Thanks to all in AeroQuad and AduCopter communities for their inspiration and ideas. I submit this photo record of my build hoping that it may help others.

Here are the custom machined aluminum parts for my frame. I used Alibre Design to capture the design and create part drawings. I am fairly new amateur metal worker and I manually machined the parts in my home shop. The arms are made from Home Depot towel bar material, quite strong and light.

My hub design was inspired by the design seen in this quad: http://forum.mikrokopter.de/topic-15135-1.html but I have corrected a flaw in that design. You can see in that design that when the hub halves are bolted together, a good, tight, consistent fit between the arm and the hub is dependent upon tight tolerances between the arm and the cavity formed by the hub pieces. With normal variances in the size of the aluminum tubing, these tolerances cannot be maintained. In my design, each arm is individually captured and clamped into its hub slot by a small piece that distributes the bolt pressure over the end of the arm. It is very tolerant of arm material size variations. A possible advantage of this design over the more common “sandwich” designs (like the “official” arducopter frame) is that an individual arm can be removed and replaced without loosening the sandwich and affecting other arms. In any case, this design is super strong and rigid and I like the way it looks.

Here is the motor end of the arm showing my leg mounting design and how it makes for a strong motor mount. The leg mount slides into the end of the arm. A bolt passes all the way through the arm, also passing through the leg and motor mounts. The bolt can then be tightened, squeezing the arm tube securely against the leg mount without worry about crushing the aluminum tube. I am using the mounting piece that came with the motor and I hope it proves strong enough. If not, I will design and machine a custom piece.

Here is the assembled basic frame. The weight at this stage was about 660g. The legs are .125” aluminum and are stronger than they look. They slope down at a 45 degree angle and are long enough that they extend horizontally about an inch or so farther than the reach of the 12inch props I am using. This provides prop protection if the copter approaches an obstacle horizontally. The legs also provide a stable platform that will not tip over if I encounter the uneven motor starts at take off that I have seen documented on the forum. The arm/leg combination does provide a pretty long lever arm that might invite damage in a hard, one-legged landing, but I mitigate this as shown below with my “bow” approach.

Here is the completed copter ready for initial test flight. With its 4400maH battery, the ready-to-fly weight is 1455g.

Although kind of hard to see, this picture shows the bow design achieved by tying the end of the legs opposite each other together with 4 runs of high test fishing line. This should greatly reduce the likelihood of bending the legs, at least in one direction and it lends a sort of shock absorber action to the overall frame. The dabs of paint on the props were applied when balancing them using a magnetic prop balancer. The electronics pod is plastic ware from the dollar store. I have seen other designs, e.g., using CD packages, where the flat part is mounted to the frame and the pod part is fitted over the electronics after they are all wired up and secured somehow. I prefer this approach where the pod is mounted to the frame, providing an enclosure that the electronics can be placed into and then it is closed up by putting on the lid as in normal use. A hole in the bottom of the container is aligned with the hole in the middle of the hub for the ESC control wires to pass through.

This detail shows the foot approach using rubber hemispheres that are actually “popper” toys from Michaels attached with cable ties. This view also shows how the fishing line is attached.

This view show the battery mounting approach enables easy battery swap. Bolts inside the pod hold it to the frame by screwing into the threaded spacers that hold the battery mounting bits. One of the ESC control cables can be seen passing through the hub center hole.

This view shows how the motor power system and 30A ESCs are installed. I use a split power system with separate batteries for the ESC/Motors and the electronics pod. Only the ground and control signal wires from the ESCs are connected to the Mega while the +5V lines from the ESCs are not connected. With this approach, you can power the pod electronics from the either the pod battery or from the USB connection when loading software or configuring with the configurator (pod battery not connected when USB connected). You can also connect the main battery to the ESCs when the USB is connected without worry of “blowing your outputs”. This way, during testing, you can run motors just fine while the USB is connected. The power distribution from main battery to ESCs is via a harness I soldered up using 2 screw lugs for the center connections of the +V and ground “stars”.

The fishing line is also visible at the in the picture.

Above the battery, you can see the battery voltage monitor affixed to the bottom of the pod. It plugs into the main battery’s balance connector. The battery connects to the power harness with bullet connectors.

I mounted the electronics in foam packing inside the pod as shown here. Connecting the connectors shown applies power from the pod battery to the +V input of the Mega. Power from the Mega’s voltage regulator powers the RC receiver via the Mega-to-receiver servo type connections.

This view shows the IMU and its connections to the receiver, a Spektrum 2.4Gz unit. It shows how I mounted the magnetometer directly to the IMU as shown in the arducopter manual as an alternative configuration. After mounting it this way, I searched through the alpha arducoper code and libraries for the “set_orientation” calls that are described in the manual but could not find them. It is not clear to me if the orientation of the magnetometer shown in my picture is correct for the alpha code or not- I have seen conflicting information about the “standard” configuration of the magnetometer in relation to the IMU.

From the other side, you can see the 610maH pod battery peeking out from under the board stack. Both the receiver and pod battery are tucked into their own cavities in the foam.

This picture shows the Mega at the bottom of the board stack and gives a better idea of how the electronics are enclosed and protected by the hollowed out foam. The foam pieces were those that the motors came packaged in. I was originally intending this to only be a temporary approach, but after seeing how it came out and how secure and protected everything in the pod is, I may keep it permanently.

Another view showing everything all buttoned up. The wires to the left of the pod run from the main receiver to its satellite receiver. On the right you can see how the +5V lines from the ESCs were removed from the servo connectors that connect to the Mega and covered with heatshrink. I did this rather than just cutting them so I could return the ESCs to their original configuration if needed.

The prop mounting hardware is shown here. The lower two pieces came with the motor but I machined the doughnut piece at the top on my lathe. I made this piece to avoid having to ream out the prop hole to fit the non-threaded portion of the motor shaft. The next picture shows how these parts were used to mount the 12X3.8 props.

First test flights to follow!!