Page 1 of 6 12345 ... LastLast
Results 1 to 15 of 88

Thread: Metrology for the cabinetmaker

  1. #1
    Join Date
    Aug 2013
    Location
    Princeton, NJ
    Posts
    7,254
    Blog Entries
    7

    Metrology for the cabinetmaker

    Wanting to spur a bit of discussion, I wrote up this post and I'm curious to see if it is of any use.



    Metrology for the Cabinetmaker – Introduction

    This post is intended as the beginning of a conversation on inspection equipment and measurement standards rather than a conclusive how-to. My equipment, techniques and process are in a continuous evolution. These techniques have been working effectively for me and so I have decided to share them, however the I do not run a metrology lab, I run a wood shop, so be aware that these techniques may be incorrect or incomplete. I introduce this only to encourage thought and spark further discussion.

    The reader may wonder as to why fine measurements make any difference at all in a wood shop. After all, wood moves with change in relative humidity and so one may conclude that minor gaps are inevitable. Wood movement is one of the greater challenges of woodworking and yet we can plan for it and build accordingly, engineering considerations for wood movement into our product.

    As example, looking at the typical mortise and tenon joint. Wood moves across the grain, so if the mortise were cut in the typical fashion it would shrink and swell across the width. At the point of production it may be .500″ wide, and months later it may be .510″ wide. The tenon, if produced from the same batch of material in the same grain orientation at the same time will have moved in the same fashion, so it too will have moved from .500″ wide to .510″ wide maintaining the fit between parts. Along the height of the board it will gap at the top and bottom of the mortise. This can be partially countered by lightly compressing the wood during the fitting process. The joint can draw-bored to retain it tight against the tenon shoulder and the joint can be fully shouldered joint to hide the potential gaps created at top and bottom of the tenon that are not solved by compression.

    Gaps are also created by errors in setup and cutout. These are due to a host of circumstances including layout error, machine setup error, and tool bit deflection. Repair of these unexpected circumstances often requires custom fitting, remaking of parts or other time consuming approaches to repair or otherwise alleviate this circumstance. To avoid minor gaps an in consistencies created by these errors the first step is to remove uncertainty where we can find it. This requires a level of precision counted in thousandths of an inch for our machinery setup. Good technique combined with accurate equipment helps to produce accurate results, as example a saw that cuts squarely can also be used to make accurate length parts as both sides of the part can be make the same length. Accuracy is an important component in efficiency of process and it begins with the checking tools themselves before they are used to inspect the equipment or work piece.

    Knowing that a square is actually square and that 12″ is actually 12″ is a worthwhile venture when the goal is efficiency. These questions become increasingly difficult to answer without the proper equipment and so it is for this reason that I feel the woodshop benefits from having a basic kit of highly accurate tools used as standards for inspection. This kit is commonly referred to as a metrology kit in the machine shop.

    Metrology is the scientific study of measurements with an aim to create a common understanding and agreement of units. This agreement allows the industrialized world to create products with confidence that parts from independent sources will interact properly. For this reason many have adopted the metric system, a system of standards based on realizable values, as their countries have industrialized. If not directly replacing their own systems many countries have standardized their systems of measure on the metric system. As example, the standard for a foot (Imperial) is exactly 30.48 cm, the shaku (Japan) is 30.3cm and the chi (China) is 32cm.

    Modern standards which create the basic building blocks of the metric system are derived from specific references which remain constant. As example, a meter is the length of the path travelled by light in a vacuum during a time interval of 1/299792458 of a second. These such references form the primary standards upon which international systems base their physical standards. In the US these standards are maintained by NIST. Quality measuring tools are calibrated to NIST standards and certain forms of work require periodically recalibrating tools to those standards.

    In the woodshop we can compare a part to the space it must consume, mark and cut, or we can measure the space and build the part using that measurement as a reference, neither operation requires an international standard. However, if our work involves using multiple tools, each reliant upon a consistent series of measurements then we are best served by forming a shop standard and that shop standard may as well be derived from the international standards.

    In a practical sense one need only compare two tape measures to realize that a lack of standards can complicate things quite rapidly. If you’re using a few different steel rules and a set of digital calipers then it’s best if they all agree with one another. This is further complicated when the machinery itself has built in gauges and those gauges are relied upon. Consistency in measuring devices is helpful in avoiding transfer errors.

    Below I will detail a few of the measuring devices which are useful in the workshop, both highly accurate tools and moderate or low accuracy tools. They’re all necessary at times and their varying degree of accuracy or simplicity each make them useful in different ways.

    Please note that in machine shop’s which utilize imperial measurements, the basic unit of measure is one thousandth of an inch (.001″), commonly referred to as a thousandth, or ‘a thou’. So, for example .500″ would be read out-loud as ‘five-hundred thousandths’. Finer than a thousandth is a ten-thousandth of an inch, or ‘a tenth’, this is confusing to many as a tenth is .100″, but the tenth as commonly referenced in a machine shop is one ‘tenth’ of a thousandth or .0001″. Still, it would be unusual to call out .5001″ as five-thousand and one tenths, instead it would be referred to as ‘a tenth over five hundred thou’ or similarly clear way of providing that it is not dead on .500″ but a tenth over.

    Surfaces

    In precise measuring, life begins with the surface plate. A surface plate is a true flat surface, often to a degree of precision measured in tenth thousandths of an inch (shop grade) or hundredth thousandths (Laboratory grade). These plates are calibrated by differential electronic level and are lapped flat, this plate arrived with the readout provided by the manufacturer. This surface is the basis for most comparative measuring so an accurate plate is a nice thing to have and fairly inexpensive.

    The surface plate, along with all of the other precision tools are covered when not in use.
    Click image for larger version. img_3772-1.jpg

    Immediately following the surface plate are the straight edges. The straight edge is a crucial piece of equipment in the workshop. This simple device allows the ability to compare surfaces to a known true reference. There are many types of straight edges, shown here is a beveled edge straight edge purpose built for checking plane soles.

    Taking little for granted, I have checked this straight edge for flatness and parallelism. The process is done by applying marking compound to the surface plate then taking an imprint of the edge. I followed that by taking a sweep over the straight edge with an indicator in search of errors.

    Straight edges must be supported along their length when they are stored, or they can be stored flat in a toolbox so long as they’re protected from other objects banging into them.
    Click image for larger version. img_3776.jpg

    Another type of straight edge, the camel back. Rather than being ground flat, these are sometimes scraped by hand. This one in particular scraped by a very talented operator, by hand. This device is accurate to a few tenth-thousandths of an inch.
    Click image for larger version. img_3747.jpg

    Specific to woodworking are winding sticks, wooden sticks used to define twist or ‘wind’ (think wind like a watch, not like the weather). These are used to give a basic understanding of surface defects. These are compared to the surface plate to ensure that they read accurately.
    Click image for larger version. img_3774.jpg

    Angularity

    The angles referenced most often in a typical woodshop are 90 degrees and 45 degrees. To check 90 degrees a tool know as an angle block can be used. These plates are best accurately scraped flat on their surfaces, square about their main faces and the edges square to the main faces and also flat. They can be used in conjunction with one another for comparing a square surface on three faces at one time.
    Click image for larger version. img_3816.jpg

    Next up is the master square, this square is a precision ground engineer’s square certified to a high degree of angular accuracy. This is an A grade Mitutoyo square, which agrees with the angle plates above. This is mainly used for comparison with other squares and occasionally brought to the work but it is never scribed against.
    Click image for larger version. img_3807.jpg

    A multitude of squares can be utilized in the workshop including various sizes and calibrations of try squares intended for checking work and creating layouts. Next is a smaller master square, made by Starrett, that is inspected accurate to .0001″ and it agrees with my angle plates.
    Click image for larger version. s5BocW8fTaO2CECGElKeyg-1294621666-1564330418144.jpg

    Finally, I have another precision square which is of a quality that can be scribed against and used for other similar tasks without much concern. It is highly accurate, but it is the one square in the shop that I will retune periodically to ensure it remains accurate.

    Angular measurements outside of 90 degrees, are read and transferred in variety of ways depending on the degree of precision required. Measurements can be provided by protractor, angle blocks, fixed angles or sine bars. The most basic measurement kit is shown here; a bevel gauge which is used to compare and transfer angles and a simple protractor. A high quality bevel gauge can be quite reliable and this example, made by Chris Vesper has proven highly reliable.
    Click image for larger version. img_3782.jpg

    For the rest of the article see Page 2
    Last edited by Lee Schierer; 08-06-2019 at 5:02 PM. Reason: copied to SMC with permission from Brian
    Bumbling forward into the unknown.

  2. #2
    I think with the advent of CNC is so many shops "milling" wood has taken on a completely new meaning.

    But my method of work for ww'ing is more related to consistency than accuracy.

    A few example:

    I'm not as interested in the absolute length of a board as I am that certain parts are exact same length. So my method of work is "measure once, cut X times" using a stop block.

    Or that an angle is not necessarily exactly a particular degree, but rather close within a range of error, and the matching angle is dialed in to be complementary.

    Or while flatness is crucial for a lapping plate, its not the same for a table top.

    That said, precision is necessary for accurate machine set up.

    Which got me to wondering - is there a difference between accuracy and precision so I looked them up.

    Accuracy: "the degree to which the result of a measurement, calculation, or specification conforms to the correct value or a standard".

    Precision: "the quality, condition, or fact of being exact and accurate"

    Well that sure cleared things up LOL.

  3. #3
    Join Date
    Sep 2016
    Location
    Modesto, CA, USA
    Posts
    9,876
    To my mind accuracy is not really needed. What we need is repeatability. Tests have shown the harbor freight digital calipers are actually very accurate. Most of them are good enough for metal working and all are good enough for wood where 0.0001 accuracy is not needed since 0.001 is plenty good enough for wood.

    Accuracy: "the degree to which the result of a measurement, calculation, or specification conforms to the correct value or a standard".

    Precision: "the quality, condition, or fact of being exact and accurate"

    Well that sure cleared things up LOL.[/QUOTE]

  4. #4
    Join Date
    Aug 2013
    Location
    Princeton, NJ
    Posts
    7,254
    Blog Entries
    7
    Quote Originally Posted by Robert Engel View Post
    I think with the advent of CNC is so many shops "milling" wood has taken on a completely new meaning.

    But my method of work for ww'ing is more related to consistency than accuracy.

    A few example:

    I'm not as interested in the absolute length of a board as I am that certain parts are exact same length. So my method of work is "measure once, cut X times" using a stop block.

    Or that an angle is not necessarily exactly a particular degree, but rather close within a range of error, and the matching angle is dialed in to be complementary.

    Or while flatness is crucial for a lapping plate, its not the same for a table top.

    That said, precision is necessary for accurate machine set up.

    Which got me to wondering - is there a difference between accuracy and precision so I looked them up.

    Accuracy: "the degree to which the result of a measurement, calculation, or specification conforms to the correct value or a standard".

    Precision: "the quality, condition, or fact of being exact and accurate"

    Well that sure cleared things up LOL.
    Thanks for the detailed reply. I believe the mathematical description of precision is along the lines of your description of consistency. Ideally both have practical limitations depending on their application and material and both have application.

    As I understand it, accuracy can be practically defined as to how close your reading is to the actual number. What 2" is actually and how it compares to what you define as 2" and how close they are is accuracy.

    Precision is how similar things are, so in a production run, I'm precise if all of my parts are within my range of tolerance, I'm imprecise if they fall out of that range.

    You're using good practice and I certainly agree with your approach. I'm suggesting to augment good practice with accurate measuring, because there is an expediency found in accuracy.

    Take your example of a saw cutting to a stop block. If the blade is not square, then the parts may be different sizes when you're finished, because as you flip the parts you're cutting a trapezoid rather than a rectangle. If you're doing it without paying very close attention to the error (good practice) than some parts will be trapezoids and some will be parallelograms.

    What I'm pursuant of is, is the ability to get the error small enough that it doesn't have a meaningful effect, such that when cutting large quantities of parts I can have the error small enough that I dont need to pay strict attention to orientation from very early on through to the end of the process. IE I want interchangeable parts in large quantity rather than specific parts, and I want to still retain 100% of the quality that is provided when each part is hand fitted.

    This is unlikely to ever occur in my work, but it's nice to greatly reduce the amount of hand fitting which is seriously time consuming and brings little value over cutting precisely the first time.
    Bumbling forward into the unknown.

  5. #5
    Join Date
    Aug 2013
    Location
    Princeton, NJ
    Posts
    7,254
    Blog Entries
    7
    Quote Originally Posted by Bill Dufour View Post
    To my mind accuracy is not really needed. What we need is repeatability. Tests have shown the harbor freight digital calipers are actually very accurate. Most of them are good enough for metal working and all are good enough for wood where 0.0001 accuracy is not needed since 0.001 is plenty good enough for wood.

    Accuracy: "the degree to which the result of a measurement, calculation, or specification conforms to the correct value or a standard".

    Precision: "the quality, condition, or fact of being exact and accurate"

    Well that sure cleared things up LOL.
    Personally, I prefer to buy good once. I have one set of inexpensive calipers, a few years old, floating around the garage and the head is beginning to die, it gives weird errors at random and so it's going to find it's way out the window one day. I've had a set of mitutoyo calipers for 20 years that still measure accurate.
    Last edited by Dennis Peacock; 08-12-2019 at 9:56 AM.
    Bumbling forward into the unknown.

  6. #6
    I'm going to go out on a limb, here, and say that I think you have a small preference for Mitutoyo instruments

    I assume you acquired some of those instruments used, and obviously people mark their tools, but it's still slightly disconcerting to see my somewhat-unusually-spelled last name on your comparator stand!

    I enjoyed the post very much, and am looking forward to the next installment! If I could take the liberty of trying to summarize your broader point, I think you're saying that: given the right tools and techniques, you can achieve hand-fitted-quality in less time, and gain the advantage of interchangeability and repeatability.

    Thanks for writing this up and sharing!

  7. #7
    Join Date
    Oct 2016
    Location
    Ogden, UT
    Posts
    1,659
    Blog Entries
    1
    Your measurement tools compete against professional high quality machine shops. I work in a high end machining environment and we use many of these tools. Some of our work requires .0001" discretion.

    In our case, if we want true accuracy we go into a CMM room. This room is kept at 70 F. For metal, the humidity isn't important as far as I'm aware. The part should be left overnight if it is really critical. All measurement tools stay in that room. I guess what I'm trying to say is that your measurements will change overnight with wood unless you are in a conditioned air location. I don't have conditioned air, but luckily live in a desert.

    I agree with your approach and I am investing in better measuring equipment over time. Hand fitting is a huge waste of time and I also don't find it acceptable practice to just "expect handwork". For my cut off saw, I use a Bosch miter saw. I take a machinist square and set it to the blade so that there is no light. However, the blade most definitely deflects. I think this must be the reason behind saws like the upcut saw. Big diameter blade and I'm assuming it is thicker too. Bigger Arbor probably with bigger motor shaft bearings, etc. It's interesting seeing the difference in accuracy between commercial and residential.

    A next step for me is to get a 3 phase converter to get into the commercial tooling space. I've also been eyeing knee mills for help with woodworking and fab'ing.

  8. #8
    Join Date
    Sep 2009
    Location
    Putney, Vermont
    Posts
    1,042
    There is much that can be discussed in trying to achieve accuracy in producing things that have many parts.
    While working with wood is not usually considered something that requires the extremes of accuracy that you are referring to in using tenth indicators and surface plates, gage blocks etc., there are certain areas of production that would benefit you in producing accurate parts in a production type atmosphere.

    Having accurate jigs and fixtures, accurate machine setup, blueprints of parts and assemblies that use geometrical tolerancing, so as to determine the correct datums to work off, of and the degree of accuracy needed in that particular aspect of that part are all important aspects of manufacturing something. Experience and technique are also big factors.

    One can only be as good, as what they have been exposed to, and learned from that exposure. That is why apprenticeships are so important in producing excellent craftsman. And why so many are drawn to places like this website.

    How far one wishes to delve into this area in wood working is questionable, but I guess interesting, if you are into that sort of thing.

    Mark Duginsky made a good point in his you tube video on machine setup in the wood shop. Basically it was to the point that a good straight edge, square,jigs and fixtures, and using paper as shims to achieve accurate adjustments in producing good parts was all one really needed to make accurate things in the wood shop.

  9. #9
    Join Date
    Jan 2017
    Location
    Marina del Rey, Ca
    Posts
    1,934
    So well done and excellent photos. Bravo!

    I especially like the sliding T-bevel and protractor.
    "Anything seems possible when you don't know what you're doing."

  10. #10
    Personally, such accuracy and precision is beyond my needs; but I'm a hobbyist woodworker with the luxury of time to sneak up on fits or creatively design around errors. I think the salient point in his whole article is seeking efficiency. Once you know your ability to measure is perfect, it becomes possible to design completely by math. Anything less requires you to add some element of human judgment to the equation, and possibly recalibration/recalculation along the way.

  11. #11
    Take your example of a saw cutting to a stop block. If the blade is not square, then the parts may be different sizes when you're finished, because as you flip the parts you're cutting a trapezoid rather than a rectangle. If you're doing it without paying very close attention to the error (good practice) than some parts will be trapezoids and some will be parallelograms.
    Yes, that is why I said it is different with one's machines. The need to be dialed in as accurately as possible.

    Thanks for the excellent distinction between "accuracy" and "precision".

    This precisely explains why cutting a whole bunch of parts consistently, but to an inaccurate size proves it all goes back to the dark matter (what used to be my gray matter).


  12. #12
    Join Date
    Mar 2003
    Location
    SE PA - Central Bucks County
    Posts
    65,679
    Excellent writing, Brian!
    --

    The most expensive tool is the one you buy "cheaply" and often...

  13. #13
    Im with Bill and Dan they locked it down out of the gate. The need for extreme precision either comes from ones internal neuroscies or the financial need for production and repeatability. For years I argued about people being too fussy with thousandths in woodworking. Until we move the shop into the CNC world. Now they mean everything. They mean a job going together smoothly or perhaps painfully, or perhaps not at all. The thousandths will always matter to the neurotic (which CNC will turn you into, trust me, I know). I believe there is still zero need for thousandths for the person making on a creative level. Those thousandths are dealt with in their hands, feel, fit, and can be on one side or the other and still come out with a phenomenal result.

    This chase for "metrology" in wood, to me, comes from a sycophancia with the youtube metal working world. The Tom Liptons, Stephan Gotswinters, Robin Renzettis, all of which I salivate for their next videos yet settle for lowly instagram posts. While they drive for millionths because its a passion, I dont think any of them would pursue it if there weren't a profitable income from it. They are producing parts to a spec'. Peter Stanton makes parts to a spec'. He makes a swivel hose replacement to a lesser spec though still higher than we are capable of because its done in his sleep.

    Woodworking requires this only if you need it to satisfy your neurosis OR to satisfy having a bunch of inconsistent parts be able to be assembled by, anyone.

    Either is fine.
    Last edited by Mark Bolton; 07-31-2019 at 5:48 PM.

  14. #14
    Join Date
    Sep 2016
    Location
    Modesto, CA, USA
    Posts
    9,876
    I rmember reading years ago that the human eye can look at a plane surface and tell if two parts, fitted together to make that plane, are off by 1/1000 of an inch. Of course hit it with a fine piece of sandpaper and that glue line is gone from sight.
    I find it is easier to provide slight setbacks so it is obvious the two surfaces are not supposed to be co-planer. Such as when legs meet a table apron I set the apron back 1/8 " If I get it wrong by 1/000 no one can tell. Same idea with a raised panel door I set the panel below flush or proud of flush by 1/4 -1/16"
    Bill D

  15. #15
    Join Date
    Jul 2007
    Location
    Inkerman, Ontario, Canada
    Posts
    1,389
    This is where you really get to know people.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •