Horn & mixed-order XOver. Please criticize!

[hasselbaink]

Hello folks,

I am crossing an 18Sound 12NMB420 at 800 Hz to a Coral M-100 driver on a 400Hz tractrix horn. The 12" midrange sits in a semi open baffle enclosure (semi-cardoid).



Right now they are crossed passively using a 2. order LR topology. As you can see the mid overlaps the Coral quite a lot while the breakup region is still rather pronounced (1.5 to 3 kHz).

1/6 octave smoothing - please ignore everything above 5 kHz as my soundcard exhibits weird low pass behavior above this frequency.


What I am proposing is a mixed order crossover: 3. order mid-woofer and 1. order compression driver. Why this topology? Well, I wish to retain the 2. order type central vertical lobe pattern. To me this yields a very constant and seamless sound nicely showcasing what the drivers can do. I listen from 14 feet away (4.5 meters) and the sound is very coherent over a wide sweet spot.

You are all welcome to suggest the pros and cons of mixed order XO in this particular application.

Cheers
Kris

[sqlkev]

i'm currently testing a midrange horn + woofer myself and getting good results w/ mixed order crossover, (actively though).

Looks like you need to tame down that peak w/ either a notch or a higher order xover. Do you have any simulation tools for this passive project?

[hasselbaink]

Alas, no I have not. LspCAD would have been nice.
I would do the notch but only as a last resort.

Btw. the two bass-mid curves are with and without stuffing the U shaped baffle. Guess which one is which.

[Michael Chua]

Are you able to do a summed response with both drivers at 12dB?

Green plot with stuffing.

[Pano]

Nice looking rig! The mid reminds me of Gary Pimm's mid cardioid. Works well!

At first glance, I'd say you need a lot more underlap. But I can't really see well enough. Do you have your measurement files for each driver? If you also have them in a text format (like FRD) that would be great.

Impedance sweeps will be very handy if you are going to do this passive.

If you have that info, I'd be happy to help. It can take a lot of hours and hand banging to get it right. Another set of eyes can sure help!

[hasselbaink]

This is the summed response:

1/6 oct. smoothing


*Green plot* is the correct answer.

I have the imp. plot somewhere and will upload them shortly.

CHeers

[Pano]

Cool! Do you have the raw FR curves- with no crossover? That would be a big help.

[hasselbaink]

Wish I had the amplitude in a single 1/3 octave smoothed curve.

18Sound mid. in free air.


18Sound & Coral M-100 combined


All in all a nice plot.

[hasselbaink]

Originally Posted by panomaniac

Cool! Do you have the raw FR curves- with no crossover? That would be a big help.

The response I've measured is virtually a dead ringer for the one on the product sheet here 12NMB420 - High Output MB Neodymium Driver.

[Michael Chua]

Can you post a plot with only 3 curves.
1) final summed response
2) LP
3) HP

You should be able to see whether there are any issues with the crossover with both at 12dB.