Crossover question for Al

[drewb]

I am thinking about using the ALK crossover on my Speakerlab clones. The drivers are identical to the Klipsch except for the tweeter. I have the wa4000 tweeter. This is a "constant directivity" speaker. After much research I found out that they require a 6db upward pole zeroed at the crossover point. This matches my observations that they are weak in the extreme highs.

I have been unable to find an example of this "equalization" in a passive crossover. I am aware that you must pay for an upward tilt by lowering the output. This matters little to me with the power of the khorns.

How would you alter your crossover to accomplish this?

Thanks,

Drew

[Al Klappenberger]

Drew,

Your assumptions are pretty-much correct. You can use my "A" network in the

Speakerlab clones with no problems, but the constant directivity tweeter will

not work in a Khorn.

I used a set of Beyma CP09 constant directivity tweeters in my "Heresy on

steroids" speakers using the equalization you suggest. The problem is that

the sensitivity starts out high at low frequency and drops off as you go up

at a 6 dB / octave rate. This means the sensitivity of the rest of the

speaker will determine the highest point that can be equalized flat. The

Khorn woofer is so sensitive that it will nearly equal the sensitivity of the

tweeter at the lowest frequency of it response, so you can't equalize it. The

big Heresys I built had an efficiency of about 90 dB / 1 meter with 1W input.

This was exactly right to flaten the response to beyond 20 KHz. Working with

the Khorn woofer I would guess it could only be made flat to about 10 KHz. I

would have to dig out the curves on the CP09 to be sure though. It certainly

wouldn't go all the way to 20 KHz and very likely not even to 15 KHz.

I would suggest that you simply get another set of tweeters. I used Beyma

CP25s in my Belles. These will take you to 20KHz with no equalization at all

and have much better dispersion then the T-35.

Al K.

[djk]

I would build a modified typeA Klipschorn network with the better quality parts that Al uses.

Woofer fed thru 2.5mH choke

Squawker fed thru 12µF+1µF cap to terminal #3 on a T2A autoformer, output to squawker on terminal #2.

Tweeter fed thru 1µF cap off terminal #5.

Common ground for everything including #0

[dhsettim]

DJK suggests using different T2A taps than those shown on the "classic" typeA schematic--I guess giving different attenuation, so would he be willing to comment on his reasons. Thanks. Tim

[djk]

The attenuation of the mid driver is about the same as a type A or AA, the level for the tweeter is boosted 6dB and the frequency response is contoured for the WA4000 tweeter which has a constant directivity horn.

[soundog]

Al-

Please tell me more about the Beyma CP25s Tweeter - is that a direct substitution for the T35? Where do you buy it? Cost? Any change in ALK crossover settings?

[Al Klappenberger]

Soundog,

The Beyma CP25 is made in Spain. The cost is around $120 or so each direct from Beyma USA. You can get all the specks from the Beyma web site (http://www.beyma.com).

As you can see by the attached picture it is a lot biger than the T-35 / K77 but it is the same Zo and roughly the same sensitivity. It gas a wider response and better dispersion though.

You can so see them in my Belles along with Altec 811b horns and JBL drivers by clicking "My stereo stuff" below.

Al K.

[drewb]

DJK uses the autoformer to boost the tweeter 6db. This is an interesting approach. However my research shows that constant directivity horns have a 6db slope downward. If this is true boosting the tweeter 6db will not solve the problem. As the frequency goes up the amount of boost required would increase. Is this the case? Any input here would be appreciated.

Al, thanks for the schematic. You are the king of the extreme xover games. 70-120db slopes are definitely out there. I searched but could not find any conclusions on extreme slopes. Do they solve the Khorn crossover point timing issues?

You were right that the crossover schematic was to busy to easily isolate the tweeter equalization. Could I trouble you for a simplified sketch (like a text book 12db rlc example)?

The Beyma's look like the best I have seen so far for the khorns. Unfortunately I don't have the real estate or wallet for them.

Thanks to all for some very helpful responses.

[Al Klappenberger]

Drew,

Boosting the level to the tweeter with a reversed autoformer is a good idea

but has a big down side. You can boost the level a lot more than 6 dB, but

the down side is that the impedance drops like a rock looking back to the

network as you increase the boost. The equalizer tends to raise the impedance

back up though. This makes designing the tweeter / squawker very difficult.

That's why I haven't done that. Making impedance matched stuff is just my

thing. It may not be important to other people. It didn't seem to be

important to PWK!

The equalizer is simply a .27 uF cap and .24 mHy inductor in series with the

tweeter and a 100 Ohm resistor connected accost the L-C.

-*-L--C-*-

'--R---'

Al K.

[djk]

". However my research shows that constant directivity horns have a 6db slope downward. If this is true boosting the tweeter 6db will not solve the problem. As the frequency goes up the amount of boost required would increase. Is this the case? Any input here would be appreciated."

Do you have any idea what Xl=1/2PiFC means?

A 1µF cap causes a 6dB/oct slope below 20Khz, IOW, the drive to the tweeter is increasing at a 6dB/oct rate up to that frequency.

Draw it out on a sheet of paper if you can't visualize it. If the response of the tweeter is falling at 6dB/oct and the drive signal is rising at 6dB/oct, the combined response is flat.

If you're lazy, you can always buy a burned out pair of tweeters from eBay and swap the horns onto yours. That way you can use a stock network.