Open Source Aircraft Design for CNC?

[Jeffrey Meyer]

Got it.

[JNicol]

Hi Jeffrey, Send me an email to john_nicol at hotmail dot com

[JNicol]

Hi Everyone,

We are continuing our planning for the Open Source Aircraft initiative and would like to ask the community for assistance in filling the following roles to help us with this planning. We are after:

• Aircraft Designer/Aeronautical Engineer to assist in putting the groups ideas into action;
• Wordpress developer and administrator for the .org website.

If you are interested, send me a PM and I will send more details. Thanks!

Regards,

[spungey]

Sorry about the delay replying. It's been a crazy month.

We had a discussion like this over on homebuiltairplanes.com a while ago. The general consensus was that the OS model makes a lot less sense for hardware than it does for software, where anybody can test the latest build without any significant investment. With an aircraft design, it's much more difficult... who approves the latest "official version? And how do you approve a design change without building and verifying a prototype? But to wait for a prototype to be verified could take a long time....

The OS model makes sense, but you must first abandon the idea that anyone approves anything. Let me say it again. No .. one .. approves .. anything. (Except where approval is needed now.) When you go open source, then you also--to a certain extent--forego traditional ownership rights.

There's two ways to look at this. The first is that it's your A**! Ultimately you and you alone are the one who will pay if a corner gets cut, literally or figuratively. That means that you, and you alone are the ultimate arbiter of what's "approved." You don't want to contribute any design or part until it's know. Naturally a prudent builder would get some expert advice. :-) The second is the flip side. With dozens (for instance) of well-known rib patterns for various NACA or custom wings, you as a designer/customizer have more control with less cost. You are free to take an existing design and modify it without incurring the wrath of anyone except the gods of gravity. And that brings us back to the first.

In practice, the person who introduces some airplane design also specifies certain patterns, parts and behaviors when that design is contributed. Any changes to a part is a change to the design--with the same issues one encounters changing a known design now--you're making a new airplane. Hopefully in the OS world you test it and then contribute back to the community so others can reproduce and build upon your work. You could in theory gain some kind of intellectual property protection on some or all of what you do, but it'd be better for all of us if you kept the chain of non-commercial improvements intact.

Also, once enough tests and successes are demonstrated on a part, a design, a pattern, a grouping of parts ... then that part design, pattern or subsystem becomes a known quantity for the FAA examiner, making flight approval easier.

In short, OS makes a lot of sense. It's just ... different ... with hardware than it is with software.
--
Richard Johnson
open source evangelist (and now OS avionics)
http://makerplane.org

[JNicol]

Just to clarify as well, the idea with this initiative is that the "official" versions of the OS aircraft would have gone through prototyping and flight testing before releasing the plans. It would be the modifications, customizations and enhancements that are contributed that would not necessarily have been through that process. If these are accepted by the community and are tested then of course this is acknowledged and peer reviewed just like any other project. These changes would go back into the main design or be accepted as enhancements. The process is evolving as we are still in the early stages, but just be assured that any changes are not necessarily just submitted and then flown.

John

[spungey]

I left out a couple of important points from my post yesterday. (Sleep probably really does help.) :-)

First, while I did say "no one approves", I neglected to say, " ... but everyone attests." When the design is available and everyone can make, test, measure, and dissect it, then everyone has the opportunity (and in aviation, a duty) to report weaknesses and attempt to design better bits. It's like having a thousand monkeys pulling on your rib truss, and reporting weaknesses when something breaks.

I never meant to imply a callous regard to safety. I think that open source makes things safer. Safety is critical. Safety is an outgrowth of testability, measurability, and repeatability. Having a publicly-known and "anyone can make one" design for every part actually improves safety of the part. It's the same thing we discovered about encryption -- the public algorithms are all more secure than the secret ones because all of the dark corners have lots of lights shining on them.


Second is the issue of authenticity. Suppose I design a special aileron hinge (for instance) and I test it on my airplane. It works great and analysis says it should be good for 10,000 +/- 200 hours MTBF. Then I make that design, plans, and instructions available. How do you know the design you got is actually the one I sent out? After all anyone can tinker with it, right?

The answer is "digital signature." Anyone providing a part, design, etc. needs to provide a signature on the work. Techniques are known and available that prove that any document is identical in every respect to an intended original. This guarantees that (if you verify it) you have the right part. In essence this is approval. (Here is where I eat crow and say, "OK, the creator approves it ... and notes the precise conditions under which it works, how well test it is, etc.")

Related to this of course is your responsibility, when you choose to make and use that part, to actually run some numbers and make sure it will actually work with your particular airplane and not prematurely fail. Once you do, then you can report how well it works in your design and make life that much easier for the next guy.

Hope that clears things up, or at least provokes some discussion. :-)

[JNicol]

Hi Everyone,

It has been a while, but just to update people that have been following this thread. We launched "The Hangar Workshop" today on the MakerPlane website www.makerplane.org. This is a repository for open source aviation related projects. We have several open source avionics hardware projects kindly provided by Matjaz Vidmar. These need TLC to complete their documentation and some may require rework, but it is a great start to the database! We also have a placeholder for the MakerPlane v1.0 LSA. This repository is a free resource for anyone that wants to start up an open source aviation project and manage it with a team. It has tools including bug and version tracking, document and file storage, forum and task management tools and so on.

Also, we will be at AirVenture next month with our very own booth and will have a couple of forum slots to talk about MakerPlane! At the booth we will have a scale model of MPv1.0 as well as a CNC machine. Please stop by and have a chat with us!

[Kiwi ZK-CKE]

Just a demonstration about how CNC machining can build wooden aircraft parts - here a a video of parts for a DH mosquito rebuild being produced in New Zealand. If the technology can produce a 54ft span wooden wing, a small homebuilt should be a piece of cake!

[JNicol]

[QUOTE=Kiwi ZK-CKE;17574]Just a demonstration about how CNC machining can build wooden aircraft parts - here a a video of parts for a DH mosquito rebuild being produced in New Zealand. If the technology can produce a 54ft span wooden wing, a small homebuilt should be a piece of cake!

That is just awesome! That is exactly what we are trying to accomplish with the first design. Optimize it for CNC from the start and provide the appropriate CAD files ready to be converted to g-code for the specific CNC machine that people have. We are also looking to provide g-code for those CNC routers that we will be using in-house for those that have the same ones.

BTW, I am an ex-pat Kiwi......

[Brian E. Evans]

One of the difficulties of generating G code for different CNC machines is that not all machines are equal . Some use line numbers , others do not, some can use higher level codes, others reduce the inputs to X Y and Z steps in two of three thousanth of an inch, with files that are horrendusly long. There is no one size fits all here, which is really what is trying to be achieved.
Brian Evans.