Ted,
    Your question about tungsten carbide tips.  I have used ones off of saw blades as you are contemplating.  They can be easily removed by heating with a propane torch and then re-brazed on whatever your pendulum attachment position should be, if of steel.
    The research pendulum that I built several years ago performs well by using diamond on sapphire.  The diamond points were taken out of old phonograph stylus units and the sapphire flats are probably not too difficult to come by.  In my case they were surplused from old mass-balances (common to weigh chemicals in chemistry class before the advent of digital balances).
    The problem with the saw-blade tips is that they are designed to cut, not rock on a flat surface. The sharp points will dig into most surfaces.
    A pendulum that I have on the pier here at Mercer that works quite well uses ball-point pen extracts resting on sapphire.  I can't tell a significant difference between its performance and the one that used diamond points.

Angel,
  I like your thoughts about magnetic 'floaters'.  The first time I built a system to do the same system as the system described in the webpage you referenced was when I was a student more than 40 years ago.  The Jesse Beams Laboratory at the University of Virginia is named in honor of the professor who perfected high angular speed experiments by this means.  He studied the diffusion of atoms in solids by an elegant (very non-conventional type of centrifuge)-- spinning ball bearings, by means of a rotating magnetic
field in a vacuum, to a high enough angular velocity that the tensile strength of the metal would begin to be called into question.  I understand that he was able (in vacuum) to actually spin them until they flew apart!
    Things that hang below magnets are fascinating for reason of their 'defiance' of Earnshaw's theorem.  In the case of permanent magnets, there can be no stable equilibrium of the type demonstrated in this webpage.
   There is an example of a stable equilibrium that looks similar, that requires no feedback.  Serendipity by a single individual (not the prowess of a well-funded corporation, Geoff) figured into its discovery by my brother Palmer Peters.  It is called the 'suspension effect' of high Tc superconductors.
http://query.nytimes.com/gst/fullpage.html?sec=health&res=940DEFD7143CF933A1575AC0A96E948260
The flux-pinning that results when you add silver to the cuprate superconductors gives them fascinating properties.  Palmer worked with M.K. Wu sho supplied him with the samples and who was the first to make the first high Tc sample (Y-BaCuo).
   About your sketch, Angel:
Would it be possible for you to have the points contacting the magnets in a 'pull away' manner as opposed to 'push-against' one?  This would reduce the normal force and thus the friction.
   Randall