Another designer using Cauer filters

[Dennis H]

Edit: well dang, Shawn must have gotten in touch with Al. I see he posted to another thread and introduced himself. I'll leave this one up anyway. Maybe we can get a "guru chat" going.

Hey Jon, do you know a guy named Al Klappenberger? The Joseph filters came up on another forum. I mentioned your stuff and someone else mentioned Al's stuff. Looks like he's using Cauer filters in his replacement XO designs for vintage Klipsch stuff. I read through his whitepaper and his design target is 120dB/octave initial with the stopband ripple down 25dB. Maybe that makes sense for what he's doing as his primary goal is to overcome the time alignment problems of the old horns.

The crossover I'm using is designed by Al Klappenberger. He is also an EE...specializing in filter design... primarily in the microwave area. I think his And Jon's filter designs are similar to each other. Not to many people doing filter design like this in the audio world.

Al details the filter designs on his site at:

http://www.alkeng.com/

In the Klipsch Upgrade section then in the Extreme Slope and in the DIY designs area. The ES600T is the network I'm using in my speakers.

Shawn

[JonMarsh]

Well, that's pretty cool, nice to see someone else doing something like this.

The thing about a cauer-elliptic is that it is just a topology; what you get from it is totally dependent on how it's tuned, and it can emulate more conventional filters over a pretty useful range.

I like to see more attentuaion in the stop band region before the first ripple back, and more attenuation in that slope, but maybe that's just my application goals. He's probably using fewer parts (which would explain the difference- the curves look more like the LP on the EM 7.2 MK2 with the low component count crossover).



The "Whitepaper" was interseting, but I was disappointed that he didn't have any plots of the modified speakers total frequency response... the proof of the pudding is in the eating, you know!

It's a hoot to see all those old drivers which were in common use back when I was a teenager...

[Dennis H]

I looked at one of his circuits and it's the same 5-component topology as yours. Standard 4th order layout with an extra cap. Pics are part of a 3-way.

Edit: Jon, I think that pic you posted has a linear frequency scale.

[Al Klappenberge]

Guys,

Well, there is a trade off between slope rate and minimum attenuation. I wanted to get down to 20 dB or more as fast as I could. If you move the "notches" away from the crossover frequency the ultmate attenuation goes up (arc top drops) but the slope is reduced. The idea is to minimize the "window" where the two adjacent drivers are both making a sound and can interfere with each other. Of course, adding parts will buy you both, but that was the best I could do with that parts count.

I do have lots of accoustic plots, if you watnto see some of them. I did them mainly to tests phasing between drivers to see that the crossover didn't have glitches.

Al K.

[Al Klappenberge]

Guys,

Here's the in-phase and out-of-phase test on the 600 Hz netwrok (I think it's the one you have the plot posted for above). The speaker was a Belle Klipsch woofer, an Altec 902 driver on an Altec 811b horn. The plot is only over the crossover reigon. The verticl scale is relitive. I figured out how to calibrate it in actual SPL units sometime later. But it's around 100 dB / 1 meter for 2.8V

Al K.