I do not have intimate knowledge of the upgrade progression of the coaxes; they came after my time at Thiel. I do know how the path worked. Jim constantly isolated the weakest link of the products at large and worked on a solution to implement in the candidate model which surfaced via that process. So every newer product had advancements over its older forebear. And every new model had multiple driver advancements in its recipe.
The model progression (excluding the SCS series) of the coax is:
1995 CS7 - 1996 CS6 - 1998 CS2.3 - 1999 CS7.2 - 2003 CS2.4 - 2007 CS3.7
The SCS and its permutations also added knowledge and solutions in a parallel timeline. So, the CS6 is near the beginning and the 2.4 near the end with 7 intervening years. The advancements affect all areas: materials, geometry, motor performance, etc. which all serve to simplify distortion modes, which allow simpler crossover circuits to correct those misbehaviors. Jim consolidated driver and crossover behavior, working toward a goal of more ideal system behavior. I directly experienced such improvements through the first 20 years of innovation. They are significant in every model development.
Part of the elegance of the 2.4 is the mechanical crossover between the midrange and tweeter. That idea percolated since the 1990 SCS, our first unitized driver. The pipe dream of a non-electronic crossover gradually took form as we developed the chops to jump from imagination to reality. The 2.3 - 2.4 small coax still had to be controlled. That took 15 XO parts. The CS6 coax took 40.
Regarding the notch filter upgrade which Duramax mentions. Thiel's need to control drivers through 7 octaves vs most speakers' 3-4, necessitates some big caps. We used PP film caps where feasible. And we developed that yellow1uF styrene/tin bypass cap for the 1989 CS5 which spread to all models, to good effect. We also limited our largest cap value to 100uF and ganged them for values up to 400+uF. The CS7.2 has such a 400uF bank in an upper midrange notch filter. Experience shows it to affect performance in its operating range. I have replaced that bank of 4x100uF electrolytixs+1x1uFs/t with two versions now under test and trials. A is 8x50uF el+1x1uF cascading multi-section PP (details proprietary.) That is what Duramax is trying. Next version is the same new 1uF core bypass plus 6x40uF els + 6x10uF PP film caps. He and I will test those next. Note that these solutions would be too expensive for classic Thiel's cost/performance plateau. But they do serve in learning new directions. Today I am working with such an upgrade in the 300uF SCS4 woofer series feed cap. I'll post outcomes when I have something to share.