http://www.youtube.com/watch?v=AvsyH4dtlTE

Like many of my projects this began with the chance finding of a lovely old brass bowl in a thrift shop. I paid a dollar for it and walked out thinking "I love the patina - it would make a great lamp..." and when I got home tossed it into the box of brass bits and pieces collected over the years from my frequent visits to the recycling plants, the junkyard and the thrift shops I frequent. I forgot about the bowl for nearly a year and then had a hankering to make a floating arm desk lamp in the Steampunk genre and re-discovered the bowl. This provided the impetus and I went looking through my boxes for other bits and pieces which could be re-purposed to provide key elements of the construction. Many of my projects start this way and in a sense, this is what makes them difficult to present as a how-to-do-it instructable. So much of this depends on what bits and pieces you may have to hand or can find, that even planning is no more than a vaguely realized "Concept Drawing". The actual details will vary depending on what components one can find, or, in hitting a problem, how one chooses to solve the problem.
I am a self-confessed amateur, learning as I go and am indebted to the online maker community for all the open-handed and generous sharing of tips and techniques from which I have learned so much.
The tools used here include a drill press, angle grinder, hand file, taps and dies and a large and a small soldering torch. I like the Steampunk genre because it allows me to use the materials I most like working with: well-seasoned hardwoods, brass, copper and steel, as well as using simple mechanisms such as springs, electro-mechanical or clockwork, rather than "black box" electonics. (Not to decry electronica - just not in Steampunk!) Aesthetically, It also allows a certain "industrial functionality" of finish, which makes the crudity of finish and the overall roughness of my pieces almost acceptable.
In all my projects I have two criteria: the piece must work as planned and it must meet my own aesthetic standards, what I call pretty functionality. Half the fun is in solving the problems as they arise and as there's never only one solution, the project could go off in any number of ways at any stage, which is what makes it interesting. I'd love to hear about anyone else's version of this sort of construction.
Note: To make the mechanism clearer, check out the video on http://www.youtube.com/watch?v=AvsyH4dtlTE shown above..

Step 1: Lamp Base

To carry the weight, especially when deployed in the extended position, a "floating arm" device needs a stable base, which should be quite heavy in order to prevent tipping over. I used a nice chunk of decades old Australian Jarrah, a reddish hardwood used in railway sleepers in the late 19th and early 20th century here in New Zealand. This piece I got for free from a local scrapyard and I estimate that is must be at least 80-plus years old. It has not been stained, simply sanded and then given three coats of Polyurethane, sanding with increasingly fine grades of paper between coats. The base has been hollowed out using a spade bit on a drill press. The finished space, which has to take the mains step-down transformer ( the light is a 12v DC Halogen) and the wiring for the on-off switch is rough, but serviceable. Would have been better with a router or a mill, but I don't have either of those.
The brass legs are made from blank brass rod - a scrapyard find, thickened at each end by soldering a matching piece and then grinding to a teardrop shape with an angle grinder. The construct was then drilled to take the threaded end of an industrial electrical contact ( For some reason, I quite often find these at the local metal recycler and always snap them up when I find them as they seem to be infinitely adaptable). The threaded pin was then shaped by filing in a drill press ( in this case acting as a turning device, like a poor man's mini lathe).

Step 2: Rotary Mechanism and Lock

Any balanced arm lamp needs to be able to extend and retract, but also rotate. In this case, this was achieved by using a brass quick-release hose clamp as the central rotatory component. Given a proper array of tools, I might have used a machined bearing with a broader surface to prevent the irritating laxity and hence, wobble, that comes with using such a narrow pivot, but one has to work with what one has available.
The lamp base is curved and the arm is a long lever, therefore it will fall away from a selected position. Accordingly a locking mechanism using a toothed wheel and a toothed pawl were devised. These are shown in the images attached. The arc of rotation of the lamp was limited by arresting pillars as shown and symmetry was maintained by using identical gas-line connectors to provide the base for the locking mechanism and the electrical wiring egress point from the base to the arm of the lamp.

Step 3: The "Skeleton" of the Arm

Modeled to some extent on the human arm, the skeleton of the arm required an upper arm, a forearm, connected through an elbow joint, and a hand, the actual lamp itself, connected through a wrist joint. Springs, the rigging discussed in the next section, then provided the muscles, which would move these joints.
The skeleton was put together using components found in my boxes of scrapyard finds.

Step 4: The Rigging or Muscles.

The springs, and balancing of the forces which their addition generates are what enables the arm to 'float". The Anglepoise lamp, invented nearly a century ago is a beautiful example of the elegant simplicity which a functional design of this form can take. The current design uses a multitude of springs, with cables and outriggers to direct the forces. Despite fairly wide searching I couldn't find any sources which could give me a principled approached to this aspect of the design. Consequently, I took a leaf from the Creator's handbook and used my own arm as a model, trying to match the muscle attachment points around the joints. In the end though, the actual points of attachment were determined empirically, or by trial and error.
The outriggers were added to help disperse the springs, for aesthetics and because they were fun to do.
About the only principles I can enunciate after all this messing about are these:

1.In order for a joint to "float" in a selected position, the forces trying to extend the joint, must match the forces trying to flex the joint.

2. Start by trying to animate or float one joint. The more joints you add, the more complex will the rigging become and true balance harder to achieve.

3. The arm should be heaviest and bulkiest at the base or shoulder and get lighter towards the hand. failure to do this will make the entire construct top-heavy and liable to topple over.

Step 5: Problem Solving

While it was relatively easy to join the skeletal elements, once the riggi ng was added, the constant force of the springs across the joints proved too much for some of the junctions and these gave way. The challenge was then to reinforce and support the joins while maintaining some semblance of aesthetics and functionality. This section shows the solutions to two of the junction failure problems: