
I've gone through 7 pages and jumped to Mython's summary post on introduction. I don't have time to sift through all this with a crying baby on my lap just to see what I've said wrong. But from what I've come across so far in the video of crossovers, and Bilavideo has said so far about crossovers in the first 7 pages, I can't see where my high level concept is wrong in what crossovers are for.
Think I'll just wait for your summarised thread.
I owe you guys an apology; lately, I've been running around half-annoyed, and I think I just ran out of patience in that single moment. I didn't mean to single you guys out. However, as I mentioned, crossover implementation with BAs and/or dynamic drivers is very complex, and every single component, whether it be the driver itself (which already has four different factory options for "tuning", including three damping options and a back-venting option, as well as different coil impedances), the material of the sound tubing (for conventional CIEMs, they use hearing aid tubing, which is available in different densities of silicone, and they will decay acoustic energy as a linear function of distance from the driver spout), the value of the acoustic (physical) damper, the diameter of the tubing, the exit bore diameter with respect to the sound tubing diameter (horn effects), the list goes on and on. Also, because of the distortion characteristics of BAs, traditional LC low-pass circuits are difficult to implement in crossovers, since conventionally-wound inductors will blow up the distortion to unacceptable levels. Thus, the design of the crossover network itself is not as straightforward as it is in loudspeakers, and traditionally only uses SMD resistors and capacitors, which is limiting if you want to build a phase-coherent crossover at the SPLs that you want. You can start to get the picture about how easy it is to misattribute what one is hearing with how something is "tuned", so that's why I ventured to say that talking about crossovers is a bit of a moot point, unless people really, truly understand it. I can't even say I understand it fully; I was lucky to have attended an extremely rigorous engineering-oriented school, yet I studied neuroscience there, and now I have addition medical knowledge as well; this education helped me more quickly understand the applications of electroacoustic engineering in this small, but specific field of multi-driver IEMs. At the same time, while IEM design doesn't require complete and absolute command of any single subject, it is equal parts physics, neuroscience, material science, and electrical engineering.
In the past I've been guilty of spewing out crazy theories and guesses, but lately, I've begun to keep those thoughts to myself, because I find that they just confuse others more than they help.

It's a lot more complicated in an IEM where tube length can offer acoustic resistace along with dedicated filters. Crossovers can be a mix of acoustic and electrical resistance. There are also much less common series networks.
For instance, most versions of twfk have a treble peak and mid dip. If you resist the treble peak the dip can become more severe. Running a dynamic driver in parallel with a normal mid bump and (not so) simply limiting it's highs could allow you to make the system nicely linear and still fast where needed. TWFKs also get used as purely double high or double mid frequency units so what's actually happening here is any body's guess unless we're told or look for ourselves. Could be a 1.5 way, 2 way, 2.5 way, 3 way or even crossoverless and why tomscy is not wasting time speculating. Have we discussed venting.
Multi driver anything is as much art as science. Makes you wonder how many iterations Dale played with before he got it to sing. Proof is in the result which seems really good.
Thanks for helping out. :-)
Edited by tomscy2000 - 2/11/13 at 10:14am


























