Skip to Content.
Sympa Menu

machinist - [machinist] Making Accurate Straight-Edges from Scratch by John A. Swensen

machinist@lists.ibiblio.org

Subject: Machinist

List archive

Chronological Thread  
  • From: Lawrence London <lfljvenaura@gmail.com>
  • To: machinist@lists.ibiblio.org
  • Subject: [machinist] Making Accurate Straight-Edges from Scratch by John A. Swensen
  • Date: Fri, 2 Jan 2015 00:02:23 -0500

http://www.practicalmachinist.com/vb/general/making-accurate-straight-edges-scratch-john-swensen-295849/

Thread: Making Accurate Straight-Edges from Scratch by John A. Swensen

  1. 12-14-2014, 10:06 PM #1
    LFLondon's Avatar
    LFLondon is online now Hot Rolled
    Join Date
    Dec 2007
    Location
    North Carolina
    Posts
    669

    Default Making Accurate Straight-Edges from Scratch by John A. Swensen

    Making Accurate Straight-Edges from Scratch by John A. Swensen
    https://home.comcast.net/~jaswensen/...ight_edge.html

    "I adapted the techniques for making straight-edges from those described for making accurate surfaces, as described in Wayne R. Moore's "Foundations of Mechanical Accuracy", Moore Special Tool Co., 1970. This fascinating book describes how reference tools and machines with accuracies on the order of millionths of an inch are built."

    "The fundamental principle used to create precision straight-edges is that the only way that three curves can match each other is if the curves are all straight lines. By repeatedly comparing the three developing straight-edges with each other and averaging their shapes where they don't match, the three developing straight-edges will become arbitrarily straight."

    "Why Three Straight-Edges Are Necessary "
    Jorgo and Antarctica like this.
  2. 12-14-2014, 10:18 PM #2
    dgfoster's Avatar
    dgfoster is online now Stainless
    Join Date
    Jun 2008
    Location
    Bellingham, WA
    Posts
    1,993

    Default

    Quote Originally Posted by LFLondon View Post
    Making Accurate Straight-Edges from Scratch by John A. Swensen
    https://home.comcast.net/~jaswensen/...ight_edge.html

    "I adapted the techniques for making straight-edges from those described for making accurate surfaces, as described in Wayne R. Moore's "Foundations of Mechanical Accuracy", Moore Special Tool Co., 1970. This fascinating book describes how reference tools and machines with accuracies on the order of millionths of an inch are built."

    "The fundamental principle used to create precision straight-edges is that the only way that three curves can match each other is if the curves are all straight lines. By repeatedly comparing the three developing straight-edges with each other and averaging their shapes where they don't match, the three developing straight-edges will become arbitrarily straight."

    "Why Three Straight-Edges Are Necessary "
    Is there more that you intended to post on this?

    Denis
  3. 12-14-2014, 10:45 PM #3
    LFLondon's Avatar
    LFLondon is online now Hot Rolled
    Join Date
    Dec 2007
    Location
    North Carolina
    Posts
    669

    Default

    Quote Originally Posted by dgfoster View Post
    Is there more that you intended to post on this?

    Denis
    No. The full article is at Swensen's web page and I am sure there is plenty more across the Web from other sources. I thought it might be useful to some.
  4. 12-14-2014, 11:27 PM #4
    dgfoster's Avatar
    dgfoster is online now Stainless
    Join Date
    Jun 2008
    Location
    Bellingham, WA
    Posts
    1,993

    Default

    Quote Originally Posted by LFLondon View Post
    No. The full article is at Swensen's web page and I am sure there is plenty more across the Web from other sources. I thought it might be useful to some.
    Oh. (puzzled look)
    Denis
    1HobbyMachinist and Pete F like this.
  5. 12-15-2014, 12:06 AM #5
    Ironwoodsmith is offline Aluminum
    Join Date
    May 2009
    Location
    Friday Harbor, WA
    Posts
    78

    Default

    I thought it was a good post. I certainly learned something about straight edges.

    thanks
  6. 12-15-2014, 12:57 AM #6
    Join Date
    Aug 2010
    Location
    Mobile, Alabama
    Posts
    196

    Default

    Not exactly a new idea, is it?
  7. 12-15-2014, 01:25 AM #7
    Forrest Addy is online now Diamond
    Join Date
    Dec 2000
    Location
    Bremerton WA USA
    Posts
    9,413

    Default

    The OP needs to look back in the archives for the acheivements of Steven Thomas, D G Fortser and some others whose names I will recall in the middle of the night.

    They've made straight edges by starting with their basic concept, particular requirements, and research. They made a pattern, got castings made then machined and scraped them.

    There's another fellow who make patterns to make wax segments that when joined, dipped, sanded, and melted out produced a box straight edge casting molds that when cast and stress relieved producted castings similar to that showcased in Moore's "Fundamentals of Mechanical Accuracy" I think the same guy cooked up some fine cast iron set-up squares. He took my scraping class in Oregin in 2007 as I recall. Beautiful work. There's also a slug of us regular posters who've collectively scraped in acres (seeming) of reference tooling.

    It's good when a guy takes the bull by the horns and made his own reference tooling in complince with if not certified to commercial standards. He should do this knowing he's not breaking new ground however laudible and successfull his efforts. Meet the challege even if you have to re-invent the wheel. Do so, and your entitled to strut and preen a little. Those of us who have been there will applaud your arrival.
    You and Antarctica like this.
  8. 12-15-2014, 04:29 AM #8
    LFLondon's Avatar
    LFLondon is online now Hot Rolled
    Join Date
    Dec 2007
    Location
    North Carolina
    Posts
    669

    Default

    Quote Originally Posted by Forrest Addy View Post
    The OP needs to look back in the archives for the acheivements of Steven Thomas, D G Fortser and some others whose names I will recall in the middle of the night.

    They've made straight edges by starting with their basic concept, particular requirements, and research. They made a pattern, got castings made then machined and scraped them.

    There's another fellow who make patterns to make wax segments that when joined, dipped, sanded, and melted out produced a box straight edge casting molds that when cast and stress relieved producted castings similar to that showcased in Moore's "Fundamentals of Mechanical Accuracy" I think the same guy cooked up some fine cast iron set-up squares. He took my scraping class in Oregin in 2007 as I recall. Beautiful work. There's also a slug of us regular posters who've collectively scraped in acres (seeming) of reference tooling.

    It's good when a guy takes the bull by the horns and made his own reference tooling in complince with if not certified to commercial standards. He should do this knowing he's not breaking new ground however laudible and successfull his efforts. Meet the challege even if you have to re-invent the wheel. Do so, and your entitled to strut and preen a little. Those of us who have been there will applaud your arrival.
    Thanks a million for your reply, absolutely amazing and inspirational! That is where the rubber meets the road to say the least, that anyone can tread the sacred ground of making their own accurate reference tooling. I was not aware of those particular accomplishments of Foster and Thomas and will search PM for their posts on this subject. Much credit to them and to you. I need to spend some time exploring the world of scraping and pick up a few tools to experiment with in my shop.
    LFLondon
  9. 12-15-2014, 05:50 AM #9
    Pete F is online now Titanium
    Join Date
    Jul 2008
    Location
    Sydney, Australia
    Posts
    2,496

    Default

    It's interesting this has come up now. Down here long straight edges are hard to come by. I have a 3' and that will probably do me, but I did wonder just the other day if it may be possible to make up a longer straight edge with a weldment, and then attach a shoe(s) on the bottom of cast iron to form the straight edge itself that could be scraped in. It seems to me it could save an expensive one off casting that in itself could be of dubious quality.

    Has anyone ever seen this type of thing done before?
    anoldcrank likes this.
  10. 12-15-2014, 08:07 AM #10
    Mark Rand is offline Titanium
    Join Date
    Jul 2007
    Location
    Rugby, Warwickshire. England
    Posts
    3,103

    Default

    He doesn't seem to show how to avoid twist...
  11. 12-15-2014, 09:43 AM #11
    Richard King is online now Stainless
    Join Date
    Jul 2005
    Location
    Cottage Grove, MN 55016
    Posts
    1,949

    Default

    Quote Originally Posted by Pete F View Post
    It's interesting this has come up now. Down here long straight edges are hard to come by. I have a 3' and that will probably do me, but I did wonder just the other day if it may be possible to make up a longer straight edge with a weldment, and then attach a shoe(s) on the bottom of cast iron to form the straight edge itself that could be scraped in. It seems to me it could save an expensive one off casting that in itself could be of dubious quality.

    Has anyone ever seen this type of thing done before?
    I have thought about extending cast iron plates but not SE's over the years that way Pete or setting 2 or 3 granite plates side by side to make longer ones when all you have are short ones. I never have gotten around to trying. Not enough spare time yet. I have read about people trying to use steel scraping straight-edge and again have not tried it myself. I suppose if your not holding tight tolerances it would work if you can control the temp so they don't grow. I have seen steel weldment machine tools come and go. Never heard of one being very stable. Pete I would be interested in what you discover if you try it.

    A old friend patented a portable mill expandable round way by using what looked like tapered sections that fit together and held with a device that resembled a D mount lathe chuck. Maybe you can make a expandable SE like that?

    I have thought about experimenting with making a cast straight-edge and cylindrical grinder beds with a threaded steel rod (or cast Iron), like a press brake back gage that has a straightness adjusting bar. When I scrape the bed of a grinder concave when the base is shorter then the table I have thought it would be nice to be able to adjust it concave, as scraping it to tenths is a pain. The bending contraption would help in non temperature controlled areas. As the temp outside changes the SE bends and you would adjust the bolt to compensate.....something for you young inventors to think about.

    I have seen this issue working in Missouri where it is cool at night and early morning and by noon it's 100 F degree's and your straight edge changes. Not a lot but a few tenths. On one job years ago at a Federal Mogel shop in SE MO, we had 2 straight-edges. One for AM and one for PM. They were identical Brown & Sharpe 48" er's. It seemed silly then and now but that worked better then scraping the one after lunch everyday. lol.. The centerless grinders we were rebuilding had mass and didn't change as fast as the SE's They eventually air conditioned the rebuilding shop.

    The Moore book has been around for years and showed how back before modern methods of testing were invented they used 3 plate method. The only way there was to achieve flatness way back then. I wonder who came up with the idea? I recall when I was a teenager lapping 3 lapping plates together for Control Data for some project they had (my Dad's Company and we contracted to many companies rebuilding equipment). That was when we didn't own a Granite plate and all we had was hand scraped surface plates. That was a real pain scraping plates back then with HSS (1960's) I can only imagine how hard it would have been in the early years...

    When teaching machine building /Scraping in Taiwan one of my students at YCM did an experiment and scraped 3 HKA-24 straight-edges (King-Way or my Family Brand Straight-edge) with a BIAX power scraper and got the 3 he scraped together to 60 PPI . When he was finished they tested them in their Metrology lab and they were all identical and within .00004" flatness. They are a work of art, and are now on display in the company show room.

    I met Greg Dermer from Portland who told me he made Moore duplication patterns out of wax Straight-edges, when he took one of my scraping classes a few years ago. He told me he wanted to learn to scrape 40 points per inch using a hand scraper and a BIAX. He also made some nice non ribbed angle straight-edges. Bushe Precision makes an Aluminum with cast iron scraped sub plate epoxied on Straight-edge A6100 Series - Aluminum Parallel Straight Edges On Busch Precision, Inc..

    Thanks LF for starting the thread. It's nice to remember the old days and not re-write history. Rich
    old_dave likes this.
  12. 12-15-2014, 10:52 AM #12
    DMF_TomB is offline Titanium
    Join Date
    Dec 2008
    Location
    NY, USA
    Posts
    2,784

    Default milling

    I often mill parts to .0005" per 40" flatness and straightness tolerance.
    You would only need to scrape if you wanted straighter
    .
    On a cnc mill I often uses a marker to make sure I mill last .001", often
    the mark is removed only partially on some of it which usually means
    the milling is removing only .0002", getting finer is difficult because
    of metal deflection and cnc servo errors. The cnc mill often oscillates
    .0001" or more. Still the need to hand scrape for finer tolerances
    Richard King likes this.
  13. 12-15-2014, 11:23 AM #13
    LFLondon's Avatar
    LFLondon is online now Hot Rolled
    Join Date
    Dec 2007
    Location
    North Carolina
    Posts
    669

    Default

    Quote Originally Posted by Pete F View Post
    It's interesting this has come up now. Down here long straight edges are hard to come by. I have a 3' and that will probably do me, but I did wonder just the other day if it may be possible to make up a longer straight edge with a weldment, and then attach a shoe(s) on the bottom of cast iron to form the straight edge itself that could be scraped in. It seems to me it could save an expensive one off casting that in itself could be of dubious quality.

    Has anyone ever seen this type of thing done before?
    Not much experience here but it occurred to me that with a welded piece the welded zone and adjoining area affected by heat might have a different crystalline structure (even after post weld heat treatment and stress relief) and that might cause unwanted minute deformation of the tool with temperature changes rendering it useless except for low tolerance work.
    Richard King likes this.
  14. 12-15-2014, 11:30 AM #14
    Forrest Addy is online now Diamond
    Join Date
    Dec 2000
    Location
    Bremerton WA USA
    Posts
    9,413

    Default

    A temperature compensated fabricated steel/cast iron straight edge. Hm. Interested concept, Richard. There's enough difference betweeen the thermal expansion rates of cast iron (varies with composition ) and steel (pretty constant) to make composite metal scrape reference tooling an unfulfilled dream. Many times I've looked at a steel tube camelback faced with cast iron and got bogged down with the expansion problem. Weight and stiffness advantages await but the thermal stability problem trumps all.

    Richard's musings and a flash recollection of a cased clock's thermal compensated pendulum (where brass and steel elements are subtractively combined to produce a pendulum whose length and period does not change with temperature) led me to musings of my own.

    Imagine a straight edge whose upper limb is comprised of brass and steel in a way that its expansion rate exactly equals that of the cast iron shoe.

    The proportions of brass to steel would depend on the particular cast iron's expansion rate. As I said, that rate varies slightly enough to screw up the straighness tolerance over a scraped reference tool's likely thermal range in service (40 to 100 degees F).


    For example, look at generic thermal expansion coefficients (in/in/degree F 10E-6)

    Brass 11.3
    Steel 7.8
    Cast iron 5.8

    These figures will vary depending with the actual materials selected.

    So brass and steel will have to be combined subtractively so their combined rates equal that of the cast iorn selected.

    One could work the math to obtain the proportions but the length of steel subtractively compensated by a length of brass to equal that of a unit of cast iron will have to be determined experimentally from samples of the actual materials. Book values are all over the map.

    Before the nay-sayers start shouting down the concept of a - um - "tri-metal" straight edge they should consider accurate thermometers based on bi-metal strips have been in use for generations. I'm proposing the bi-metal principle be applied in the opposite sense. I vaguely perceive a clever design and metalurgical stability are just over the horiaon.

    The design of the posited structure will have to incorporate elements leading to stability and rigidity equalling that of an equivalent cast iron straight edge but with reduced weight, in range of a home workshop rescources, etc.

    A tall order but do-able in my fevered imagination. Hmm indeed.
  15. 12-15-2014, 11:46 AM #15
    waynes is offline Cast Iron
    Join Date
    Mar 2011
    Location
    Trenton, On
    Posts
    339

    Default

    I suspect the straight edge varies due to changing temperatures - if given time to stabilize at any particular temperature, I would expect it would end up straight.

    If correct, having the base with a very similar cross section to the arch would probably prevent most of the problem - though it may make the straight edge rather heavy and unwieldy.
  16. 12-15-2014, 11:46 AM #16
    DMF_TomB is offline Titanium
    Join Date
    Dec 2008
    Location
    NY, USA
    Posts
    2,784

    Default

    i remember a straight linear rail bolted to a concrete floor under a skylight window at roof.
    .
    the next day the linear rail was bowed or curved when checked with optics. we had wanted to
    use a non contact capacitance sensor connected to a lap top to see if a 20 ton cast iron wheel
    was round.
    .
    we waited til the next day when it was cloudy and the linear rail was straight again. we measured the
    20 ton wheel and found it was not circular or octagon but partially both. the theory being when the
    wheel was ground the grinder as on the side. when the wheel turns the part between the spokes is
    on the top of wheel during rotation that part sags from the weight.
    .
    basically the wheel was changing shape every time it rotated. the wheel was in a air pressurized
    chamber at only top and the had a air pressure gage showing the pressure fluctuation every time wheel rotated
    the gap at wheel sides sealing the leaking air was changing so the air leakage was changing as it
    rotated
    .
    parts sag from their own weight. a straight edge flat when horizontal may not be flat when held
    vertical. sort of like a 20 foot long 1" round steel bar sags more when held horizontally than when
    held vertical, when someone says they have a straight edge flat and straight within .0002" per 40"
    i just think maybe if held horizontally it is straight. when same straightedge is turned on its side
    and horizontal again it may bend differently and no longer be straight.
    .
    we used to use a precision invar tape measure. it depended a lot on whether it was continuously
    supported, sorted on ends only, supported on 3 points, 4 points, etc. there was a hundreds of
    pages long length compensation table showing true values as the support method was change
    Demon73 likes this.
  17. 12-15-2014, 11:48 AM #17
    Richard King is online now Stainless
    Join Date
    Jul 2005
    Location
    Cottage Grove, MN 55016
    Posts
    1,949

    Default

    Per Dennis's post #4 on Forrest's last post #14.

    Oh. (puzzled look)

    Rich
  18. 12-15-2014, 12:08 PM #18
    Richard King is online now Stainless
    Join Date
    Jul 2005
    Location
    Cottage Grove, MN 55016
    Posts
    1,949

    Default

    Quote Originally Posted by DMF_TomB View Post
    i remember a straight linear rail bolted to a concrete floor under a skylight window at roof.
    .
    the next day the linear rail was bowed or curved when checked with optics. we had wanted to
    use a non contact capacitance sensor connected to a lap top to see if a 20 ton cast iron wheel
    was round.
    .
    we waited til the next day when it was cloudy and the linear rail was straight again. we measured the
    20 ton wheel and found it was not circular or octagon but partially both. the theory being when the
    wheel was ground the grinder as on the side. when the wheel turns the part between the spokes is
    on the top of wheel during rotation that part sags from the weight.
    .
    basically the wheel was changing shape every time it rotated. the wheel was in a air pressurized
    chamber at only top and the had a air pressure gage showing the pressure fluctuation every time wheel rotated
    the gap at wheel sides sealing the leaking air was changing so the air leakage was changing as it
    rotated
    .
    parts sag from their own weight. a straight edge flat when horizontal may not be flat when held
    vertical. sort of like a 20 foot long 1" round steel bar sags more when held horizontally than when
    held vertical, when someone says they have a straight edge flat and straight within .0002" per 40"
    i just think maybe if held horizontally it is straight. when same straightedge is turned on its side
    and horizontal again it may bend differently and no longer be straight.
    .
    we used to use a precision invar tape measure. it depended a lot on whether it was continuously
    supported, sorted on ends only, supported on 3 points, 4 points, etc. there was a hundreds of
    pages long length compensation table showing true values as the support method was change
    Tom it must have be a learning experience working at one of the most precision machine builders, Gleason. I enjoy reading your posts (now that I know you..lol) What your telling us reminded me of Lyon Precision and their work with Air Bearings Inc. They have developed products and employee's have written books detailing spindle rotation and testing it.. Here you can and watch this and understand it. http://www.lionprecision.com/sea/revolutionVideo.html &
    Precision Spindle Metrology: A book by Eric Marsh

    Reminds me of when I was touring a plant in Taichung and one of the engineers described a situation they had. They were building CNC lathes and would move them in an assembly line. One station was for scraping and alignment, then it would move to electrical install and then guarding and final testing run and at that location the machine was out of tolerance. I figured it out at once..The machine was under a huge window and the sun was shinning on it. We took a temp gun and the front of the machine was hotter from the sunshine then the back that was in the shade. Rich
  19. 12-15-2014, 12:27 PM #19
    DMF_TomB is offline Titanium
    Join Date
    Dec 2008
    Location
    NY, USA
    Posts
    2,784

    Default

    Quote Originally Posted by Richard King View Post
    Tom it must have be a learning experience working at one of the most precision machine builders, Gleason. I enjoy reading your posts (now that I know you..lol) What your telling us reminded me of Lyon Precision and their work with Air Bearings Inc. They have developed products and employee's have written books detailing spindle rotation and testing it.. Here you can and watch this and understand it. Revolution in Spindle Measurement Video Documentary &
    Precision Spindle Metrology: A book by Eric Marsh

    Reminds me of when I was touring a plant in Taichung and one of the engineers described a situation they had. They were building CNC lathes and would move them in an assembly line. One station was for scraping and alignment, then it would move to electrical install and then guarding and final testing run and at that location the machine was out of tolerance. I figured it out at once..The machine was under a huge window and the sun was shinning on it. We took a temp gun and the front of the machine was hotter from the sunshine then the back that was in the shade. Rich
    .
    i used to work for Kodak. they had many big multi story tall machines that moved from temperature
    changes or different weight on a floor
    .
    anybody looking through a optical level with a fork truck driving by cannot miss the floor moving.
    .
    when taking some measurements i had to ask people to stand away from a optical scale as their
    weight standing on the floor was making the floor go down distorting the optical scale reading
    Richard King likes this.
  20. 12-15-2014, 12:37 PM #20
    Richard King is online now Stainless
    Join Date
    Jul 2005
    Location
    Cottage Grove, MN 55016
    Posts
    1,949

    Default

    I do that with the classes Tom, We put a level on the granite plate and have the class crowd around behind the plate. Have one student watch the bubble..usually a Starrett 198 .0005" / 12" and then we all walk to the other side of the plate and the bubble moves. It is so interesting when you share the stories of where you worked. Thanks. It would be a great service to all or to the history books if all of the machinists were to share all these stories so younger generations can read about them. Have to Thank Milacron for giving us this forum! Rich
    Demon73 likes this.



  • [machinist] Making Accurate Straight-Edges from Scratch by John A. Swensen, Lawrence London, 01/02/2015

Archive powered by MHonArc 2.6.24.

Top of Page