Ask the doctor

This page is for quick questions that I can answer for you without extensive research. If your question is general and non-proprietary in nature, I may post it here along with my answer, so others can benefit from it. That way this page evolves as a Q & A page.

Aliases and requests for anonymity are fine and will be respected, but I obviously need to know who you are.

You can either E-mail your question, or contact me by phone, fax or mail.

Q:

You stated that there is a growing need in the metrology consulting field.

How do you start? I would think the insurance expense would be large, in addition to all the equipment. I love to measure parts, work with CAD & CMM. I think it's a great occupation and I wish I could do it full time. What would my best course of action be?

I'm researching for a Tech Writing Course, trying to justify upgrading/reworking our 20yr old manual CMM. We're not a huge company but we do a considerable amount of our own sheet metal, plastics, & welding fabrication and could really USE a new machine. I've wondered about suggesting it to my employer (21yrs/16 in QA) that we purchase new CMM and do some outside consulting to offset the cost, but it seems so far fetched.

Any advice would be very helpful.

thanks for such a wonderful website.
Knowledge is everything. Thanks for sharing!

cindy walter

A:

Cindy,

Thank you for your nice words about my website - I am glad you find it helpful.

The consulting I do is more in the "how-to" area. I provide training seminars and I work as an assessor of accredited laboratories - I do not provide measurement services.

I used to manage the standards laboratory for a large diesel engine manufacturer, so I have been in the business of measuring parts and justifying equipment purchase.

I understand and share your love for measuring parts. I find it so fascinating to be able to measure things that are invisible to the naked eye.

Regarding your questions:

I think your justification would come in several areas:

  1. Higher accuracy. If you upgrade the machine with CNC control, you will be able to make the touches much more repeatable and accurate than you can do it manually.
  2. Better productivity. Running the machine under CNC control will allow you to measure parts much quicker than you can do it manually.
  3. More features. If your software is as old as the machine, then upgrading the software would take you from the equivalent of "Pong" to a modern videogame - there is a tremendous difference in what you can measure and how you can get it presented with graphics etc.

As for doing outside measurement work, I would start by talking to customers and suppliers and even competitors in your area and try to get a feel for whether there is a market. Once you have determined that, you can start looking at what you would have to charge to offset the cost of the machine and how many hours you would have to sell each month, etc.

You may have seen that I am publishing a newsletter. I am currently working on the next issue. If you send me your address, I will put you on my mailing list.

Best Regards,
Henrik S Nielsen

Q:

We presently send our measuring and test equipment (calipers, dial indicators, height gages etc.) out to a subcontractor for calibration, but we wish to do some of the calibration in-house in the future. In order to do this I need to document a procedure. Is there a reference I can go to that will give me a step-by-step method of physically performing in-house calibration for each specific instrument.

Thanks.

A:

Keith,

Unfotunately, I am not aware of a good source for "canned" calibration procedures for the type of gages that you mention. I also think it is slightly dangerous to go that route, if those procedures say something different than you are actually intending to do - auditors love to find that kind of discrepancies. I know, because I am one.

The best thing you can do is get out the standards for each type of gage, look through them and note all the requirements that are given in the standard.

Then decide for each requirement if it is something you are going to do for each calibration or just as an incoming inspection (flatness of the anvils of a micrometer is an example of what I would only do as an incoming inspection). From that you can write the step-by-step instructions, including what to measure and the tolerance for each requirement.

I know it is a little more work up front, but I believe it will pay of by giving you procedures that are describing exactly what you do and that you have much more ownership in than what canned procedures you may be able to find. One way of not making this task too daunting is to have a plan that says for example that you will generate one procedure each month, so you can gradually go from sending things out to doing them in-house. It will also help you avoid being swamped by all your gages suddenly going in-house.

I hope this helps.

Best regards,
Henrik

Q:

Is there a defined number that indicates whether the linearity is good or bad? Like the criteria used for Gage R&R (Less than 30%)?

Thanks for your help,
Paula Miller

A:

Ms. Miller,

I am not aware of any "official" source that gives you that information. In some work I did for my previous employer, Cummins Engine Co., I set up a study that evaluated bias and linearity together as they are very hard to define independently of one another. In the study we looked for the largest deviation from the true value within the range of interest.

Our criteria for that number - like the 10%/30% for GR&R - was that 5% was acceptable and 10% was marginal. The reason for the much tighter values is that a systematic off-set (bias) is much more detrimental to your process capability than added variation (repeatability) is.

While there is no 1:1 translation between the two, we calculated that the 5%/10% had about the same impact as the 10%/30% for R&R in typical cases, so we figured that these were reasonable limits to apply.

Using these limits on real life measuring processes has shown that this is often a much tougher criteria to live up to than the R&R criteria. It also showed that we had some gaging that was able to meet the R&R criteria, but still hurt us by making us mistarget manufacturing processes because of their bias.

I hope this helps. Best regards,
Henrik S. Nielsen

Q:

Hi, I'm trying to understand the concept of cylindricity, which was very recently introduced to me. I can only find reference to the symbol, but I've found no clear definition or the impact of cylindricity on a shaft or cylinder. Any information that could be provided would be greatly appreciated. Thanks.

A:

Dear Ms. Hollis,

You are very correct, there is no clear, official definition of what cylindricity is.

The conceptual definition is that you take two co-axial cylinders, one being fully inside the material, but just touching it and the other fully outside the material but just touching it. You then align the axis of these cylinders such that the difference in their radii is the least possible (this is a minimum zone fit). The difference in radii is then the cylindricity.

The easiest way to understand cylindricity in practical terms is as the radial distance from the highest peak to the lowest valley when you take the combined effects of roundness, straightness and taper into account.

There are a couple of things to be aware of and look out for when calling out and measuring cylindricity. There is nothing in this definition that tells you how many points to take and what filters to use, if you are using a digital instrument, such as a CMM. With different settings, you can easily double your value or cut it in half, so you have to keep track of your settings, if you want to compare values.

Another issue for CMMs is that they perform the fit slightly different from what I outline above. They calculate a best fit (least squares) cylinder through all the points taken and use the axis of that cylinder to calculate the cylindricity from. This gives a higher value than the correct fit, but it is generally not possible to put a percentage on the increase.

There is a draft standard on cylindricity on its way through the ISO system. It will be ISO 12180, but it is still only a draft (I chair the committe working on it). As far as I am aware, there is no effort under way in the ANSI/ASME system.

Please let me know, if I can be of further assistance.

Yours sincerely

Henrik S. Nielsen

Q:

I recently attended a seminar put on by JTI Systems called PMAP(Process Measurement Assurance Program). It focused on measurement uncertainties and how they can affect the final product. What the seminar failed to give me was a process flow that I could follow when setting up a program. I guess the seminar description mislead me.

I am looking for a flow chart, step-by-step process, or some sort of cookbook format that I can follow when setting up a calibration plan for a particular measurement tool. Is there something like this out there?

Here is one example I currently have: We have a process that dispenses a caulking on top of a circuit. This caulking is cured and then a height checker automatically measures the height of the bead. The resulting measurements are used to control the process by an automatic feedback loop. If a part exceeds an upper or lower limit it is scrapped. It is easy to see that the integrity of the measurements is very important, therefore we have a standard that calibrates this height checker. At the moment, this height checker is not calibrated on a regular scheduled basis. It has no calibration sticker. There is no calibration procedure. The standard also lacks this. I need to set up an entire calibration program that is traceable. I also need to make sure that the uncertainties in this entire process do not eat up a significant portion of my product specifications.

I believe that I could create some sort of process flowchart that would cover most of these needs, however I am afraid of two things: 1)would I miss something really important? 2)Is there already a process out there that spells these things out for me and I would be painfully re-inventing the wheel?

So again I ask...is there any well thought out process documented on how to set up a calibration/measurement assurance program?

Thanks for your time. Please forward this request to anyone who may be able to help me out.

Regards,

Matt Martel

A:

(answer by snail mail with enclosures)

Dear Matt,

Thanks for your question. Yes, there is a process like the one you describe. It is called the PUMA (Procedure for Uncertainty Management) method and it is outlined in the technical report:

ISO/TR 14253-2: Geometrical Product Specifications (GPS) - Inspection by measurement of workpieces and measuring instruments - Part 2: Guide to the estimation of uncertainty in GPS measurement, in calibration of measuring equipment and in product verification.

As opposed to the title, the process is very simple. I have included a diagram of it as well as a chart of distributions that are used in estimating uncertainty contributors.

Using this method you would decide on a target uncertainty, which is the maximum acceptable uncertainty for your measurement. Then you would set up an uncertainty budget for the measurement, where the uncertainty of the standard you are using is one of the contributors. When you then start tweaking the uncertainty budget to optimize the cost of the measurement while staying within the target uncertainty, you determine the maximum allowable uncertainty for the standard. This in turn becomes the target uncertainty for the calibration of the standard. In this fashion you can work your way backwards through your traceability chain and connect everything with uncertainty budgets. On a macro scale you can determine all the capability requirements for your company's calibration laboratory this way.

Allow me a shameless plug: I teach this method as part of the seminar "Managing Measurement Uncertainty" (I have conveniently included a description of the seminar as well). I would of course also be more than happy to assist you with this work on a contract basis, but I have the impression that you mostly looking to get a little pointer in the right direction to do it yourself. I hope this information does this for you. Finally I have included a copy of the first issue of my newsletter Metrology INSIGHT. One of the articles in it may also be of help to you.

Best regards, Henrik

Q:

Dr. Nielsen,

What is a fixed gauge? I am doing some basic research and have come across a reference that they are "expensive, but measure a single, unfiorm product at high rates of speed. For certain parts they are more cost-effective than CMMs". Can you refer me to a supplier or better yet a web site where I can see a picture and read some text.

Thank you,
James Suc

A:

James,

A fixed gage is usually a gage that is made to simulate the mating part that the workpiece it measures has to fit together with. If the workpiece fits the gage it passes the inspection, if not then it fails the inspection. If, for example, you are making wheels for a car, the fixed gage may simulate the hub you mount the wheel on with cylindrical pins instead of bolts, placed at the nominal locations of the bolts an with a diameter that assures that the bolts will fit through the holes in the wheel.

The fixed gage is fast, since it is usually a one-step operation to check a part. However it does not give variable data, that can be used for SPC or capability studies. Since it is custom made for each part number it is fairly expensive, but generally cheaper than a CMM. However, where the CMM can be reprogrammed for different parts and easily accomodates engineering changes, the fixed gage has to be either reworked or replaced, if the part it measures is changed.

Edmunds Gages in Farmington, CT and Nelson Precision in Cincinnati, OH are a couple of fixed gage manufacturers.

Thanks, Henrik

Q:

Thank you so much for your response. One last quick question: I also have visited the Marposs web site. I assume some of these gauges are "fixed" too. But some seem "flexible" in that they measure several dimensions but I'm not sure if they can produce variable data or if it's go/no-go. I saw a photo of a crankshaft measuring gauge that looked like it could be reconfigured (maybe not so easily!) to measure different shaft-like parts. Does this classify as a fixed or flexible gauge.

Thank you in advance. James Suc

A:

James,

The Marposs gages are generally like "Meccano" kits, where you can build gages to fit your needs from standard components, thus I would not classify them as fixed. Also, they tend to have sensors in them, which give you variable data. That is generally not considered a fixed gage. Most fixed gages measure several dimensions or relationships between several features, so measuring several dimensions is not a disqualifying criteria.

Best regards, Henrik

Q:

Hi!

I´m from México, I´m studying Mechaninal engeenering and I am trying to get some info about METROLOGY, is it a career, a grade, a specialty?

Thanks,
al588299

A:

Hi,

Metrology is the science of measurement, just like psychology is the science of the mind and meteorology is the science of predicting the weather.

It is a specialty within several engineering disciplines. Personally I am a mechanical engineer, so I specialize in dimensional measurements of workpieces.

Is it a career? Well, I try to make it a career. I have been working in the field for over 10 years and I find quite a high demand for this specialty, and it is increasing as tolerances get tighter and tighter.

You might want to visit: http://www.cpm.uncc.edu/metrology_lab/index.htm which is the home page of the metrology laboratory at University of North Carolina at Charlotte. There is some good information and pictures from their laboratory there.

Best regards,
Henrik

Q:

Henrik,
I am a quality engineer and am responsible for implementing and maintaining a complete gage control system. I use gage blocks for calibrating thickness gages, calipers, etc., but I am a little foggy on calculating and using things like accuracy, linearity, stability, bias, etc. I need to put in a system that is compliant with QS9000 as well as ISO9000. I need a good technique for calibration procedure and a good gage control software system to get this done. I have the Measurement Systems Analysis manual by AIAG, but it mainly deals with R&R.

Could you help me?
SG in TN

A:

Dear SG,

Here is a high-level checklist of the things you need in order to live up to the requirements of ISO/QS 9000.

  • High-level procedures describing how you run your gage quality system.
  •  
    • How do you buy gages?
    • Put them in service?
    • Calibrate and maintain them?
  • Documentation of the qualifications of the people performing the calibrations.
  •  
    • This should be your definition of "qualified"
  • A gage recall system, that will keep track of all your gages and recall them at appropriate intervals.
  •  
    • The system should preferably allow you to retain "as found" and "as left" readings.
  • Documentation that your outside calibration suppliers are either accredited, or that you have surveyed/audited them.
  •  
    • This is a big issue in the new QS9000 as you have noticed.
  • Procedures for calibrating all the types of gages you calibrate in-house.
  •  
    • What do you do?
    • What is the acceptance criteria?
    • What do you do, if a gage fails?
  • Procedures for when and how you are doing R&R, accuracy, linearity, stability, etc. studies. This is separate from calibration.
  •  
    • You can probably define a process where you do it annually and keep the auditor happy.
    • You will probably want a spreadsheet or other piece of software to crunch the numbers and then record the final results either in your gage recall software or another database (or you can do that part on paper).

I hope this helps as a checklist to plan your work from. If you have any further questions, or need further assistance, please do not hesitate to contact me.

Best regards,
Henrik