EAA 292 FACETMOBILE
The following page details the construction of a Facetmobile experimental airplane by a group of builders from the Experimental Aircraft Association (EAA) Chapter 292 located in Independence, Oregon. The original FMX-4 Facetmobile was designed and built by Barnaby Wainfan, Rick Dean, and Lynn Wainfan back in 1993 and made quite an appearance at Oshkosh in 1994. It is a lifting body aircraft; in other words it has no wings, the fuselage provides the lift. Even with the success of this design, no other aircraft were produced and no plans or kits were ever made available (with the exception of countless radio controlled model versions of the design).
Enter a group of active builders from EAA 292. This group decided to re-engineer the design (with Wainfan's knowledge of their project) and then not only build one flying example, but at least six. EAA 292 has a track record of "group builds" of aircraft, and this includes (13) Nieuport 11 replicas back in 2002, and (5) Airco/deHavalland DH2 replicas currently nearing completion. The goal of the re-engineer is not to redesign the aircraft, but refine its construction details... and then of course end up with six flying aircraft!
And this may be a good place to introduce the group of builders. They are Robert "Bob" Ingle, Ernie Moreno, Vince Homer, Denny Fuhrman, Bill Hoffman, Curt Anderson, and Bill Higdon. This page is being written by me, Robert Haines (hey, it's my website). Now on to the project.
Please note: It is my understanding that the Wainfans are not directly involved with this project, so please respect their time by directing any questions or comments only to members of this build group.
The basic structure of the aircraft is aluminum tubes connected with riveted gussets. The Wainfan prototype was hand cut and built, and because so, the exactness of how the tubes fit together was subject to hand cutting and grinding. Many of the connection points have multiple tubes clustered and nested, and almost none of the angles are simple right angles. In other words, this design is not one that can be built easily by hand and is very difficut to get all of those tube connections accurate.
And this is where I got involved. I'm a member of EAA 292 and I'm already building my own airplane. However, I have a CNC and love to take on new challenges, so I volunteered my time and my CNC. The hope was to use my CNC to cut the ends of the tubes with a profile that would allow them to fit together precisely. Once the tubes are all notched and predrilled, then use the CNC to cut accurate metal gussets. Hopefully, everything will line up, including all of the predrilled holes, and the tube fuselage will simple Cleco together almost ready for riveting.
The first step was to layout the fuselage in a 3D CAD model. In this fashion, the layout could be reviewed by the group, and then tube lengths and angles could be acquired from the model. Bob Ingle and Ernie Moreno took field measurement of the original prototype, and then Bob developed the CAD model.
The next big step was to develop the software to create the CNC cutting files. Using a spreadsheet, a master list of all of the nodes and their location in 3D space were listed. Then, a second spreadsheet page defined all of the tubes using these nodes. This tube definition included starting node, ending node, and then tubes which this tube needed to be notched to clear. This page also included information about where drill holes needed to be located. While Bob collected the node and tube information, I developed the Visual Basic code that calculated the tube coping profiles and CNC cutting files.
OK, it was a little untraditional to use the Visual Basic integrated into Microsoft Excel, but our goal wasn't to develop a distributable piece of software. By manipulating cells through subroutines, you can quickly see what your code is doing (the spreadsheets hold the input and output information and it becomes easy to see what happened). So code development was easier using Excel programming (although I'm sure I'll get booted out of any "cool kid club" for using such low-rent software).
The Visual Basic subroutine started with the node information and by use of the dot product and cross product (remember Vector Algebra?) calculated the angles and rotation between all of the tubes. Then, a coping path was developed by using a common formula for tube coping. There are many tube coping calculators on the internet, and a few sites post the math. I started with the basic formula that uses the diameters of the uncut tube, the cut tube, and the angle between them and provides a distance from the end of the cut tube at a specific degree of rotation around the diameter of the cut tube (whew!). Then calculated that for all 360 degrees of the cut tube. Then calculated that for both the inside and ouside diameters of the cut tube. Then, if additional uncut tubes were present, calculated thier inside and outside offsets with respect to the cut tube. Then compared all of the offsets and combined the deepest cuts for the rotation of the cut tube. Then added a cutter diameter offset. And when that was done, it provided a cutting path around the cut tube. I then generated a subroutine that created a GCode cutting file for a tube coping setup on my CNC. And so with one button, the guys could go from initial tube node information to finished cutting file.
The screenshot below show the initial information of the tubes with the resultant coping path shown on the graph. The two sides of the screen shows both the front end and back end of the cut tube. Using the cutting file, the CNC first drills all of the holes on the tube, then copes the front end, and then cuts the part free by coping the back end. Five minutes, one completed tube.
My CNC was setup with a fourth stepper motor for tube rotation. This held the back end of the tube, and the tube was positioned in a long precision groove cut in a plywood fixture. You may see this fixture along the left side of the CNC table:
After two weeks of software work, we were finally ready to cut something. PVC water service pipe was used to develop the process and create a scale model. The PVC cut nicely, but we had to insert a long 3/8" steel rod to hold the tube flat.
The next phase of work was to build a 1/3 scale model of the Facetmobile. It was noted that even at scale, the angles, notches, and gussets would be the same as the full sized prototype.
Below we have Denny, Bob, Ernie, and Vince starting the model:
Initially, we held the model together with tape and rubber bands, but it became quickly evident that we needed actual gussets. After a few gussets were cut from scrap aluminum, the model simply snapped together:
After a week or so, we had most of the tubes included in the model. By this time, every tube had to be cut two or three times due to changes or corrections: which tube was cut vs. uncut and orientation and location of drill hole locations. Coordination of the details was a massive undertaking.
The original plan was to go directly from 1/3 scale PVC to full scale aluminum first prototype. However, this is such a complicated project that it was decided to build a full size PVC model.