EDMs to finish the cavities was a slightly different story with the same ending. It would solve the problem of access because they are reverse images of the holes, cavities, ribs, and contours and, so, have no walls or other obstructions around them. It also would provide a kind of in-process feedback mechanism because any corrections to the electrodes occur before actually making the finishing cuts on the cavity. The problem of collecting and processing millions of points and thousands of surfaces quickly and economically would

Typical Die-Tech aluminum body side die-casting. “This industry has become more of a science than a craft,” Berry observes. “Even though we were a small company, we started employing the most advanced production and computer technologies we could afford.” Consequently, a full complement of computer-aided design and manufacturing (CAD/CAM) systems, computer numerically controlled (CNC) machine tools, and automated electrical discharge machines (EDMs) move both digital data and physical product through the 25,000-ft2 facility, from conception to delivery.

remain unsolved, however. Laser Scanning is Rad The LC 50 laser scanning head and Focus software changed the situation radically, prompting Berry to invest in a CMM equipped with the Metris technology. With help from Focus Scan, the head can collect 19,200 datum points per second along any three-dimensional surface in its line of sight. Focus Inspection then compares the cloud of points to the CAD model and generates a color-coded map of deviations from nominal. In a few days, quality control can provide production a visual analysis of an entire surface. “Laser scanning changes the economics by gathering vastly more data in an extremely compressed time frame,” says Berry. Die-Tech’s CMM uses a Metris LC50 laser scanning head to measure an electrode that an EDM in the shop will use to finish an injection mold for producing housing pivots for windshield wipers. The turnaround is fast enough for the shop to use this feedback to produce perfect electrodes between the roughing and finishing processes while the roughed block is offsite at a heat-treatment facility for hardening. Meanwhile, the CAD department designs the electrodes, the machining department cuts them, and the quality control department scans them. “By inserting laser scanning into the process, we can use the electrode measurements and setup information to perform a kind of simulation of the process,” says Berry. This simulation using Focus Inspection validates both the electrodes and the setup instructions for the EDMs, the two primary sources of variation in electroerosion. The CAD technicians use the setup data to put the electrode into the specified position and orientation and then compare the electrode to the corresponding parts of the CAD model of the mold. The color-coded map of the comparison leads machinists to any minute interference and clearance caused by an imprecisely machined or positioned electrode for correction before any cutting takes place on the mold cavity. By the time the roughed mold block returns from heat-treating, the electrodes are within tolerance and the setup instructions are ready. Feedback from the simulation has neutralized the primary sources of variation, and the electroerosion process itself is extremely accurate and repeatable. “So we have reduced the dimensional errors in the production of our molds to the tolerances of our equipment, which normally exceed the requirements of customers,” reports Berry. A Good Investment

Consequently, EDM operators can have the confidence to run all of their jobs untended and as fast as allowable. “In the past, there were certain complex EDM operations that they hesitated from doing without being there because they were critical and had to be just right,” explains Berry. Now, the operators have the confidence not only to run their machines untended, either overnight or while they tend other machines during the day, but also at full speed. Berry estimates that the confidence created by Metris’ software squeezes about 10% from the lead-time, helping Die-Tech to keep its three to four-week lead-times well below the industry’s more typical five to six weeks. As far as he is concerned, any scanning on the tool to prove that it adheres to the customers’ specifications is free. “They might ask for 16 or so points for verification,” he says. “We can give them a million, proving overwhelmingly that the tool is incredibly accurate.” And now, that capability figures prominently in the business plan. “We don’t send anything out the door that we expect to come back for anything other than customer-paid engineering changes,” says Berry. “We haven’t budgeted for a mold’s coming back to make things right.” He also has confidence— the confidence to eliminate that line from the budget. He knows that laser scanning has transformed measuring on the CMM from a cost into a value-added process that shortens lead-time, trims cost, and guarantees quality.

Chris Berry, the owner’s eldest son, uses Focus Inspection to check an electrode for finishing an injection mold for producing housing pivots for windshield wipers. He is comparing a cloud of measurement points collected by the laser scanner to the original CAD file.

Employment
Contact info for employment: Joseph Mackiewicz mank5@comcast.net Location Southeastern Michigan and Mid-West for work Terms of Employment Contract programmer to full time Pay Structure Hourly or Salary Software MeasureMax software Contact info for employment: Steven Majetich n8pvb@yahoo.com Location Youngstown,Ohio and Western Pennsylvania region Terms of Employment Looking for Full time work Pay Structure Hourly or Salary Software Val-Meas 200 on HP-200 machines. LK-DMIS 5.22, CAMIO 5.22

Introducing a new website

This site will be available on May 1st, offering a Classified section, Employment, Events and Tradeshow listings, and a CMM / Vision industry business directory. So look forward to this site and the CMM Quarterly website at the beginning of May.

NEWS BULLETIN F A R O Te c h n o l o g i e s I n c .

FARO

1 2 5 Te c h n o l o g y P l a c e L a k e M a r y, F L 3 2 7 4 6 THE MEASURE OF SUCCESS
Darin Sahler, Global PR Manager Darin.Sahler@faro.com, 407-333-9911

FARO’s All-New Software – CAM2 Q – Speeds, Simplifies Measurement
January 21, 2007, Lake Mary, FL – FARO Technologies,

engine that creates customized reports in various formats, including Microsoft Word and Excel, HTML, and PDF. • Efficiency: CAM2 Q can eliminate the need to move the device when measuring large parts, because multiple users can now measure with multiple FaroArms and FARO Laser Trackers simultaneously – even on the same location. Also, every measured point can be viewed, deleted, or remeasured. Users don’t have to start over if new features are added to a prototype. • Simplicity: CAM2 Q simplifies setup, training, usability and programming. Whereas other software can involve lengthy setup procedures, Q can be installed and ready to measure with FARO hardware right out of the box. It’s also the easiest measurement program to learn because the software and training were developed simultaneously. “We focused on learning, as opposed to just training,” Carvalho said. “Rather than just showing how the features work and expecting users to figure out how to apply it, we teach them how to solve their specific metrology problems.” In addition, the new intuitive user interface reduces the learning curve and measurement time for inexperienced users. It also allows them to complete the job in the fewest number of mouse clicks possible. Finally, users can quickly create, edit and run programs in the same window rather than opening a new application or separate screen. Also, CAM2 Q allows programmers to plan and record measurement/operation sequences that the software then programs as easy-to-follow, on-screen instructions for other users. • Live Data: CAM2 Q eliminates the lag time between measuring and reporting that is common with other

Inc. (Nasdaq: FARO), the world leader in portable computer-aided measurement hardware and software, announced it will release its all-new metrology software: CAM2 Q. “Simply put, we’ve advanced the functionality and training while making it even more intuitive for any user,” FARO President and CEO Jay Freeland said. “Whether they’re a seasoned metrology pro or using our portable CMMs for the first time, CAM2 Q puts the user in control.” Continuing FARO’s tradition of having the fastest software on the market, CAM2 Q has been built from the ground up with a best-of-breed CAD translator and the Parasolid CAD engine. But according to Antonio Carvalho, FARO’s Director of Software Engineering, the most impressive facets of CAM2 Q are not features of the software, but are “the core design aspects and user benefits — freedom, efficiency and simplicity — in perfect balance, resulting in new software that’s in a class of its own.” • Freedom: CAM2 Q can easily change/convert between different measurement units (inches, mm, microns, angstroms, etc.) or coordinate systems (cylindrical, polar, Cartesian, etc.) at any time without re-measuring. There are no limits to the number of units the user can define. The all-new user interface is fully customizable, so users can create a work environment that best meets their needs. Place windows and toolbar layouts based on the priorities of the project – not a preset format – and the software saves multiple users’ choices. And unlike other measurement software, CAM2 Q can add, change or measure features at any time – in any order – so the project can flow logically as the task demands, without being restricted to a preset order of operation. It also has a new, versatile reporting

metrology software. When a feature is measured, users can instantly see exact dimensions and geometric properties. The easy-to-read reports allow production decisions to be made immediately. • Confidence: CAM2 Q incorporates market-leading CAD translators, as well as NIST- and PTB-certified fitting routines, which certify that its geometry calculations are accurate to 0.0005mm (0.00002”). That’s one half micron; 500 times smaller than the width of a human hair. “CAM2 Q’s unique combination of power and simplicity provides a confidence-inspiring user experience with FARO’s portable CMMs,” Freeland said. “No other software/hardware package in the industry simplifies measurement while providing so much capability. The end result for our customers: better products, more costeffective processes, and a competitive advantage.” CAM2 Q will be available on February 5, 2008. About FARO With more than 14,600 installations and 7,000 customers globally, FARO Technologies, Inc. designs, develops, and markets portable, computerized measurement devices and software used to create digital models – or to perform evaluations against an existing model – for anything requiring highly detailed 3-D measurements, including part and assembly inspection, factory planning and asset documentation, as well as specialized applications ranging from surveying, recreating accident sites and crime scenes to digitally preserving historical sites. FARO’s technology increases productivity by dramatically reducing the amount of on-site measuring time, and the various industry-specific software packages enable users to process and present their results quickly and more effectively. Principal products include the world’s best-selling portable measurement arm – the FaroArm; the world’s best-selling laser tracker – the FARO Laser Tracker; the FARO Laser ScanArm; FARO Laser Scanner LS; the FARO Gage, Gage-PLUS and PowerGAGE; and the CAM2 family of advanced CAD-based measurement and reporting software. FARO Technologies is ISO-9001 certified and ISO-17025 laboratory registered. Learn more at www.faro.com

News from Renishaw
For Immediate Release

REVO™ achieving great reductions in CMM scanning time for early adopters

Revolutionary ultra-high-speed scanning technology meets manufacturer needs for both faster throughput and more data points November 2007 — Renishaw’s revolutionary, ultra-high-speed scanning system for coordinate measuring machines is achieving tremendous productivity gains with early adopters. World-class manufacturers have retrofitted existing CMMs with the new Renscan5™ system and infinite positioning REVO™ head. They’ve applied the system to speeding inspection on some of their more complex and time-consuming parts. Following are results that they’ve realized, compared to previous practice: • Jet engine blisk — 922% throughput improvement. The Inspection sequence comprised 9 section scans of the airfoil profile, 8 longitudinal scans of the blade, 2 scans of the root profile and 1 scan of the annulus profile. Conventional 3- axis scanning 10 mm/s REVOTM scanning 500 mm/s 46 minutes (1 blade). 22 hrs. 14 min (all 29 blades)

4 min. 30 sec. (1 blade) 2 hrs. 10 min. 30 sec. (29 blades)

• Automotive cylinder head — 680% throughput improvement. The inspection sequence comprised 12 valve seats and 3 circular scans in each of 12 valve guide bores. 3-axis axis scanning 15 mm/s REVOTM scanning 400 mm/s for valve seats 50 mm/s for valve guide bores 29 min. 13 sec. 3 min. 42 sec.

• Automatic transmission valve body – 155% throughput improvement. The inspection sequence comprised twelve 6mmdiameter holes, six 5mm-diameter holes, 25 points on the gasket face, 45 spool bores, 3 shot scan on the gasket face and 6 points on the spool face profiles.

Three axis scanning 15 mm/s REVOTM scanning 500 mm/s

18 min. 5 sec. 7 min. 5 sec.

“These developments are revolutionizing measurement throughput,” states Denis Zayia. Renishaw’s CMM product manager. “Besides major reductions in cycle times, Renscan5 and REVO make it possible to obtain far greater data point coverage. REVO’s low-mass, low-inertia design allows ultra-high-speed data capture — up to 4000 points a second compared to 200-300 data points for conventional scanning.” The agility and point-taking productivity of the Rensca5/REVO dynamic measurement system is stimulating the creativity of metrology professionals, says Zayia. Helical scanning, for example, can greatly speed the measurement of bores, while generating thousands of data points to determine roundness, concentricity or taper. In comparative testing, he notes, Renscan5 reduced form measurement scanning of a cylinder bore from 90 seconds to just 2.5, enabling all cylinders of a V-8 engine block to be scanned in less time than a single cylinder by conventional methods. A leading automobile manufacturer, he notes, is applying the system’s massive data capture capabilities to measure engines for wear analysis and deviations of the cylindricity of each cylinder in its spatial orientation. Wear is portrayed in visual graphics similar to a thermal scan. “Faster inspection is especially vital on large, complex, highvalue parts with many critical features,” notes Zayia. “CMM inspection can be a major bottleneck to efforts to speed throughput and gain Lean efficiencies. Form measurement of complex parts and critical geometries for functional fits can demand many thousands of data points. Needing to produce and document parts to ever-higher precision, ever-tighter tolerances, manufacturers are looking to CMM speed for a solution.” As a practical matter, conventional three-axis CMMs scan at rates of 5 to 15 mm/sec in order to hold accuracy, notes Zayia. This is to avoid high acc/dec rates and rapid axis changes that can induce inertia errors, causing deterioration in measurement accuracy. “CMM inspection has been stuck in that time warp for over two decades,” he says. Created under Renishaw’s longest and largest development program, Renscan5 blows away the speed limits, he says. The Renscan5 enabling technology encompasses a range of breakthrough 5-axis scanning products that measure at up to 500 mm and 4000 data points per second, while virtually eliminating the measurement errors normally associated with existing three-axis scanning systems. A 3-D measuring device it is own right, the REVO head features two rotary axes – one in vertical plane, one in horizontal — to give infinite rotation and positioning capability, notes Zayia. The REVO

measuring head performs synchronized Y and Z axis motion to quickly follow changes in part geometry during inspection routines, he stresses, avoiding the dynamic errors caused when moving the larger mass of a CMM structure. Where X axis motion is required for the probing routine, he says, Renscan5 moves the CMM at a constant velocity along a constant vector as measurements are being taken, removing the acceleration/deceleration inertia errors incurred in conventional scanning. Contact Jeff Seliga or Renishaw Inc. 5277 Trillium Blvd. Hoffman Estates, IL 60192 Tel: 847.286.9953 Fax: 847.286.9974 jeffrey.seliga@renishaw.com

Roger Rude Kemble & Rude Communications, Inc. 4242 Airport Rd. Cincinnati, OH 45226 Tel: 513.871.4042 Fax: 513.871.4895 rrude@kemblerude.com

Wenzel
non-contact and touch trigger, point-to-point or scanning, and connects to a large number of worldwide measuring systems of different design and brands.” Wenzel GmbH is the 4th largest CMM builder in the world and manufacturers its CMMs intrinsically accurate, negating the need for the huge error compensation files to achieve quoted accuracies which have become the norm in past years from most CMM manufacturers. The black granite used on Wenzel CMMs is processed inhouse at their impressive manufacturing facilities in Germany, not imported from China prefinished…which is now the industry norm. Xspect Solutions, Inc. headquartered in Wixom, Michigan, and is a wholly owned subsidiary of Wenzel Gmbh of Germany. Xspect Solutions is the Number 3 supplier of CMMs in the North American metrology market including; new Wenzel CMMs, and is the world’s largest supplier of pre-owned CMM equipment. Xspect Solutions also supplies new Wenzel CMMs with its OpenDMIS software to the North American markets.

News Release
WENZEL GmbH LAUNCHES OpenDMIS™ CMM SOFTWARE IN THE ASIAN MARKET

Wenzel will market OpenDMIS along with the existing Wenzel Metromec CMM software.

As a part of Wenzel GmbH global marketing strategy, the OpenDMIS™ CMM software that was designed and developed by Xspect Solutions and became a Wenzel product when they acquired Xspect Solutions in the fall of 2007, will now be marketed in Asia as a complement to Wenzel’s own Metromec CMM software that is well-known for its powerful performance and native geometry CAD access. Frank Wenzel, Wenzel GmbH president, says, “We now have the ability to offer our Asian CMM customers the best of both worlds. OpenDMIS is the second most popular software in the United States and will be offered to Asian customers of Wenzel bridge-type CMMs and as an upgrade to existing CMM users. OpenDMIS will targert customers looking for unparalleled CMM usability and short training curves. For the other range of Wenzel CMMs, we will continue to provide the well-known Metromec CMM software that allows the user to measure geometry and free form surfaces in one software package. Metromec speaks more than 14 different languages and supports a wide range of probe systems including;

OpenDMIS software. Send all Inquiries To: Keith Mills, Xspect Solutions, Inc. 47000 Liberty Drive Wixom, Michigan 48393 Tel: (248) 295-4300 kmills@XspectSolutions.com


				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:730
posted:1/20/2010
language:English
pages:30
Description: This is the Spring 2008 issue of CMM Quarterly