ALIII's and Crossovers

Since I have a jones for a pair of ALIII's you are likely to see me ask a lot of questions. All speakers are compromises and the ALIII's have a great combination of performance and WAF. Not perfect, but probably as close as I can get with a reasonable amount of money. I do promise to search the forum first so hopefully won't start a topic that's already been discussed to death.

As soon as I get my hands on some, I'll be able to transition from "picking your brains" mode to "contributing with new information" mode.

That said, looking at pictures of the crossovers posted in one of the threads (viewtopic.php?f=15&t=7773&p=86233&hilit=alIII#p86233), I see lots of potential for some improvements. I say potential, because Mr. Carver always seems to do the impossible, and I wouldn't be surprised to for him to take advantage of a property of what would normally be considered a design flaw.

Here are some of this things I see (easiest first, hardest last):
* Replace electrolytic caps with metalized polypropolene or other
* Replace wire wound resistors with non-inductive resistors
* Add high quality bypass caps to existing caps
* Reposition inductors and change orientation to prevent magnetic interactions.
* Replace inductors with high quality Solen and Erse inductors. Going this route would also mean matching DCR's and likely result in adding non-inductive resistors in series with the replacements.
* Replace all caps with biased capacitor pairs

Anything beyond minor surgery would likely be best done separate from the stock crossover. I usually like to retain the stock crossover completely intact, and build a separate network. This also allows me to make side by side measurements and listening comparisons. Considering what these ribbons are cabable of I think a completely redone network (biased caps, new inductors properly arranged, non-inductive resistors) might be in order.

Has anybody been down this road and what were your results like?

It would probably be better to eliminate the crossover altogether and do it all actively with an active external crossover and a parametric equalizer. A specially built box (see Clearwater) to do both would probably cost less than the changes you recommend. Getting metalized poly caps with enough capacitance to replace the NP electrolytics would be a special make (and they'd be huge). Those Erse (or Solen) inductors don't come cheap either, and would have to be custom wound for the inductance (and resistance) necessary.

You can get a Rane AC-22 for < $100. I understand with Marchand, you can specify a couple of EQ points to be built into a crossover they offer.

BillD,
Yikes, you are correct about the caps! I somehow mentally lapsed about the sizes needed for a 150 Hz crossover. A biased 150 Hz network would be quite large.

It sounds like active is the way to go. When I get a pair to look at, I'll map the voltage drive of the built in crossover. If it's just a simple crossover at 150 Hz, with xDb per octave, then using an active crossover will work just fine. Without a schematic to look at It's difficult to see if some shaping was done in the passive crossover.

I haven't been very satisfied with the noise levels of most active XO's I've used, but that is mostly due to the use of high efficiency speakers.

I characterized the crossover for the Silvers (48" ribbon). Bob's crossovers are funny. In this silver crossover, the corner frequency of the low end is 145 Hz and the high is 167 Hz. The lows and highs are 180° out-of-phase, so the woofers are hooked up backwards.

Silver Crossover.JPG (80.08 KiB) Viewed 450 times