Nice link. Don't know how I missed that after so many threads on the matter. Here's another post:
Originally Posted by IpodHappy
As a physicist, I've spent a bit of time thinking about some of the factors which might impact audio performance of some of the IEMs. Here are my thoughts relating materials science with what might be happening. Please note that these hypotheses are not validated.
There are several changes which come from using your electronics:
1) With newer ceramic chip capacitors, there is some shifting of value under potential loading or aging due to instabilities in the dielectric materials. These capacitors are also not linear with voltage. I don't know what values are being used for current 2 or 3 driver cross overs or their body sizes, so I don't know if they are using the most recent dielectric materials. Older electrolytic capacitors with polymer dielectrics also changed during forming, a process through which the dipoles would align with the electric field potential.
2) With sufficient time, even at room temperature, Cu, eutectic and Pb-free solders (SbAgCu or SAC) will anneal, changing their electrical characteristics slightly. Plated Cu is especially susceptible (Cu traces on circuit boards). Vacancies are driven to the grain boundary surfaces, resulting is slight shifts of resistance, but no more than would occur due to standard strains (Cu foils are often used as strain gauges since the Cu resistance will shift under strain).
3) If enough current is passed through the electrical conductors, the electron wind bombarding the atoms will result in movement of the conductor atoms, again largely along grain boundaries, in an effect called electromigration. Hopefully, none of your electronics are designed so poorly as to experience this effect.
4) Epoxies rely on chemical bonding, called cross linking, during cure which results in densification and increased modulus (stiffening). During the manufacture of products using epoxies, there is often a push to reduce production time by reducing the time allocated for curing. This means there may be additional curing occurring in the field. How might this affect headphones? In a dynamic driver, the voice coil is often glued to the diaphragm with epoxy. As the epoxy continues to cure, the bond between the coil and diaphragm will strengthen, resulting in more energy transferred to the driver and less absorbed in viscoelastic deformation and internal friction. This would likely result in better high frequency energy transfer between the coil and driver, improving high frequency efficiency.
5) The molecular chains in rolled or cast plastic films (driver diaphragms) tend to align to the rolling direction or surface on which they are cast. This gives them a bit different modulus in plane compared to out of plane, and definitely impacts their coefficient of thermal expansion (CTE) in each of these orientations. Vibrations during speaker use create micro fatigue in these molecular chains, resulting in slight re-arrangement of the molecules. Usually, this results in a hardening of the material, caused work hardening. I would expect this to be a second order effect compared to the continued cross linking which might occur over extended time periods as mentioned in (5), but again would improve high frequency efficiency.
6) In designs with metal speaker diaphragms (some BA drivers), the metal will also experience work hardening with cyclical stresses. This would result in a slight hardening which would again tend to make the driver more efficient at reproducing high frequencies.
Just some thoughts.
I recommend searching prior threads on the matter since this topic gets out of control.