A Zobel on your tweeter ?

[Al Klappenberger]

Having seen a Zobel network applied to a piezo tweeter in another thread I got curios if a Zobel might be of some advantage to an ordinary tweeter, or the K55 squawker for that matter. A well designed filter intended to be terminated in a pure resistive load will get very screwed up if it’s loaded with a complex impedance that includes inductance. R-C Zobels are often used for woofer filters but I always assumed tweeter filters didn’t need it because their impedance was good enough to load a filter by itself. I decided to do a little research to see if that was true. I plotted the complex impedance of Bob Crites CT125 tweeter. It’s roughly 7.5 Ohms in series with about 38 uHy. (MICRO Henrys) at 10 Khz. That’s not band. With a Zobel of 10 Ohms in series with 0.68 uFd connected across the tweeter terminals the inductance cancels out nicely. I made plots of the response of the high channel output of my ES5800 network loaded by the CT125 tweeter with and without the Zobel connected. There’s a rise of about 4 dB at 20 Khz withOUT the Zobel. With it operating the response is flat like it is supposed to be. I also plotted the acoustic output of the tweeter with the Zobel in and out of the circuit. The rise shows up in the response but only above about 12 KHz in any significant amount. I really doubt that difference will be worth the effort to implement it.

Next I did the same investigation using the K55 squawker driver.

Al K.

BTW: Don't assume that the acoustic plot of Bob's tweeter is really like the plot I show. It's NOT! The plot was taken with the mike laying on the bench and the tweeter several inches away. Just look at the difference between the plots.

[Al Klappenberger]

I did the same plot of impedance and ES5800 network output using a Zobel of 13 Ohms in series with 1.2 uF across a K55V driver on a K500 horn (Belle Klipsch squawker horn). Again the inductance tuned out nicely. In this case the network was looking into the K55 through the combination of a 3619 transformer (taps X-4) and the 10 Ohm swamping resistor it normally sees with my networks. This combination swamped out the need for the Zobel entirely. I didn’t bother to plot the acoustic output. Without the swamping resistor, the stock Klipsch networks would probably be susceptible to the inductance and might benefit from the Zobel. I didn’t bother to test that situation though.

Al K.

[freddyi]

Hi AL - since I've got K55V laying around not hooked up - here's one on an Altec 511 - maybe a Zoebel would help a bit with cap and inductor rolloff to keep K55V out of the tweeter region and its own double dip region -? - might be a waste of parts and money

http://img35.imageshack.us/img35/7086/k55v511z.gif

[Al Klappenberger]

Freddy,

NAH! It won't do a thing! I have been thinking on this since the post. This is just an academic exercise that virtually confirms why Zobels are not normally used except on woofer filters. Adding one to a K55 driven by a Klipsch "A series" network is silly. An example of why is to think about the 4500 Hz mod Bob Crites adds to the AA. It's an inductor in series with the driver. It simply adds to and overwhelms the tiny 0.2 mHy voice coil inductance of the K55! If you want to attenuate the 9 kHz glitch you need to filter it off. There's how to do it. A Zobel will just cancel the series inductance making the glitch worse!

Al K.

[Al Klappenberger]

Freddy,

If you really want to get rid of the 9 KHz glitch, here's a filter that will NAIL it! All you need is three 1.5 uF caps and a .2 mHy inductor connected right in the cable going from the network to the K55 driver. It shouldn't upset the midrange at all.

Al K

[freddyi]

wow ! - that is a good notch - -thanks!