Thinner makes sense wrt the relationship between rigidity/brittleness and flexibilty if one is concerned w/ excursion. Does that necessarily translate to better fatigue characteristics though? Does that necessarily hold for lateral stress when tensioning? This would put a constant, disproportionate stress on varying areas of the driver. In fact, it seems too complicated and dependent on the particular variables and an elements modulus and other factors to make a sweeping generalization right? The T50RP uses massive traces and even bigger ones for the original version. Both are rock solid w/ nearly never any driver failures. They also use smaller drivers than Audeze as well. The fact remains that the issues w/ the failure of LCD3 drivers and signature changes over time are not what I'd consider normal, let alone exemplary performance when compared to Fostex, Hifiman or even the LCD2. If thinner traces make the LCD3 drivers more reliable, then something else is wrong. Perhaps the film thickness.
It would also depend on the grain direction of the material. Here we use rolled&annealed copper and the grain direction can be an important factor with regards to a dynamic bending application. But I'm not sure how much the drivers are "flexing" here or even what metal is being used. General rule of thumb with regards to dielectric/substrate/film thickness, the thinner the better. You might want to review some of the data on the Hifiman / LCD-2 failures. The LCD-2s were put on hold a few months back for a similar issue (not just the cracked wood). Hifiman headphones have had their issues as well (though to a lesser degree than what Audeze went through this past winter with both the LCD-2/3). What the causes are, I haven't the faintest clue and hypothesizing without any real data is very difficult unless one has some sixth sense IMO.
Edited by MacedonianHero - 5/23/12 at 1:29pm