I also have DDS Eng 90-1 constant directivity waveguides that will cover from about 1100 up with 1" compression drivers.

Conceptually, my goals are to match and blend as many aspects here as possible while keeping crossover point above the vocal range, also paying particular attention to dispersion at crossover.

Right now I'm thinking that what I want to occur is that as the Azura's begin to beam, I want the waveguide to start taking over.

Initially, I had been thinking that I would do a first order crossover at a much higher point - say 4k or so, but have been told that was very difficult to do well, due to the 3db/octave rolloff characteristics of CD waveguides.

I don't know that it matters, and I don't see what I could do about it anyhow, but from center to center of the horns will be about 29" apart - one reason I think I have to be sure I do it right.

So where and how do you think I should start planning my crossover?

Regards

Ken L

• Spice
  
• Crossover Document
  
• Crossover Electronics 101 Lab Handhout
  
• Top octave compensation for constant directivity equalization
  
• Two-way system
  
• Horn Phase
  
• Baffle spacing, phase angles and time alignment, revisited

Chris

./atten.pl -8
driver Z: [8.00] ohms
Series resistor: [4.82] ohms
Parallel resistor: [5.29] ohms

#!/usr/bin/perl -w

if ( ! @ARGV ) {
$drop = -6; # dB
}
else {
$drop = shift(@ARGV);
}

$driver_Z = 8; # ohms

# split the load between the driver and the resistor
$Rp = (10 ** ($drop / 20 )) * ($driver_Z/(1 - (10 ** ($drop / 20 ))));

# calc the the combined impeadance of driver and parallel
# resistor, and subtract from original Z.

$Rs = $driver_Z - ( 1 / ( ( 1/$Rp ) + (1/$driver_Z )));

printf(" driver Z: [%3.2f] ohms\n Series resistor: [%3.2f] ohms\n Parallel resistor: [%3.2f] ohms\n",$driver_Z,$Rs,$Rp);

One can easily calculate the attenuation of a voltage divider formed with resistors. Where things become a little bit tricky is when the reactive impedance of the driver and of the crossover components is figured in. You can actually get peaking from reactive forces that is greater than attenuation provided by the resistors in the L-pad. Such peaking can form a narrow peak, or it can form a very wide band, having influence wider than an octave. This is dependent on the ratios of resistance to reactance, and so the easiest and best way to analyze the whole picture is with a circuit modeling tool like Spice.

I've read thru the crossover document and some of the other stuff.

I regret to say that I no longer seem to have the ability to process that sort of info as well as I used to.

Reading thru again may help clarify it somewhat.

regards

Ken L

The on-topic question I had is about the crossovers. Doesn't including
a significant series resistor alter the crossover frequency of the cap(s)
and coil? The reason I ask is that you've suggested in the past that the
Eminence 1.6K crossovers are OK to use along with the HF comp
components verses making up the whole x-over by hand.

Thx, Chris

About the crossover, yes, resistance changes the load and so it alters the transfer function. The crossover point and damping (Q) are affected. in fact, that's precisely what we want, in the case of the π crossover, we want the components R1/R2/C1 to alter the response of the base filter. We needed to create top-octave compensation that is the conjugate of the driver's power response. To do that requires 6dB/octave augmentaton after an initial shelf of flat response. We also need about 10dB padding. And it just so happens that these things shift the base filters enough that they sum nicely given the baffle spacing of the drivers. All these things are needed of the crossover, otherwise, it would not work for our purpose.

I found the culprit setup.exe under Docs&Settings/userID/temp.
Once I deleted that, the installer actually ran the right setup.exe.
I love Windows.... NOT!

How does one go about creating a Spice model of their drivers?
It seems to be something about making an RC tank circuit
that resonates at Fs, but aren't there a range of values that return
the same center freq? And how does Zmax fit into the picture?

Thx, Chris

About making a virtual circuit that models mechanical resonance, the idea is to find parallel L,C and R that acts as the mechanical resonator does. R will be set by Zmax, so that one is easy. Then, the values of C and L will be the same at resonance, and since Q is the ratio of reactive impedance to resistive impedance, you will find a the value of inductor and capacitor with reactances equal to Zmax / Qms at Fr.