Welcome to Our Community

Some features disabled for guests. Register Today.

Composite plates for Sphinx build

Discussion in 'CNC Mills/Routers' started by Steve Kingsley, May 14, 2017.

  1. Steve Kingsley

    Steve Kingsley Well-Known
    Builder

    Joined:
    May 28, 2016
    Messages:
    7
    Likes Received:
    4
    So I had this idea to try and mix the use of 3D printing and very basic workshop tools to build the plate set for a CNC build based on the Sphinx C-Beam design from @Kyo.

    My thought process went like this:
    • I have a 3D printer, how can I use it for this?
    • Thick metal plates are expensive (ish) at least where I am in the UK.
    • My workshop tools are at the cheap end of the spectrum. A small bandsaw, a Ryobi drill press and a Ryobi plunge router. They aren't that accurate. (by "they" I mean in my hands they aren't)
    • Cutting 6mm or thicker is going to be hard and slow with a lot of "cleaning up", and I hate filing.
    • I'm fascinated by stressed skin constructions like the torsion box top for my bench.
    The plan then, was to print a "core" from PLA and glue thin aluminium sheet to each side to act as the skins.

    First the "core":
    I took the design and made a minor change to the bearing hole as I have a 10mm lead screw I want to use. I also opted to go for 10mm thick to really be sure the concept would work and be stiff. I then used slic3r to produce a sliced print with zero top and bottom faces and a honeycomb infill. I had to scale the infill level to ensure that all of the holes were captive by the hexagons. I'm figuring that the perimeters around the holes will give good strength when in compression with the wheels. The picture shows it printing, though this is the second one as I forgot the first time.

    [​IMG]

    I made a mistake at this point. The print bowed a fair bit when I took it off the printer and I decided to "fix" that by running it under hot water and pressing it against a flat surface. It worked to correct the bow but warped the part in a way I didn't spot till the end :(

    The faces:
    I wanted to be reassured that the skin wasn't too flimsy so I went with 1.5mm thick anodised aluminium. I'd read somewhere in my searching (here I think) that the anodised finish may provide a better gluing surface due to its porosity??, plus it looks nice, so off to eBay - £8.40 for a sheet 400 x 400 x 1.5mm (vs about £45 for 6mm)
    I was happily able to cut the sheet to the rough shape with the bandsaw, leaving a little for clean up on the router. Yes, my plan was to clean the edges up to size with the router mounted in a table and use a flush trim bit to ride on the PLA core as a template. I glued the first rough cut face to the core with my "carefully selected" adhesive of choice: Gorilla glue - the original polyurethane stuff.

    Quick tangent: the glue selection.
    I went on a long and involved Google search for the best thing to use to glue PLA to aluminium. This turned up one theme, that not much likes to stick to PLA... or aluminium... I was not deterred. My first plan was to use epoxy and I came across lots of people suggesting that CA is fine for PLA to PLA. Loctite 401 came up a lot as a good glue for bonding aluminium (expensive! I'm looking to save money here) but looking at the data sheets it's not mentioned for the PLA side and I was also hoping for something with a longer working time. All of this research led me to think that my best bet was just to try something and see. I had this bottle of Gorilla glue staring me in the face which had recently seen action in the repair to the front of a model plane (EPO foam to Balsa) and it seemed very strong there.

    The glue up:
    I wanted to ensure I got the bow out of the core and generally that the resulting panel would be as flat as it could be. My bench is pretty flat so I then used a mirror tile (from the 3D printer) as "flatter" surface on that (that's how they make ). I lay the aluminium down and spread a thin coat of glue over the surface. Knowing that you need moisture to activate the gorilla glue, I soaked the PLA under the tap for a bit and toweled it off. I figured that enough would be absorbed or at the least wick into the gaps in the printed surface.
    I lay the printed core on the top and covered with some grease proof paper, in case the glue foamed up too much, then a 1/2" bit of birch ply, clamps and a heavy tool box for good measure. Wait 2 hours and presto

    [​IMG]

    The glue didn't foam up so well, maybe it dried to much while I was faffing with clamps and stuff.

    Trim and repeat:
    I used the router to trim the face to size. This worked pretty well but I think the bearing got a little warm at times and wanted to sink into the PLA rather than follow it. I then glued up the other side...

    [​IMG]

    This foamed up much better as I made it wetter this time. A final trim, though at this point I swapped to the spindle which had now arrived (Kress 1050 FME-P) as the runout on my router looked awful! The trim bit was also running on the aluminium from the other face here so felt a bit safer.

    So with the final trim and holes yet to be drilled through, it looks like this

    [​IMG]
    [​IMG]

    Provisional Conclusion:
    I'm yet to add the holes but my first impression is that this is a really stiff plate, I mean I have no way to really test that but just applying brute force in the hand it has no discernible flex to it. I'll update with some pics when the holes are in.

    Oh and the error I introduced at the start by heating the PLA core under hot water was that the thing isn't square!!! arrghh!! I'm hoping to mitigate this in the hole layout. It was dumb and I won't do it again, the second core has now come off the printer and is as square as my eye can tell with an engineers square.

    It's not a quick process so the "cost saving" may be purely a "cash in the pocket" thing but the time spent is fun hobby time not work production time. Also if I scale to 12mm (~1/2") my only cost change is the extra couple of hours on the printer, which is over night and autonomous at this point, and the PLA which is very small (hmm I should weigh it).

    If anyone has made it to the end, well done, and I'd love to get some feedback on this idea and if anyone can see any issues for using parts made this way on my first CNC build?

    Cheers
     
    Mark Carew and Kyo like this.
  2. Grantman

    Grantman Veteran
    Builder

    Joined:
    Apr 10, 2017
    Messages:
    57
    Likes Received:
    28
    I believe you'll need better glue. PU works best with porous materials so 30min epoxy might be a better choice. Isopropyl alcohol will release/soften the glue if you want to try again.
     
  3. JustinTime

    JustinTime Master
    Builder

    Joined:
    Dec 18, 2013
    Messages:
    917
    Likes Received:
    180
    Steve, maybe you should have drilled the holes before you glued the second skin? It would be easier than to try to hit the perfect spot now.
     
  4. Steve Kingsley

    Steve Kingsley Well-Known
    Builder

    Joined:
    May 28, 2016
    Messages:
    7
    Likes Received:
    4
    Thanks for the feedback.

    @Grantman On the glue front I think this one will stay as is, it's all trimmed up. I've got the core for the second one ready so I could try epoxy, I do have some 30 min. I could mark them and see if either one fails. To be honest, given the direction of force and the fact it's also clamped together in so many places, I'm having a hard time picturing the glue failing.

    @JustinTime My original plan was exactly that but I felt that holes in the first core print looked a little off. I had the nozzle size set wrong in slic3r (I usually use Cura). As it turns out that was a spot of luck due to messing up the squareness. I think I'll do that with the second one and use it as a drilling template for this one. That way they'll be spot on with each other. The PLA holes are a bit under size so should get cleaned up and align with the faces.

    We'll see how that goes
    Thanks again
     
  5. Steve Kingsley

    Steve Kingsley Well-Known
    Builder

    Joined:
    May 28, 2016
    Messages:
    7
    Likes Received:
    4
  6. Kyo

    Kyo Master
    Staff Member Resident Builder Builder

    Joined:
    Feb 27, 2014
    Messages:
    592
    Likes Received:
    503
    Very cool, What a excellent way to tackle this with the tools you have on hand. Reminds me of the research I did into Vacuum bagged honey comb carbon layups. Stressed panels ect. When correctly bonded this kind of plate can be extremely strong ! I would not have thought to try this with 3d printed parts. I look forward to watching as your project progresses.
     
  7. Steve Kingsley

    Steve Kingsley Well-Known
    Builder

    Joined:
    May 28, 2016
    Messages:
    7
    Likes Received:
    4
    Thanks @Kyo it's turning out pretty well so far. One problem on this first set is the counter bores for the flanged bearing and the screw heads on the X plate. I'm going to need to have a look at the X axis plate and see if the counter bores are needed for clearance or what. I was thinking about incorporating spacers and threaded inserts into the panel build up.

    Here's a picture with the holes drilled. On the second plate, I did drill through the first side using the printed core as the template. I then glued up the second face and drilled through from the first side. The second pic shows it screwed to the C-Beam, alignment was spot on! I didn't get chance to trim this second face as it got too late to make that much noise :)

    [​IMG]
    [​IMG]

    Cheers
     
    Kyo and JustinTime like this.
  8. Kyo

    Kyo Master
    Staff Member Resident Builder Builder

    Joined:
    Feb 27, 2014
    Messages:
    592
    Likes Received:
    503
    For the leadscrew bearings, This is much the same as laser / water jet cut plates and why I spec'd flanged bearings. Don't worry about the pocket or recess. Simply cut a 16mm dia. hole all the way through and the press fit / bearing flange will retain the bearing no need for milling the pockets.

    On the X-axis plates the rear x-axis plate does not need the counter bores at all. They only provide a little extra thread engagement with the length of screws I was using. The front x-axis plate can use either counter bores or stand offs. The purpose is to allow the Z-axis to sit flush with out interference from the bolt heads.

    edit:
    Yeah now I am thinking about it. that would work. Before gluing up the plates to the core. some heat set inserts like these "Here" can be inserted in the front x-axis plate layup. The bolts for the wheels can come in from the rear and screw into the plate to secure as long as the bolt does not protrude beyond the face of the plate no counter bore or spacers would be required to mount the z-axis assembly flush at all.
     
    #8 Kyo, May 17, 2017
    Last edited: May 17, 2017

Share This Page

  • About Us

    The OpenBuilds Team is dedicated helping you to Dream it - Build it - Share it! Collaborate on our forums and be sure to visit the Part Store for all your Building needs!
  • Like us on Facebook

  • Support Open Source FairShare Program!

    OpenBuilds FairShare Give Back Program provides resources to Open Source projects, developers and schools around the world. Invest in your future by helping others develop theirs!

    Donate to FairShare!