This discussion is focusing on melees.
The general rule of thumb: Taking multiple toughnesses on a toon follows a law of diminishing returns. The idea being, that since 1 toughness opens up the enhancements, the first toughness = Total (HP from toughness + HP from Enhancements), and therefore taking the rest is a very quickly downward bump to your HP pool, or is considered no longer worth it.
Thoughts on this:
1. One Issue. Looking at the extra HP from an in-the-box only standpoint like this precludes the relative values of toughnesses to different classes. So for example, a 14 con rogue is going to gain a bigger percentage boost to total HP than a 14 con ranger, since the ranger is getting two extra HP/level already anyhow, and therefore has a larger total pool to begin with.
Caveat:
Any melee that is going to sacrifice a significant amount of offensive feats for HP is going to gimp themselves. It doesn't matter if you have the survivability of a mountain, if you can't break through the wet paper cloth separating you from your enemy. I am taking this as a given that an 8 Str intimitank is worthless in most content regardless of its defenses.
Fun Math:
First, why not take the toughness enhancements as a build given? I certainly think melee builds that don't take toughness at least once are rare enough that this could be considered "assumed" HP just as most people assume the build will wear a Con 6 item endgame. There might be exceptions; but most builds take advantage of them.
Second. Two points on constitution = 20 HP endgame. This means that every toughness feat could be considered equal to sacrificing a feat for 2.2 points of Con. Now, bumping up Con with ability points is a direct diminishing return (since it costs 2 then 3 build points per ability point raise). Let's take a common example of a build starting with 14 Con.
1 toughness = 2.2 Con = ~4 additional build points
2 toughness = 4.4 Con = ~11 additional build points
3 toughness = 6.6 Con = <technically infinite build points since this is impossible to do at creation and therefore akin to dividing by zero; however we can extrapolate as we like>
This is the fun part of the math. This is entirely right; completely the opposite of conventional wisdom; but just as true as conventional wisdom. The math is coherent within the framework. Therefore, if I wanted to say that taking toughnesses multiple times is actually more beneficial than taking it just once, especially on low hit die bards, rogues, etc to include the top issue - there would be math to support this.
Yes I was bored ; but at the same time, I felt it beneficial to verbalize why, in the absence of a direct impact DPS feat, it could often be a good idea to "pile on" toughnesses on a toon and not regret it in the slightest. Don't believe it is a horrible idea just because of misleading math; when we can come up with misleading math for just the opposite point of view.
PS If you are a true believer in the general rule of thumb, and firmly convinced that your math *is* right, you have my permission to dismiss me as a heretic and move on, or go read some general semantics.