The first step was to make a plaster casting of my head. On this I sculpted the basic features for Karl's head using oil-based clays: his muzzle, nose bridge, eyes, etc, and I made sure there was space above my head for mechanisms.

Next I coated the clay with a thin layer of liquid latex and let it dry; this provides a separator to keep the fibreglass resin from reacting with the clay.

The next step was to use fibreglass resin and gelcoat (for thickness) with small fibreglass patches to build up a 3-5mm-thick "skull" over the model, let it cure, and dig out the clay.

I then Dremelled out holes for the ear mechanisms and the area around the eyes. I cut and installed aluminum brackets for mounting the antlers and the ear servos. I cut up some Dr. Scholl's in-sole pads and hotmelt-glued them in in layers to make the helmet fit snugly and pad my head properly.

Next I went to a hobby store and bought a clear plastic sphere used for making Christmas decorations, and painted on the inside Karl's eyes; first the white sclera, then the varyingly-shaded irises. I used masking tape to mask the pupils; they do not get painted and are clear. These were mounted on steel brackets and screwed into the skull.

Next I bought a tiny black and white video camera, the PC-14XP, and I sealed it in a shrinkwrap sleeve to keep out moisture. Then I bought a small pocket television, making sure to find one that had an external video input, and took it apart, re-soldering the boards together in a geometry that would fit inside the muzzle of the skull. Each board was encased in a hotmelt glue-sealed mylar casing, again, to keep out moisture generated by breath and perspiration. The camera was mounted inside the left eye and the TV monitor was mounted in the muzzle.

Next I sewed a stretchy "sock" of lycra to perfectly fit the head and neck shape, and made a foam "neckbrace" to pad my neck diameter out to that of my head so that his neck would be uniform. I carefully cut pieces of white and brown fur and hotmelt glued them to the "sock" so that there were no gaps and so the grain of the fur lay in the right direction. I then put a 1/4" clipping guide on a pair of electric clippers and shaved the fur down uniformly short on his whole head, letting it be longer farther down his neck.

The ears were based on the plastic forms used by taxidermists-- they come pre-shaped just like a deer's ear, and are available from Jonas Supply in Broomfield, Colorado (their catalog is amazing). I carefully stuck fur to them and mounted small aluminum brackets in the bases that screw down to the servo shafts.

Next, enlisting Lots Of Help from a friend (Hi Persephone!), we sewed the velvet cloak and satin tunic. This took a lot longer than two lines can sum up and I would never have been able to do it without her help. :)

The custom-built leather gauntlets, the white knight's belt (on the backside of which is mounted the computer and power systems, described below), and the spur straps were all crafted by a very talented leather worker, CJ, of CJ's Leather in Denver, Colorado. CJ does phenomenal custom work quickly and at a very reasonable price.

The sword, which really doesn't even have a blade and is actually bolted into the scabbard to quell the fears of airline and ConFurence security, was crafted by a friend and talented sword builder, Doug Houdek.

The neatest bit in the costume is the Animatronic Sequencing Computer. For as long as I can remember, I've wanted to make a robot or animatronic toy. I spent months researching various analog options and I finally hit on the idea of using normal hobby R/C servos and a digital controller system-- the ACS. This is an amazingly simple 2-processor circuit that allows the costume to display easily selected sets of emotions through discreet finger movements.


NOTE: the 6VDC power supply must be able to supply a good deal of current, or else the computer and the servos must have separate battery packs. This is because when the servos are moving and loaded, they can draw so much power that it resets the computer. This Is Bad. :)
Though the design shown only uses 2 servo-motors, the ACS is capable of simultaneously driving up to EIGHT of them. It is also possible to increase the number of hand-switches (described below). The software is currently configured to drive 4 emotions, with 6 settings each (every three seconds the computer chooses a new setting within the emotion bank to keep the face active and "alive"), and 2 motors. All of these are customizeable in the BASIC code that drives the ACS.

To operate the ACS, just turn it on-- its EEPROM memory keeps the program and settings stored forever even when the power is off. The fingerswitches cause the ears to move to different emotion settings. If the box is opened and the "Next Emotion" button (see below) is pressed when the ACS is turned on, the computer enters Program mode.

Click here to see the code listing.
Parallax sells the PC- or Mac-based development system for $100.

The switches to the right of U2 are the red pushbutton switches shown on the board above. They allow new animation sequences to be programmed into the ACS at any time. They represent, from top to bottom:

1 Servo Position + 2 Servo Position - 3 Next Servo 4 Next Setting 5 Next Emotion

The switches to the left of U2 are hidden in the fingertips of the left-hand gauntlet. By closing the switches in different combinations on the fore- and middle finger, different emotions are selected. Clicking the thumb button acts as "Enter" and engages that emotion. I've programmed the combinations as follows:

Middle Finger Fore Finger Emotion OPEN OPEN Normal, relaxed expression. Ears forward and attentive. OPEN CLOSED "Look right" expression. Left ear forward, right ear back. CLOSED OPEN "Look left" expression. Left ear back, right ear forward. CLOSED CLOSED Angry expression. Both ears flattened back.

Here I'm programming sequences into the computer.

I would also work to incorporate more moving facial features; one way to do this is to use smaller servos. I used Futaba servos rated at 42 ounce-inches of torque. They make a much smaller model that delivers around 35 ounce-inches, though it is about $65 instead of $20. This smaller model would still have enough torque to drive most any well-designed facial motion, and having a smaller form factor, more servos and mechanisms would have fit in the helmet.

And the single hardest thing about animatronics, I found, is designing and building good mechanisms to convey the motion from the rotating servos. One of the keys to it is properly designing the outer skin-- be it fur or latex or whatnot-- so that that motion can be seen and be realistic.