Poker Flat Research Range (PFRR), managed by University of Alaska Fairbanks is working on several unmanned aircraft project.  One of them is converting surplus Raven UASs to run on the APM 2.5 and equip them with an extended battery pack. The Ravens will be used by researchers in Alaska for coastline, environmental and wildlife studies. 

As a proof of concept, I designed and built this carbon fiber nacelle for my undergrad research class. It would cover the APM should it be located on top of the fuselage to make more room for payloads. 

PFRR had a point cloud of the raven fuselage generated from a NextEngine 3D laser scanner.  From this model, I designed a mold that accurately met the surface of the fuselage.  After adding supports to the mold so it would not deform in the vacuum bag, I printed it on our Fortus 250mc 3D printer.  When the mold was finished, it was prepped with a mold release agent and 3 layers of carbon fabric were cut and infused with resin.  After curing in the vacuum bag, I removed, trimmed and sanded the part. 

On this first iteration, the part fit very closely with no large gaps.  Any gaps can be attributed to the rough trimming which will be simplified in later designs by adding a flange to the lip of the mold. 

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones


  • I printed the mold at .254mm layer thickness.  If i had done the fine at .178mm there would be less post processing work.  With more sanding and filling and it should turn out as smooth as any other mold. For mold strength, if I were making more than one part, I would print a skirt around the sides of the mold, so that it could be turned upside down and filled with an epoxy or cement.  This way you save on material by printing a thin mold, but have the strength from the filler material you pour in the bottom of the mold. 

    If I had been more careful when removing the part I could have saved the mold, but I scarred it a bit and cut some of it away with the dremel.  Again, more time spent on this would have made it better, but it was a one shot deal.

    I like your project. Im working on a similar design for a student project. We are designing (and building) a research UAS for coastline mapping and marine wildlife surveys in the Aleutian Islands off the coast of Alaska.  

  • Thanks for all that. My prototype fuselage was just laminated over foam, which was later removed but now want to make a mould.

    If I were doing the plug out of wood I would expect to get a near perfect gloss finish after varnish etc, so would hope to get the same after finishing a plastic printed part.

    I expect the carbon part is stronger than the printed mould so it would probably get damaged easily during separation. Not a problem for a one off I suppose.

    This is the project.






  • The ribs caused by the layers of the 3D printed part aren't big enough to transfer to the fabric, but do fill with epoxy. This made it harder to remove the part from the mold and made it necessary to sand the surface of the carbon part, hence the dull finish.

    I added a few layer of mod podge to the mold as a filler then coated it with release wax. That didnt work very well. I would recommend more filler and sanding.  Also, I would make a positive mold so it would be easier to sand the mold surface. Also with that, the 3D printed surface would not be your finish surface, if you can vacuum bag it without getting wrinkles then it should turn out good.  That depends on the part, and how much you care about your surface finish though.

    when switching from positive to negative molds, remember to do a face offset in your model for the thickness of the fabric before you print (if you're really close on tolerances).

    something else to explore, print a positive and negative mold and compress the fabric in between.  Put them in a press or clamps instead of vacuum bagging.  since its 3d printed, the surfaces should match up perfectly, but sanding might mess that up.

  • I was wondering about 3d printing to make a plug for a mould. Having not handled a 3d printed part I was not expecting a great surface smoothness so would have expected to have to sand, prime, paint, anything to be used for a plug.

    Was this mould ready to laminate on straight away (after release agent?) or did it require smoothing?

    The surface finish of the carbon part looks a little dull but not bad.


  • Moderator

    Sorry for the delay, we added a little shelve inside where the APM gear sits. It's accessible from under the wing.  Wish I was a non-profit institution so I could get  a real raven or two...  Too Cool!

  • This part will not be used. In tests, the apm worked just fine on the inside. I was looking for project ideas as it was the end of the semester and carbon fiber looks good in presentations. This project was only my proof of concept that well fitting composite parts can be be designed from the point cloud and their mold 3d printed.

    The ravens were given to us by the air force for research.  Most of them came in kits in which one or two parts were broken. The good parts have been set aside and we work with some of the more "used" ones for prototyping.

    Brian, what do you mean "a platform added to the raven"?

  • Moderator

    where did you get the surplus raven?  A new system is a lot more than 45k!  Looks pretty beat up.  We put the APM2 on a platform added to the raven which sits right below the main wing.  Still have a ton of room for payload and would not need to have it outside.  If you do build in carbon, IMO it's pretty much over kill for where you're putting the APM.  On top like that it's a lot more protected by the rest of the platform so all the extra protection offered by the CF may not be worth all the issues you'll have from it shielding signals.  FG will work just as well and be a lot easier to work with the mold you made.  

    Looks good though.

  • You could always do it in fiberglass and then paint it if it interferes.  Carbon does look good with that holographic sheen.

    I'm trying to design a 3D printed wing tip mold.  Shapeways is picky about things coming to a point, has to be thick enough and that kills the trailing edge and the very tip.  My solution is divide the foil in 2 halves and add thickness, but that adds exponentially to the price.

  • I wasn't sure just how into the UAV scene you guys up there were, but this is impressive! Kinda makes me wish I wasn't the only one pioneering UAVs down here in Anchorage. Oh well, someone has to do it, I suppose! Keep up the good work!

  • Now that is a great use for a 3D printer! :-) Awesome work!

This reply was deleted.