Re: String Trees, Or No?
In line with what Philtrum said -
There are several functions a nut must perform. The obvious first is to set spacing and height above the fret plane, and keep strings in place when plucking or bending. Next is to perform as a boundary point for the wave, and this is where problems occur on Fender style headstocks with no strings trees.
When a wave travels along a string and reaches a boundary point (nut/saddle/fret), there is potential for various degrees of the energy to be reflected back, transferred through, or transmitted beyond. If a string has suitable downward pressure and the rest point is broad enough, most of the energy will be reflected back, with of course some being transmitted through to the neck as well, but very little transmitted beyond to the strings beyond the nut. In my professional opinion, a tree-less Fender style nut simply does not have enough downward pressure to create a suitably firm boundary point on the upper strings, even with the strings wound low on staggered posts.
Ideally the string should see an immediate angle of at least 5° where it first contacts the nut. Less than this and the boundary point can become somewhat blurred, extending beyond the face by a slight amount (varying with amplitude and frequency) rather than seeing it as a finite point. Then once past the initial face of the nut this angle needs to curve slowly to a slightly steeper angle in order to keep the string in contact (in a flat slot the string will arc slightly beyond the end points and float over rather than rest on the base). Without this arc and a suitable increase in angle behind the nut to the tuners, you can loose a lot of energy as well as experience back buzz.
Fender has addressed this in some models with their roller nut and staggered tuners, but this really just masks the more obvious symptoms rather than addressing the core problem. The angle is still so shallow that downward pressure at the nut is insufficient to create a firm boundary point, and this combined with the very narrow point of contact allows a good deal of energy to be lost beyond the nut. Then to address the symptoms of back buzz and excessive ringing, they include a pad behind the rollers to dampen and mute this lost energy. For some players with a light attack and/or high distortion it may not present any issues, but is still far from ideal in my opinion.
Leo did not design these necks for any special tone or ideals of engineering. It was simply a way to cut cost/waste by being able to manufacture necks from standard, readily available 5/4 maple stock. String trees were then added for obvious reason of necessity. The E/B tree is a must in my opinion, and the G/D tree an option, which if omitted can work pretty much fine as long as the G slot is cut absolutely perfect. Often times this leaves the G string with no more than 7-9° total angle when wrapped low on the post, which can work but leaves no room for slop or error in cutting the nut slot.
I will often polish out the underside of conventional stamped string trees to minimize friction for players who use a lot of tremolo and bending, but once this is done there should be no more tuning stability issues with a string tree as there would be without. It's really something I consider a must have though, at least for a vast majority of players. Some players' styles and demands can get by without it if the nut slots are cut to extremely tight tolerances, but then again, I see absolutely no reason not to have one anyway.