Tom S,
Well, yes and no. The current production is in process of being built as I write. As far as they go, the crossovers aren't even built yet. The demo units do have the new topology but they’re going to be upgraded with the Mundorf parts as soon as they begin arriving. Essentially, it’s easier to build whole new crossovers than to take the ones in those units apart and replace the parts. Essentially, everything will be “new” that goes out the door or sets in our demo areas once Mundorf gets the parts to us.
Russell,
Your units are a bit of a hybrid. They are the closest pair out there to this final design, but there are a couple of capacitor/resistor value changes that yours lack. It’s a long story but
Double Ugly loaned his original pair out to
Jim Merod, who then forwarded his pair back to us. We then upgraded that pair to the final topology and shipped them to D.U. After that pair went out to D.U.,
vonwaffen purchased his pair.
Don’t worry, just e-mail me about this and we’ll get you fixed up.
Oh, and as to how I’m feeling right now?

ooheadsoo & D.U.,
So...there are a few demented minds such as myself that are interested in such mundane little technical details?! OK, you asked for it.
To begin, a little background is in order. The biggest source of benefit and headache in our designs is the waveguide. Obviously, not every speaker has one of those things so their implementation is not as straight forward as a simple woofer & tweeter in a box. Such designs may be boring and low tech, but these days "boring" would be a nice change of pace.

Anyway, the waveguide produces an acoustic gain effect (increased SPL) at the bottom end of the tweeter's operating range. The combined tweeter/waveguide system exhibits 10dB of more output in a frequency range centered on either side of about 1.5KHz. Above 5KHz, the output is the same as if the tweeter were mounted in a flat baffle.
Well, if this big 10dB "hump" in the frequency response were not corrected for by an equalizer circuit, it would sound horrible! So...we have a passive network (engineering term - band reject filter) that is a subsection of the complete crossover. The problem arises from the fact that the reactive impedance's of this network will interact with the traditional "High-Pass" part of the crossover in a very complex way. 10dB is a lot of gain to try and attenuate over a limited band while avoiding interaction with the other components in the network. Invariably, the selection of components is a difficult process because many more than one set of values will achieve the desired "flat frequency response." The problem is that any given set of values will exhibit their own unique "terminal impedance" as seen by the driving amplifier.
The set of all possible combinations that give a flat response is just a shade shy of infinite. There are 3 possible ways to pick the best combination that yields both a flat frequency response and the smoothest electrical impedance. The first is to model it mathematically on paper. Even an EE with a P.H.D. wouldn't want to tackle that by long hand - and I sure ain't no P.H.D.
The second way is to do the modelling via computer. Problem is, you have to have the software that will allow you to import the actual frequency response from the driver, as well as it's electrical specifications. This is the 21st century way of doing it, even if you don't have to have any brains. Well, when we started building these things, that software didn't exist, or if it did, it was way too expensive. Besides, the hardware wasn't up to snuff back then anyway. Now days the stuff that's out there will do everything but get your coffee for you. So in lieu of that option, we followed the third approach.
The third option is...run some basic numbers to get component values in the ballpark. Then...fly by the seat of your pants. Use your intuition and take measurements. Adjust values and repeat - ad infinitum. Hopefully you're sanity will remain intact when you're done and the impedance won't be
too bad. If it is, back to the dungeon until it's acceptable. In a lot of cases "acceptable" is about the best you can hope for.
Well, along the way we picked up the software...and since then I've been too darn busy to learn it!!! So, I just went with what I know and toiled over my test bench until I got the darn thing right.
Specifically, 1 coil went away, 1 coil went up in value, 1 coil went way down in value, 1 capacitor went way down in value, 1 went up in value, one resistor changed in value and 1 resistor was added. Then, another tweeter compensation network of 1 coil, 1 capacitor and 1 resistor went away completely! - No longer needed. As far as what each of these parts does, we don't have the space, I don't have the time and trust me...you DON'T want to know! Hell, for the time it's cost me, I wish I didn't. Oh yeah, no offense intended but...HA! it sure wasn't as simple as a stupid little Zobel network!
Anyway...that's the scoop. Now aren't you sorry you asked?

PS. The stuffing issue is my little secret. You can't expect a girl to give them all up can you? (No!...I'm not really a girl.)
TomSThanks! Yeah, things are getting a little hectic around here lately. I guess the NuForce S-9 project has had several benefits. The time it's taken to do it has given us the time needed to refine everything else. Along the way, I guess folks have figured out that we're not as dimb as we look. aa
-Bob