On the face of it, roller nuts seem like a good idea, the principle being to reduce friction to a minimum. In practice however they rarely work as perfectly as predicted...
One of the early efforts was the Wilkinson roller nut. This had two sets of rollers, one front, one back, the string passing over the front ones and under the rear ones (except for the Jeff Beck version which had a cutaway on the E, A and D strings and just one roller). These were fitted to the the Fender Strat Plus range in the early nineties.
While these nuts occasionally worked they went wrong far more often. Rollers frequently suffered from rattling and buzzing to which the only solution was to impregnate it with a heavy engine grease of the type used to lubricate vehicle drive shafts. My concerns about the dermatological effects of this led me to use Vaseline as an alternative however on occasion neither worked and one had to resort to whacking the little soft metal plug that held the rollers in. That stopped the rattling but also caused the roller to stop rolling. Just how ineffective the roller action had been was measured by the fact that this rarely had any effect upon the tuning stability of the guitar. An additional problem was that the strings ran through deep slots in the nut's frame which offered no lateral support of harmonic damping so despite being a fair sized lump of stainless steel the effect on sustain and tone was actual detrimental. The nuts were sized for dedicated gauges as well so that if you dropped a gauge the string would develop a hollow buzz as it rattled around inside the slot; if you went up a gauge the string would jam. a friend and I once calculated that the nut slot of the roller nut offered eight times more potential friction than a conventional nut...
The Wilkinson nut was eventually replaced by the LSR roller nut. Sadly the LSR roller nut was an even poorer piece of design and is an example of what happens when someone who doesn't really know very much about how guitars work gets involved in the design process.
The idea behind the LSR nut is that the string is supported by two captive ball bearings which support the string and define it's sounding point. Bingo! no friction...
Unfortunately, no harmonic damping either.
An important function of the string is to support the string in such a way that the pulse created when the string is plucked is reflected back to create resonance and harmonic damping is the key to this. If the pulse passes through the nut instead then energy is dissipated in driving resonance between the nut and the tuner in the area which I call the "headstock harp". As well as diminishing sustain this creates enharmonic overtones which are fed back into the sounding length of the string causing it to sound out of tune or have an odd tone.
The designers' solution to this was to place a little rubber block behind the ball bearings to act as a harmonic damper however in so doing they restored all the friction they had hoped to get rid of by using the ball bearings.
Factor in the lack of individual height adjustment for strings, fiddly fixing screws and a tendency for the top string to slip through between the ball bearings and bottom out on the first fret, the cost of the part and the fitting and you might as well have just got an expert luthier to cut a conventional nut properly first...
Find out what I think of compensated nuts and the Buzz Feiten system here:
a critique of the principles behind the Buzz Feiten Tempered Tuning System
