3-Way Crossover Help!

[69stingray]

Hello Everyone. I am in the process of designing a crossover for my future 3-way speaker. I am using version 3.0 of the "Passive Crossover Designer" by Jeff Bagby, which is excellent, great use of Excel!

I am using a 3rd filter throughout and have chosen crossover frequencies of 400 Hz and 3,200 Hz. Basically, the tweeter has an Fs of 1,400 Hz, so I choose 3,200 Hz since it is 1.5 octaves above the tweeter’s Fs frequency. The 400 Hz frequency is 3 octaves below 3,200 Hz. I have a small L-Pad on the tweeter that I might even do away with, it is only 0.5 Ohms. I have also added a Zobel network to the midrange.

The attached PDF will show you my current crossover construction and my summed freq. response. I have two concerns/questions.

(1) I need to verify I labeled the components currently using the values from the Excel spreadsheet.
(2) The upper end of the band-pass increases in frequency. When I added the Zobel, this helped, but I would like to lower it more. Also, there is a dip pretty close to the crossover (approx. 3,200 Hz). The PHASE PLOT has something going on e at the same frequency. What can I do to correct this and smooth out the summed response curve?

Here are the drivers I am using:

Dayton 7” Aluminum

Audax 4"

Badgeless 1" Tweeter

[69stingray]

Putting all the images in one file was too large, had to break them up.

[69stingray]

more.....

[69stingray]

final one......

[paulhfx]

Originally posted by 69stingray
The upper end of the band-pass increases in frequency. When I added the Zobel, this helped, but I would like to lower it more. Also, there is a dip pretty close to the crossover (approx. 3,200 Hz). The PHASE PLOT has something going on e at the same frequency. What can I do to correct this and smooth out the summed response curve?

It appears that your mid and tweeter are out of phase at the crossover, which is creating that glitch in the spl graph. You also have a rising response in the 2k-3k range.

I'd leave the tweeter and woofer xover alone for the moment and rework the mid's low pass for a lower crossover point (say 2600-2800). This will reduce the rising response below the 3200Hz point, and will also impact phase (not sure exactly how - that's why we have these simulators ). Check your results with this revised xover, and switch phase on the tweeter if necessary/if it helps.

Paul

[percy]

not sure whether this will help or not but have you tried different xover orders ? For the 3-way I was modeling, LR2 looked the best. Third order gave me problems similar to what you seem to be having. You get one end of it right and the other falls apart!

Try overlapping the xover frequencies between drivers as is often the case in a 3-way.

[69stingray]

Thanks everyone. I tried changing X-Over freq. of the midrange, didn't seem to help at all until I raised it into the 4k range, then the summed response went bad. What I did try was changing the high-pass on the tweeter to a second order and that got rid of the dip due to the mid and tweeter being out of phase. I have attached my current summed plot with +/- 3 dB boundary. What does everyone think? 2k-3k is still a little hairy.

[kelticwizard]

Well, this is just the purest guesswork, and I haven't plugged these numbers into any program.

Rework the mid crossover to 2900Hz, the tweeter to 3500 Hz.

This is an attempt to lower the mid crossover by 1/8 of an octave, and to raise the tweeter's crossover by the same amount.

No promises, but it might help that hump at 3200 Hz.