I use braided sleeving to cover the servo wires. I also melt down the ends with a soldering iron to keep them from fraying. It makes the links a little more stiff, but a lot stronger-I have not experienced any pinching when using this around my wires.
I C.
Radio Shack has a kit, they might have it in a local store. In New York there were also the small one man computer stores, they have it too. Besides that you can check if the store has a web site. Any place with a modding section should have something you could use (a larger CompUSA, but not say Best Buy), but you might have to settle for UV neon green or something if you go that route. 
Also expect to be ripped off.
yeah, Im sure its more in the stores. I think if it is sold though, it would come in a neutral color like black or something. Thanx anyways, ill check around at the Source CC (radio shack in Canada)
That non-fray looks VERY nice. Yea, the cheap computer stuff need heat shrinking on the ends. That also doesn’t seem to cost that much.
yeah, i read that and Im probably gonna get some some heat shrink too.
Hi guys, please feel free to post concerns. I encourage feedback! It might make more sense if you realize I am making this servo specifically for robot arm shoulders and elbows. Any other application it proves useful in is just icing on the cake. 
I rejected the elongated “C” bracket early in the process as it would have added more weight, and been less sturdy. The actual axis assembly only takes up about one inch in the current design. If I find additional gear reduction is useful, the added length of the motor will not cause problems. I think the overall width of this version will be about the same as if I had gone the wide “C” bracket way. Anyway it’s still early. Lemme build the prototype and see how it performs.
Despite this being an old topic and more than a little off topic, I would like to introduce you to my website and video of the robots seen here.
I hope you enjoy,
-Hive
I agree!!! 
I can see it: NBW National Biped Wrestling, aired from coast to Coast and most parts of Japan. 
what are you actually using to control those with? and are you using gyros for stability control or is it all open loop?
Thanks,
I will be competing in RoboGames in less than a week - so hopefully the bots will do well. I am pretty anxious.
Eddie, while I do have a 5-axis IMU in the bots they are also controlled by 2.4ghz ps2 controllers and bluetooth (both can be used at the same time).
After RoboGames I plan on switching to Gumstix (or similar processor) and having a wireless camera along with zmp math to control the movement (which will take into account the kalman filter). A lot of work ahead, but they are working really well right now.
-Hive
all that with just a ps2 controller? Your programming skills must be impecable~! 
i think these events should have an obstacle coarse competition too,
for those of us who dont want to trash our bots and to spur a building of
more capable robots,
Why do the robots get down on their knees to fight? BTW Anyone can video a robot kicking the crap out of a stationary RoboNova, why don’t you video the RoboNova fighting back? Would make for a much more enjoyable and believable vid in my opinion. 8)
Very cool. You’ve obviously studied Japanese Robo-One bots.
Also Jim mentioned, why is it on its knees? It seems like you are really stretching the rules. More specifically, this rule:
(b) The maximum length (size) of the foot must be less than 70% of the length of the robot’s leg. The leg length is defined as the distance between the ground to the axis that connects the leg to the body of the robot.
It might also have to do with lowering the CG, as well as increasing the footprint.
By kneeling to fight, the CG is lowered, making it somewhat less prone to tip, but by kneeling and spreading the hips, the footprint within which it balances is increased from the rectangle as defined by its footpads, to the rectangle as defined by its footpads and knees, which are also spread laterally by the hips. They go from a top-heavy arrangement balanced on a small base, to a somewhat more evenly-distributed mass, on top of a base that is considerably wider and deeper. Compare trying to knock over the leaning tower of pisa to trying to knock over the Pyramids of Giza.
Actually, the robots are fighting in a very low squatting stance and are not on their knees at all - just very close. This is, as Seamus mentioned, to lower the CG. Given that the feet are also made out of aluminum they tend to slide a lot on the arena floor - lowering the CG also helps this matter.
If this were a boxing competition then we would certainly have the robots maintain an upright stance, but the RoboGames have specified this as wrestling: “two robots must defeat each other by using any combination of pushing, wrestling, boxing and martial arts techniques”. We also have a “Mid Stance” which is half way between the low stance and a full standing position - this is seen one time as the RN has its gyros enabled (it takes 4-5 hits to talk knock the RN over).
And Jim, the RN is trying to fight back, but our humanoids have real-time control and move far faster; the RN simply can’t position itself fast enough to throw a punch. Moreover, I only edited the video to cut out the time when our humanoids waited for the RN to get up.
Anyway, I hope everyone was still able to enjoy the video.
-Hive