584,830 active members*
5,897 visitors online*
Register for free
Login
Results 1 to 16 of 16
  1. #1
    Join Date
    Jul 2018
    Posts
    6248

    Gantry Design Biomimicry

    Hi All - I'd like to invite Andrew (Awerby) to co design his dream gantry with me. Andrew is a sculptor and will bring values to the design that a boring mechanical engineer won't. Andrew what natural shape or muse thing/object would suit a gantry? Describe your dream machine so we can flesh it out....Peter

  2. #2
    Join Date
    Sep 2018
    Posts
    83

    Re: Gantry Design Biomimicry

    I don't really think that biomoimicry is going to do you any great favours given that you need to meet too many criteria for a functional machine. A gantry needs to be stiff, straight, rigid. Nature has produced some fantastic solutions to problems, but precise, repeatable linear control isn't really one of them.
    However, if you're determined to follow this route, I'd look at both exo and endoskeletal structures.

  3. #3
    Join Date
    Jul 2018
    Posts
    6248

    Re: Gantry Design Biomimicry

    Hi Gerry - Thks for the thought. But I'll try not to guess the result. Its a conceptual exercise to see what happens. exo, isogrids, diagrids who knows? My work on gantries so far has show that lightweight sections need internal transverse shear elements to prevent warping of the shape, mid weight sections are torsion dominant and we know circular is best, then heavy sections become bending dominant. Nature didn't design a femur to be a straight column, its slightly spiraled to absorb shock. Plus we make gantries the same section right across because that's convenient not really required. Lets find out with some imagination and effort what it could look like.... If Andrew comes on board...Peter

    stiff - yes
    straight - only the rails
    rigid - same as stiff in my world
    and I'll add damp

  4. #4
    Join Date
    Jul 2018
    Posts
    6248

    Re: Gantry Design Biomimicry

    Hi all - Add tensegrity and geodesic structures to the list as well. Peter

    see the last few entries of this as lead in for this thread... https://www.cnczone.com/forums/diy-c...neering-5.html

    The attached steel gantry shape has 46% deflection of my benchmark 100x100x6mm steel tube with the same envelope.

    Torsion implies a plain circular section. It has to be parallel as the torque is the same along its length no matter where the tool is. It implies a minimum thickness tube to prevent wrinkling and of adequate inertia to achieve the required rigidity. An isogrid or diagrid tune could be lighter.

    Bending implies a square or rectangular section but tapered as the BM declines towards the ends. But is that true? I'll have to test that.

    Shear requirement is same along the gantry so the shear stiffness and section inertia has to be consistent along the tube to prevent local lozenging. An over critical circular tube with internal shear members may do better.

    All of these are membrane solutions and skeleton type structures maybe better. These are not usually considered as they are difficult to model and manufacture and difficult to analyse but times have changed...

    All of these considerations also imply we are seeking a minimum weight solution. cheers Peter

  5. #5
    Join Date
    Jul 2018
    Posts
    6248

    Re: Gantry Design Biomimicry

    Hi All - I built a beam model of our 1000mm long 100mm dia x 6mm thick tube and applied 1000N at various spots along it to look at trends. beams are useful as I can get std moment and shear diagrams out of them.

    1) The shear load increases by a factor of 2 from the middle to the ends. So the transverse beam area or shear support area can be much less at the middle. This implies it can be tapered towards the middle or it only needs shear support at the ends
    2) In pure bending the deflection is about half the benchmarks so we could say its half bending and half torsion deflection in this case
    3) The moment diagram shows it diminishes from 250,000Nmm to 90,000Nmm at the ends. Again this implies it can be tapered.

    So now have to do same with torsion but that's for another day. These are membrane solutions and once we have a direction we can look at skeletal or other concepts. Peter

  6. #6
    Join Date
    Jul 2018
    Posts
    6248

    Re: Gantry Design Biomimicry

    Hi all - Taking the circular design a step further and from prior models trends, making the rails as far apart as possible makes this seed shape. Its slightly stiffer then the flat face design. Down in the 0.001mm so same. So next is to do some skeletal modelling. The work so far says that torque is a big factor this points at a diagrid or isogrid. In an isogrid the longitudinal members are the main bending resistors and the windings take the torsion. In a diagrid the windings do the lot. Need to think a bit about how to model this easily. If anyone is into rhino/grasshopper that would be a great help. I've started reading up on grasshopper. Been a Rhino user for 20 years but haven't ventured in GH yet. Peter

    Also I don't want to preempt the manufacturing process, so seek the optimal geometry then solve how to do it vs the other way round.

  7. #7
    Join Date
    Apr 2004
    Posts
    5728

    Re: Gantry Design Biomimicry

    It's an interesting question, Peter. A lightweight part is what attracts me to the idea of a composite beam, and the vacuum-bag process opens up a lot of design possibilities that you don't get with a straight extrusion. Biomimicry is a good source for design ideas when the thing you're trying to create has a natural counterpart in a problem that Nature has already solved, but as Gerry points out, the need to hold a sliding vertical member that's torqued from the end as rigidly as possible to a horizontal beam isn't one that comes up a lot for her.

    The FEA analysis you've done is helpful; it points to a design for a beam that has points of attachment at each end that are as widespread as possible while narrowing at the middle, somewhat like a suspension bridge, or in bio terms, the structural lines of a spider's web. But against that one needs to consider the natural tendency for any long structure supported at the ends to sag most in the middle, while resonance - a major cause of problems in machining - is greatest midway between node points, as in a guitar string. Dampening that effect should be a major concern in designing a gantry beam. It's probably best dealt with by providing interior structures that counteract any tendency for the beam walls to oscillate at the points of maximum deflection. Of course, starting with a composite material gets you partway there in itself, since it performs better than most metals in this regard.

    If we consider the failures of suspension bridges in the face of resonant frequencies the importance of dampening becomes apparent. The Tacoma Narrows bridge collapse was the most famous, but there was also a Millennium bridge built in London that had to be closed and retrofitted with tuned liquid dampers before being put back into service. Perhaps something like that would be useful in a gantry beam, or even an active dampening system that sensed resonance and applied counter-forces to counteract them could be implemented.
    Andrew Werby
    Website

  8. #8
    Join Date
    Jul 2018
    Posts
    6248

    Re: Gantry Design Biomimicry

    Hi Andrew - If the structure is not skeletal then it can be filled with a dense liquid or water to act as a damper. Or a liquid damper can be added. I think the next step is for you to tell me a little more about your 2m gantry (and machine in general as the gantry has to connect to the machine somehow) ie its intended usage so I can scale up the current benchmarks. Currently I prefer lifting gantry designs as this makes the gantry only deal with a consistent torque as the Z axis does not move up and down. Depending on what you want to machine means the lifting gantry may not be suitable. Although I have thought thru this and can't see any disadvantages yet. Peter

    Destructive resonance is a problem with very lightweight structures but we have already made the gantry twice as stiff in the same envelope with similar mass so have moved the resonance up a bit. These things can be simulated and understood before build. Lets keep investigating...

  9. #9
    Join Date
    Jul 2018
    Posts
    6248

    Re: Gantry Design Biomimicry

    Hi All - Andrew has stepped in which is great. He sent me a couple of shots of his machine so I will scale up the design to suit. I built a torque model in the round tube and here's the vector plot of the deflection. So it indicates as expected that the gantry could be tapered to the outside. The central deflection is 0.115mm and at the edges its 0.055mm so good scope for tapering. I'll look at Andrews design and figure out the next concept.

    One reason for tapering is the more complex the structure the less likely it will have simple low freq vibration modes. Parallel sides, rectangles and thin structures have lots of freedom to vibrate. Triangular, curved tapered shapes vibrate less. Peter

  10. #10
    Join Date
    Apr 2004
    Posts
    5728

    Re: Gantry Design Biomimicry

    It seems that filling a hollow composite gantry beam with liquid would negate its major advantage, which is light weight. But I agree that casting something that's curvy and triangulated would be less likely to resonate. Are you saying that the gantry beam should be tapered so that it's widest in the middle, or widest at the ends?
    Andrew Werby
    Website

  11. #11
    Join Date
    Jul 2018
    Posts
    6248

    Re: Gantry Design Biomimicry

    Hi Andrew - Widest at the middle. The only load that creates wider at the ends is the shear load but this is not a dominant condition. I'd prefer to design a damp structure with no liquid damping. Just putting all ideas out there for now. Next step is to design a benchmark gantry then look at alternatives. Your original gantry is aluminium? Peter

  12. #12
    Join Date
    Apr 2004
    Posts
    5728

    Re: Gantry Design Biomimicry

    It seems that the main torsion would happen if the Z axis was fully extended downward and was in the middle of the beam. Any pushing back and forth at the spindle would stress the attachments of the beam to the rest of the structure, so I'd think more width and thickness there would be helpful. If there was any increase in width for the gantry beam, it would have to play toward the top, to avoid clearance problems.

    The original plan for a gantry beam in my drawing was dual square-section steel tubes that have the profile rails attached in front, with a rectangular steel tube between them, all bolted together to a steel plate on the back. I originally was considering aluminum, but it doesn't seem very stiff - making it as stiff as an equivalent steel beam would mean increasing the weight considerably, which obviates the advantage of using aluminum and would cost a lot more too. But I think composite would weigh less than aluminum and be stiffer per pound (or kilo).
    Andrew Werby
    Website

  13. #13
    Join Date
    Jul 2018
    Posts
    6248

    Re: Gantry Design Biomimicry

    Hi Andrew- so what is the size of the benchmark gantry? 300x200x12? You mention 17" high so 435H (17") x 300mm deep x 10mm thick? and how far is the tool below the gantry 400mm? Peter

  14. #14
    Join Date
    Jul 2018
    Posts
    6248

    Re: Gantry Design Biomimicry

    Hi Al - My first shot at a parametric isogrid. I usually do this in Rhino but then its hard to adjust, In fact every iteration has to be rebuilt entirely. That's why I should get into Grasshopper. The prior semicircular solution weighs 19.77kg and this weighs 11.95kg nearly half. It deflects 0.36mm vs the 0.12mm of our best so far, poor result. But its an attempt. The grillage should be behind the rails.but I couldn't figure out how to do that parametrically. I built this as two counter helixes and some longitudinals. I'll have another crack soon. My thoughts are I can have a tube laser cut with all the holes, but then I'm choosing the solution based on manufacture. I can 3D print nearly anything so shouldn't consider manufacture yet. Peter
    Attached Thumbnails Attached Thumbnails femur.JPG   isogrid 1.JPG   Iso 1.jpg  

  15. #15
    Join Date
    Jul 2018
    Posts
    6248

    Re: Gantry Design Biomimicry

    Hi All- I also did a semi circle in fibreglass (30GPa). It weighs 11kg so same as steel isogrid. It deflects 0.23mm so is ahead of the isogrid but not as good as the steel semicircle. The FG section is 20mm thick. If I went to CF (70GPa) then it would be lighter and as good as the steel semicircle but 20mm of CF is costly. Easy enough to make the section bigger to get to the required deflection but then loose apples to apples on envelope.cheers Peter

  16. #16
    Join Date
    Jul 2018
    Posts
    6248

    Re: Gantry Design Biomimicry

    Hi All - I decided to use a 250x250x12mm x 2m long SHS steel as a benchmark. I made a test model to figure out how to make a foam centre and the skin in the one model. I did this by creating a solid then shelling outward in a new configuration. This way the faying surfaces are perfect. I used an infinitely stiff lever 400mm below the gantry with a 1000N force. The benchmark deflects 0.039mm this is a static stiffness of 26N/um.

    Benchmark weighs 178kg, steel shell model weighs same, 179kg and deflects 0.035mm (its 10mm thick). The fibreglass version filled with PU 300kg foam weighs 73kg and deflects 0.115mm so needs some work 9N/um. cheers Peter

    I also ran the model where the composite one was a solid casting of EG (30GPa = E) it weighs 250kg but is stiffer. So a hollow casting will get the same stiffness and lose weight just have to figure that out.

Similar Threads

  1. Generative Shape Design - Let the Computer Design the Part
    By abender in forum Mechanical Calculations/Engineering Design
    Replies: 1
    Last Post: 06-17-2019, 04:15 AM
  2. Replies: 0
    Last Post: 09-24-2018, 10:41 PM
  3. Design Feedback For 8020 Ballscrew and Linear Rail Design
    By GisMo in forum DIY CNC Router Table Machines
    Replies: 4
    Last Post: 02-01-2018, 01:30 PM
  4. Replies: 0
    Last Post: 05-14-2013, 03:23 PM
  5. Finding Engineering Design Software For Automatic Machine Design
    By hellokitty in forum Uncategorised CAM Discussion
    Replies: 0
    Last Post: 01-06-2008, 07:39 AM

Tags for this Thread

Posting Permissions

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