Need Crossover advice, AR 8B (Teledyne) bookshelf speakers

Hello All, I recently was gifted AR 8B bookshelf speakers. Of course the foam surrounds are dry rotted and I have the kit to refoam the woofers.
While I am doing that I got inspired to tweak the crossover thinking that i can improve the sound over the original. The woofer is running free-range with no LP filter. The tweeter uses a 5.0 uF capacitor only. Per what I posted below, it sounds like this was done as a cost-cutting effort and not necessarily to deliver the best sound. I'm thinking a 12 db per octave LP crossover at 2000 Hz is a good idea to prevent distortion. According to AR, the tweeter has an "effective" rolloff of 18 db at 2000 Hz using the single 5.0 uF cap. Should a more sophisticated HP filter be used incorporating an induction coil and a cap, and should the HP filter use a 12 db/octave slope as the tweeter was designed to have a natural rolloff giving an effective 18 db/octave slope? Or should the HP filter use a 3rd order filter? Please advise, the more I look into this the more I am getting confused! Thanks!

"The two-way AR8B is the smallest and least expensive speaker system in the current Acoustic Research line. Like all AR speakers, it is an acoustic-suspension model, and it has a newly designed 6-inch woofer that crosses over at 2,000 Hz to a 1-1/4-inch “liquid cooled” cone tweeter. (The tweeter presumably uses ferrofluid or a similar material to damp and cool its voice coil.) The crossover network has been simplified and its cost reduced by designing each driver to have a natural rolloff outside its operating frequency range. As a result, only a capacitor is needed to protect the tweeter against damage from powerful low-frequency signals. The effective crossover slopes are 12 dB per octave for the woofer and 18 dB per octave for the tweeter."
 
Should a more sophisticated HP filter be used incorporating an induction coil and a cap

There is no need to incorporate an inductor in the tweeter high pass filter as the combination of the natural roll-off built into the tweeter and the electrical roll-off caused by the series capacitor supplies a more than adequate acoustic roll-off slope.

Such minimalist crossover designs can work very well provided the drivers are engineered with natural roll-offs to suit, as Acoustic Research appears to have done in this case.

1745607581695.png


In my opinion, adding additional crossover components may negatively effect the integration of the drivers in the AR8B.

To justify a crossover modification, you would need to view the frequency response of the speaker in its original state in order to reveal any potential weakness that might be solved by adding additional crossover components.
 
Hi Galu, I appreciate your reply. You bring up a valid point that I have considered. My thinking is that in order to keep costs down a very simple HP filter was used. Was hoping maybe someone has had direct experience with restoring these and have experimented with the crossover.
 
  • Like
Reactions: Galu
should the HP filter use a 12 db/octave slope as the tweeter was designed to have a natural rolloff giving an effective 18 db/octave slope?

If the tweeter has a natural roll-off of 12 dB/octave then adding a 2nd order, 12 dB/octave filter could result in an acoustic roll-off of 24 dB/octave.

Adding a 2nd order, 12 dB/octave filter to the woofer could also result in an acoustic roll-off of 24 dB/octave.

That may be a direction in which to experiment, but I fear it could result in a reproduction that is less 'alive' than that of the original configuration.

Hopefully this, my second contribution, will give a bump to your thread and encourage a vintage AR loudspeaker enthusiast to chime in!
 
2khz crossover frequency would require a pretty heavy cone and some more inductivity than usual.

2 khz would be very low crossing for a tweete with 6db.

Natural roll off at 4khz for a bass driver is more realistic
 
2 cents >
I have read that "liquid cooled" / ferro-fluid type tweeters dry-out over time and may actually create a 'goo' in the voice coil gap.
This would not only effect frequency response, but also output efficiency 😕