Design/Production Tolerances

Discussion in 'Aircraft Design / Aerodynamics / New Technology' started by highspeed, Apr 13, 2011.

Help Support HomeBuiltAirplanes Forum by donating:

  1. Apr 14, 2011 #21

    Jay Kempf

    Jay Kempf

    Jay Kempf

    Well-Known Member

    Joined:
    Apr 13, 2009
    Messages:
    3,587
    Likes Received:
    901
    Location:
    Warren, VT USA
    This is not true. I cut wood in my home shop to thousandths with band saws, table saws, routers, good fences and good cheap measuring tools. Course I did a long stint in a master pattern shop working for a journeyman master pattern maker. With a good blade in a band saw and 3M77 adhesive and a large format printer ($300 off of ebay) as well as some magnification and a good set of sanders you can get a profile to split a .005 wide line. I have done this over and over in all kinds of applications. Thin plywood over ribs can be a very accurate airfoil. It may not be as rigid as foam backed carbon or equiv but it is pretty good as long as it is supported on a short enough interval. Wood is still the original composite material. It isn't as versatile as modern composite techniques for a one off but it is very easy to use and build light strong structures with wood if you limit 3d curvature.

    My material of choice for cheap QC templates is free formica scraps from a local cabinet shop. Works great and it is rigid and dimensionally more stable than aluminum and they have barrels of the stuff. I haven't done it yet but a good dremel gantry router made out of scrap printer carriages are getting cheaper and easier to get (open source) every year. One 2D router can hit a few thousandths pretty reliably. I haven't seen anyone just put a dremel in a DesignJet printer head but we are getting closer to that every year.
     
  2. Apr 14, 2011 #22

    Dana

    Dana

    Dana

    Super Moderator

    Joined:
    Apr 4, 2007
    Messages:
    8,508
    Likes Received:
    2,918
    Location:
    CT, USA
    Even if you can cut wood to .005, it won't stay at that dimension for long... wood changes a lot with temperature and humidity variations. I used to work with composites and thermoforms for automotive applications, where we usually used CNC cut wood for prototype tooling. Even in a climate controlled shop, it was a major problem.

    -Dana

    Beware of strange places and dark dingy places, be careful while bending the law...
     
  3. Apr 14, 2011 #23

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,560
    Likes Received:
    5,220
    Location:
    Orange County, California
    It's been done in wood. Sand and fill, sand and fill... Norman and Autoreply are a bit more knowledgeable than I about this, and the methods Norman describes are the traditional ones that work on virtually any construction material. As AR said, these are methods and accuracies for racing sailplanes. Compromise is not an option.
     
  4. Apr 14, 2011 #24

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,560
    Likes Received:
    5,220
    Location:
    Orange County, California
    Yep. Wooden-winged competition sailplanes used to be reprofiled two or three times a season. My understanding (AR, confirm?) is that modern composite ships only get the treatment once a season or so.
     
  5. Apr 14, 2011 #25

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,560
    Likes Received:
    5,220
    Location:
    Orange County, California
    We're going a bit OT, but yes, this has long been a rant of mine as well, especially when it comes to defense or space work. We (the USA/Europe) over-build stuff. Everything has to be a perfect little gem. In commerical products this has a place, as AR pointed out. But on a jet fighter? A space station? Pluuuuueeeeze! In some ways, the Soviet/Russian practices are much more practical. Personally I think they tended slightly too far the other way, with some things just a little too Beverly Hillbillies for my taste. But not everything has to be engineered to the 'nth' degree. The classic (and possibly legendary) space-related story is how the NASA invested several tens of thousands of dollars developing a ball-point pen that would write in zero-G. The Soviets just gave their guys a pencil. :gig:
     
  6. Apr 14, 2011 #26

    Jay Kempf

    Jay Kempf

    Jay Kempf

    Well-Known Member

    Joined:
    Apr 13, 2009
    Messages:
    3,587
    Likes Received:
    901
    Location:
    Warren, VT USA
    Depends on the wood and of course with wood humidity and temperature are variables. So to do any sort of accurate work you have to have a stable environment. We used to use a whole mix of materials and mostly for stability we used Mahogany a lot as well as MD40 for tooling.

    That is why I said I use Formica. That stuff is remarkably stable. It is basically a phenolic sort of material at heart. This was withing the context of how do you check an airfoil for accuracy. My point was that you can do a really good job of being accurate with simple and inexpensive tools. More accurate than is needed for most HB applications. I do my own car alignments at home with simple tools and some knowledge of trig and gravity.
     
  7. Apr 14, 2011 #27

    autoreply

    autoreply

    autoreply

    Moderator

    Joined:
    Jul 8, 2009
    Messages:
    10,732
    Likes Received:
    2,539
    Location:
    Rotterdam, Netherlands
    I'm not sure I understand your re-profiling correctly. I know that several world champions didn't do anything to their wings surface, except for proper waxing in the years before reigning their title and keeping it out of the sun if not necessary. Reprofiling is - to the best of my knowledge - only done if the skin is seriously out of shape. Usually, the Gelcoat/PU is stable enough to keep intact and in proper shape and accuracy for at least a decade, allowing full laminar.
    Opinions vary about the skin, but several gliders are produced in such a way that the spar-section post-cures and the resulting shrinkage causes a distortion in the skin.
    Now we're talking about shrinking properties anyway; any ideas why steel frames are used to support molds? If you're building a mold in the classical way (plug, cover with gelcoat, glass and pull it off, then put metal frame on), wouldn't it make more sense to make the stiffening frame out of glass too? Use some foam to get the required depth (and bending stiffness) and make a messy structure of glass. It seems more logical to me, but I'm sure there are reasons to do otherwise. Any thoughts?

    Here's an example
     
  8. Apr 15, 2011 #28

    GESchwarz

    GESchwarz

    GESchwarz

    Well-Known Member

    Joined:
    Oct 23, 2007
    Messages:
    1,157
    Likes Received:
    92
    Location:
    Ventura County, California, USofA.
    The best advice I can give on "tolerancing" is...don't cut it to length until you have to.

    Leave as much material on as you can until you have to remove it for final fit. If you are designing parts as you build, it is ideal if you can make a mockup of the part, say out of cardboard or similar cheap and easy material to fab, and use that as a pattern of the real piece. Mockups are especially valuable when you are making something entirely new. They are a cheap way to discover design issues and improvement opportunities.

    It is better to cut it long and have to cut it a second time for final fit, :) than to cut it "just right" then find out it's actually a little too short/small, then of course you have to make another one. :mad2:
     
  9. Apr 15, 2011 #29

    highspeed

    highspeed

    highspeed

    Well-Known Member

    Joined:
    Apr 10, 2011
    Messages:
    275
    Likes Received:
    13
    Location:
    Spring, TX
    Measure twice, cut once...then realize you used the wrong measurement, or measured from the wrong spot. A suitable array of four letter words are appropriate and warranted in that case.
     
  10. Apr 15, 2011 #30

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,560
    Likes Received:
    5,220
    Location:
    Orange County, California
    I've read several reports of seasonal reprofiling in the past - wish I could pull one up in my memory now... IIRC, one may have been in Strojnik. I'll check when I get a chance. In the meantime, we'll leave it at this.
     
  11. Apr 15, 2011 #31

    GESchwarz

    GESchwarz

    GESchwarz

    Well-Known Member

    Joined:
    Oct 23, 2007
    Messages:
    1,157
    Likes Received:
    92
    Location:
    Ventura County, California, USofA.
    Anybody who has worked in a shop very long has done those things. How about reading the tape measure upside down and mistaking 32" for 23" or 23 3/4 for 24 1/4. Easy mistakes if you are thinking about something else.
     
  12. Apr 15, 2011 #32

    Jay Kempf

    Jay Kempf

    Jay Kempf

    Well-Known Member

    Joined:
    Apr 13, 2009
    Messages:
    3,587
    Likes Received:
    901
    Location:
    Warren, VT USA
    OR MEASURING FROM THE 1 or 2" MARK AND FORGETTING TO SUBTRACT OUT THE OFFSET!!!!!!!!!!!!!!!!! Still getting me after all these years. Also cut the top off a nice new door recently!!!!! Mark the scrap side of the material before you carry the piece to the saw... And soooo many others.

    Another thing I ran into was having people in a shop all using their own tape measures. Then finding out that none of them matched. So we put up a standard 1' and 10' mark in the shop and people had to have their tape measures calibrated to the standard and witnessed to use them. ARGH!
     
  13. Apr 15, 2011 #33

    Topaz

    Topaz

    Topaz

    Super Moderator

    Joined:
    Jul 30, 2005
    Messages:
    13,560
    Likes Received:
    5,220
    Location:
    Orange County, California
    Heh. Yeah. :roll:
     
  14. Apr 17, 2011 #34

    JMillar

    JMillar

    JMillar

    Well-Known Member

    Joined:
    Jan 18, 2007
    Messages:
    233
    Likes Received:
    0
    Location:
    Antigonish, NS, Canada
    "I cut it three times, and its still too short??"
     
  15. Apr 17, 2011 #35

    GESchwarz

    GESchwarz

    GESchwarz

    Well-Known Member

    Joined:
    Oct 23, 2007
    Messages:
    1,157
    Likes Received:
    92
    Location:
    Ventura County, California, USofA.
    The old motto "measure twice, cut once" is only good if you are measuring the right thing and measuring correctly. Sometimes it's not a measurement that is wrong, but a calculation!!! Calculations should be cross checked by an alternate method to see if you get the same result.


    Many a foul explitives have been uttered as a result of not fulling understanding or being disoriented with regard to a unique design, and therefore pressing forward into cutting metal, only to find out that I made it backward, or as someone said, measured from the wrong feature.

    Also watch out for lines that were drawn and then design was revised but the obsolete lines were not completely removed. The part is then made using the obsolete line.
     
  16. Apr 17, 2011 #36

    orion

    orion

    orion

    Well-Known Member

    Joined:
    Mar 3, 2003
    Messages:
    5,800
    Likes Received:
    135
    Location:
    Western Washington
    As we used to say at the Lazy-B Ranch (Boeing): Measure with calipers, mark it with a Crayon, then cut it with a hatchet.

    Tight tolerances, properly applied, make great parts. Improperly applied, then you either have improper fits or very expensive parts. Many years ago, before I went back to school, I was a machinist/fabricator for Boeing. One day we were to make a series of hydro-formed parts that had tolerances down to .0005", something we never really saw before on the hydro-form. The part was sort of a cubic cup shape and given the tight tolerance we figured it was an engine part. The material was a mild stainless so it was easy to form but the rejection rates were high (about 75%). We finally finished the order and shipped it out but as a follow on I was aksed to se if I could follow up for next time and see if this tolerance was really necessary.

    It wasn't - the parts were ash trays.
     
  17. Apr 18, 2011 #37

    GESchwarz

    GESchwarz

    GESchwarz

    Well-Known Member

    Joined:
    Oct 23, 2007
    Messages:
    1,157
    Likes Received:
    92
    Location:
    Ventura County, California, USofA.
    Although your example is an extreme one, it is typical example of what is going on throughout the aerospace & defense industry. On the FBM program I mentioned earlier we have to hold to .0005" on molded plastic and rubber(!) parts. Even the prime contractor recognizes that it is over toleranced, but the hoops (cost) that would have to be jumped through to loosen the tolerances, including requalification are so great that it is prohibitive. Therefore our scrap rate is relatively high...we're rejecting lots of parts we know are good, but they don't meet the print. Is that insanity?
     
  18. Apr 18, 2011 #38

    autoreply

    autoreply

    autoreply

    Moderator

    Joined:
    Jul 8, 2009
    Messages:
    10,732
    Likes Received:
    2,539
    Location:
    Rotterdam, Netherlands
    Just an easy way to compete yourself out of the market because you're too expensive I guess. But then, most of them (defense contractors) are actually government companies so they don't have to worry about that.

    As for the tolerances; I feel pretty stupid for recognizing any single error that's mentioned here from experience :speechles
     
  19. Apr 19, 2011 #39

    GESchwarz

    GESchwarz

    GESchwarz

    Well-Known Member

    Joined:
    Oct 23, 2007
    Messages:
    1,157
    Likes Received:
    92
    Location:
    Ventura County, California, USofA.
    You won't find such a thing here in the USA.

    The problem we are discussing here is simply a result of engineering's failure to apply tolerances based on actual manufacturing capability, and a certain lack due dilligence in determining the level of criticality of each and every dimension and thus the required tolerance. Instead they apply standard and arbitrary tolerances. It is then the job of the Manufacturing and Quality Engineers to catch those errors before the drawing is released. When they fail to do that, and the part goes through qualification and passes, that's it. It's a little like naming a baby, to some degree you're stuck with the decision, particularly when the customer/government requires full requalification for any changes.

    It's just a matter of doing it right the first time.

    The earlier the problem is discovered, the cheaper it is to fix it.

    Check your drawing three times, measure twice, cut once.
     

Share This Page

arrow_white