Open Source Aircraft Design for CNC?

[JNicol]

Hi Everyone,

I wanted to start a conversation about the potential of an open source aircraft design that could be built using CNC? With 8' x 4' CNC machines being relatively available (one kit can be purchased for $2,000 - $3,000) and CNC services like Ponoko that can be used to outsource, it appears to be a quick and cheap way to get an aircraft built....? My biggest problem with trying to fly is the affordability of building and owning an aircraft. I love the EAA magazine and Homebuilt magazine etc, but my eyes glaze over when I see kits that can be completed for "only" $40,000 etc. I realize that there are cheaper kits, but I was thinking about one that had plans that could be downloaded and built using a CNC machine to speed the construction up and use relatively inexpensive materials. My eyes lit up when I saw what was being achieved by open source projects such as Wikihouse: http://www.wikihouse.cc/

There is also work being done on the Volksplane to create CAD drawings of the plans in Yahoo Groups. I was wondering if there were any collaborative efforts out there to generate an open source aircraft design specifically for CNC? If not, why not! I can imagine downloading a design and cutting out the parts for construction like a giant meccano set. You could probably cut out the parts for an entire small aircraft in a few hours! I can see a group of builders purchasing a CNC machine between them and mass producing the parts for a group build. It would be really great if we could get an ultra-low cost aircraft in the works that could get more people building and flying! If there is interest, would anyone like to start something like this as a movement?



Regards,
John Nicol

[spungey]

I am 100% behind all kinds of OS efforts. We really are an OS community and should share with like-minded ones.

Have you checked into availability of CNC machines at TechShop locations around the country? Also, I wonder who would be doing the 3-d modeling and what data formats will be needed.

[JNicol]

Well, the 3D design, modelling etc would be part of the open source effort. It might be that perhaps an out-of-copyright design could be re-purposed? EDIT: actually maybe an updated Fly Baby would do the trick? The work would probably have to go into re-designing some of the components to make them easy to cut and assemble and perhaps create tabs and slots for pieces to fit together.

The formats would be standard CAD, probably dxf or whatever, with output to CNC file formats such as g-code, stl and so on. I am sure that there are CNC machines available at various Maker-spaces around the world that builders can send the files to for manufacture. The Ponoko service is only one.

Here is a picture of work that is being done to convert the Evans Volksplane into CAD:



So even this could be turned into CNC-ready files, but it hasn't been optimized for CNC as it would still be built in the traditional way. I think that there is an opportunity to make the designs more even more builder friendly. It would be neat to have a completely open source design that could be built upon and improved for the homebuilder community. I know that there are some talented designers out there and well given that there are 175000 EAA members, even a tiny .5% of interest would be almost 900 contributors.

[Eric Page]

This is something I'd be very interested in, at least as a builder. I lack the engineering knowledge or CAD skills to be a design contributor, but it certainly sounds like a great idea.

[Eric Witherspoon]

I'm having trouble envisioning exactly what you're asking for / thinking about trying to do.

So split it up:
Looking for the lowest cost.
Looking to use CNC for some reason.
Looking to generate low build time.
Looking to have builders work together to build multiple copies, with some benefit of that to be defined.
Create a new design that meets ? objectives...

My impression is that "open source" for a whole NEW aircraft design is impossible. Nobody's going to agree what the objectives are, and unless there's an arbitration board (or Office of the Chief Engineer) or the like, it will just diverge in a bunch of different directions.

Besides, if creating a new design isn't really the main objective of your inquiry, why not support one of the plans vendors out there? They worked many, many hours, built test parts, flew stuff that didn't work so well and worked it out - all to give you the rights to copy their success for just a couple hundred $. I'm sure they'd be thrilled to find out you have gathered together 10-20 like-minded individuals willing to stand on their shoulders to get into the sky.

[JNicol]

OK, so every idea starts off with a discussion and then we start drilling down to details. I am not trying to boil the ocean, as I said "I wanted to start a conversation".

If you read my post again, you will see that I said perhaps it would start off with an out-of-copyright design, or something like the EAA fly baby. That way it has already been proven and there are no major design hurdles. The re-design comes about with making the parts readily machined by CNC and able to be put together more easily. For example, instead of a rib being made up of 20 separate parts, it is cut out as a single piece from ply. Also, why CNC? I gave the example of the wikihouse. I have built my own CNC machine and it is a very quick way to fabricate parts. Saying that, I do appreciate the devils advocate to challenge and question as it makes us consider the details and alternative ways ahead.
Here is a link to a CNC kit that I think could be built and used by the average homebuilder: http://buildyourcnc.com/CNCMachineKits.aspx

I agree that starting off by trying to do a new design is crazy, that is why I suggested replicating something already designed. Having a design committee and so on will create a million dollar composite electric, solar powered truck built with recycled milk cartons. I am not sure having plans vendors is the way ahead unless they make the design available as open source. People can still make money from open source by providing the parts, printed blueprints, consulting and whatever. There are many companies that are making money from open source, so the business model is proven. I believe that it would also allow improvements and updates easily available back into the project.

So to reflect back some high-level objectives (but again, not trying to advocate any one idea this is just a discussion to see if people are interested):

1. Low cost. I don't know what the parameters are yet, but I would be hoping that it would be under $4000 minus engine and instruments as a first stab in the dark. I note that the fly baby needs about 10 sheets of plywood plus spruce, metal, hardware etc, so probably fits in that price range, but I think that redesign for CNC ply manufacture could bring that BOM down. Even if the BOM didn't come down, perhaps the cost savings are transferred to lower build time?
2. CNC, so that parts are easily manufactured. It would also mean that parts are consistent with tight tolerances, can be replicated over and over, and customized as required. (Also means that a builder can get the parts from multiple sources, not just one manufacturer.)
3. Low build time. Maybe under 100 hours for the build minus engine, instruments and perhaps fabric covering, but again, depends on the design. I am sort of hoping that it would all fit together like a big constructor set and even easier than the majority of kits out there now. Again, the wikihouse concept in execution.
4. Type? Maybe something that would fit into the Ultralight or Sport light category. My personal preference would be a two seat side by side. That would immediate cut out the fly baby, but maybe start off with a single seat existing design like the fly baby and then move to something else if it is successful.

Thanks for listening folks.



John Nicol

[rwanttaja]

Basic question I have, is can you reliably use CNC machining with wood parts? All of the CNC stuff I've previously seen have been with aluminum. Can these small "home" CNC machines carve a spruce wing spar 13 feet long?

John mentions the Fly Baby as an example of a potential open-source aircraft. Like many older wood homebuilts, the wood parts for the Fly Baby are often a "cut to fit" design rather than "cut to this exact shape and dimension." To be able to CNC them, someone would have to basically build an aircraft while taking careful dimensions for the CAD work.

Folks in the Fly Baby community have already generated CAD files for most of the steel and aluminum parts, but of course, the main structure is wood.

A better candidate would be the Murphy JDM-8, an all-metal Fly-Baby-Class airplane. However, there is usually some resistance to single-seat airplanes. The Thorp T-18 might be a better pick.

Ron Wanttaja

[Eric Witherspoon]

For your $4k airframe target, my guess at this point is riveted sheet metal will be the lowest cost. In large part because if it's 6061, it doesn't have to be painted. This keeps build time, weight, and cost down.

The other reason is, I think that $4k is the starting point for the fabric and finishes of a fabric-covered design, unless it's an open-tube-fuselage (like some versions of the Belite) with only the wing/tail being covered. So with the desire to not look excessively cheap, or not look like a pod-and-tube design, but to have a "real airplane" shape, but still fall into a "low cost" category, riveted metal will be the way to go. (See, I'm imposing my desirements - pod-and-tube may be more the trade you're looking for...)

This also falls right into line with getting some benefit from CNC - doesn't have to be matched-hole, but at least the skins can have all the pilot holes in the right places, then just clamp it together and drill through. Pod-and-tube will be a bunch of tubes and cables - not stuff that CNC is going to help with much, though the interfacing brackets may be done more quickly.

Blind rivets are quick, though more expensive than driven. Such a design, though, could allow the builder to choose - lower cost or quicker/easier.

So, here's where your $4k goes:
$1500 for 6061 aluminum sheet and angles.
$500-700 for pulled rivets (much less for driven).
$1200 for nuts/bolts/washers/cables/cotter pins/castle nuts/fairleads/wheels/tires/brakes - really. And I'm leaving out the hundreds for all the stuff that goes between the firewall and the engine...
$300 worth of steel for engine mount, control sticks, control pushrods, etc., or $2000+ for someone to weld it all up and paint it for you.

Then we still haven't got:
Couple hundred for a canopy bubble / windshield (though it might be open-cockpit w/ just a windshield)
Couple hundred for a cowling (though you might do a nose-bowl and sheet metal sides)
Couple hundred for wheel pants (maybe concede it's too slow for these to matter much...but they look so good...)
Couple hundred for gear legs (though the BK-1 has an interesting configuration of just Al bar stock with some custom axle weldments)

Anyway, say you finish off those above "non bare airframe" costs for another $2k.

Then there's an engine. Say this thing is really light/tiny (single-seater for build time, reduced parts count, and cost, of course). Even a 2-cyl VW-based engine is $3-4.5k, and 4-cyl VW's are in the $5-7k range. Pick a number - $5k for engine. Ok, we're at 11k now, and it still doesn't have a fuel tank. Or a prop. Oh, and electric start might make it more popular (add another $1k).

So, fuel tank, prop, panel, radio, seat pad, antenna, engine monitoring stuff - easy to spend another $3k there and still be very basic.

So you're looking at $14k bare-bones, no electric start. And you said you wanted a 2-seater.

A used Rans S-12 is starting to look pretty good. Quicker to get (just write a check), known performance, factory behind it for support...

But it kind of makes me wonder if there might be a way to get plans for that Murphy...

[JNicol]

Basic question I have, is can you reliably use CNC machining with wood parts? All of the CNC stuff I've previously seen have been with aluminum. Can these small "home" CNC machines carve a spruce wing spar 13 feet long?

Hi Ron,

Yes it can cut wood very well.... the wikihouse project I quote is basically 4 x 8 sheets of baltic ply. The same machines can also cut aluminum, plexy sheet and whatever....

Here is a link to a shopbot cutting wood on youtube (a 4x8 sheet used to create a workstation). Lots of videos there actually of this type of machine. http://www.youtube.com/watch?v=e6965cKlNqA&feature=related

The link that I gave above for the CNC kit, is itself cut on a CNC machine out of baltic ply and obviously needs to be of high accuracy in order to create another CNC machine. These particular machines have an accuracy of 0.0007" and it is made out of ply......

The bed of the machine is 8' long. You can cut longer pieces by indexing them. So you could cut a 13' length, or longer.

John

[JNicol]

For your $4k airframe target, my guess at this point is riveted sheet metal will be the lowest cost. In large part because if it's 6061, it doesn't have to be painted. This keeps build time, weight, and cost down.


So you're looking at $14k bare-bones, no electric start. And you said you wanted a 2-seater.

Well I think I would like to challenge the conventional thinking and push the boundaries a bit here. Maybe there is a hybrid solution of wood and aluminum I don't know. I do know that the CNC systems I am thinking of are capable of machining both.... I think it is a matter of adapting a design with this new technology in mind. I link swinging a hammer like most people in the forums, but I am also lazy I guess. If I can spew out 10 ribs in half an hour instead of in a month, then I would like to do that. Many kits are manufactured using CNC already, but a part you buy already machined might cost $200 when it is comprised of a $5 piece of metal and 10 minutes in the mill. Nothing wrong with that of course, but if I can now build my own 4' x 8' CNC for under $3000 (plus computer) and I could manufacture 99% of the parts required for my aircraft in maybe less than a week, I think that is pretty powerful! Having a really easy step-by-step construction plan would be the other really important element of course. IKEA, but with aircraft.... and without the annoying little allen keys. Yes, I said I ideally wanted a two seater, but I did also say that starting off with a one seater would probably be easier initially to prove the concept. I appear to be repeating myself....

There are no kits that I am aware of that include the CNC files to allow you to do this, so I thought that this would be a neat thing. I would personally prefer Open Source so that we are a little bit more in control of the destiny of it and let any builder that wants it, build it for themselves. There will be ample opportunity for people to make their money along the way if they want to I am sure.