A question of phase relationship

[Cosmos]

I recently set up the following with a Behringer DCX-2496 active crossover:

JBL 2226H in JBL 4507 cabinets.
JBL 2445J drivers on 350 Hz. Edgarhorns.

I am crossing at 560 Hz with a 6 db butterworth on the woofers and 24 db Butterworth on the horns. (I have already corrected to standard polarity on the woofers)

Here's my question:

A 6db slope Butterworth is 90 deg out of phase and a 24 db is 360 degrees out.

Since I can set phase relationships with the crossover, where should it be?

For whatever reason, in listening so far, it seems to sound best with with the horn set 225 deg out.. (inverted plus 45 deg)



Also, while I don't need CD compensation with this horn, the natural HF roll off of the 2445 seems to lend itself to compensation. Is there a standard fixed Lpad compenation formula to help? At the moment, I am using the parametric EQ function to add about 12db of boost centered at 14KHz...

I think I can use this:

Where R1= 9 ohms and R2=4 and bypassed with a cap.. maybe around 1 uF?


I used 4Pi crossovers when first setup. The electronic crossover is significantly better...

While I am not wild about the dispersion (or lack thereof) of the Edgarhorns, this setup shows serious promise..:yes:

 

[Zilch]

4507 cab wiring corrects for polarity; you may have re-inverted it. Check the polarity with a 1.5V battery to make sure the cone moves out with (+) to red.

For comp, you mean bypass R1 with a cap, I presume. Yes, but you've made an L-pad, attenuating ~10 dB, in addition to the HF comp. See Dickason for calculating parallel RC contour filters, or measure.

DCX has parametrics; use them for the comp.... :yes:

 

[Gibsonian]

The DCX2496 has a feature that will adjust for phase correction. You need a mic, but it's a great feature. It's what you need to be doing with your setup.

 

[westend]

Watching this thread with interest, I'm planning on buying the Behringer crossover for biamping Altecs and would like to know what configuration you end up with. The actives do a good job controlling the phase relationships, IMO.

 

[djnagle]

 

[Wayne Parham]

I recently set up the following with a Behringer DCX-2496 active crossover:

JBL 2226H in JBL 4507 cabinets.
JBL 2445J drivers on 350 Hz. Edgarhorns.

...

I used 4Pi crossovers when first setup. The electronic crossover is significantly better...

While I am not wild about the dispersion (or lack thereof) of the Edgarhorns, this setup shows serious promise..:yes:

The 4Pi crossover wouldn't work for this application because it is designed for a DI-matched speaker, matching a direct radiating 15" midwoofer to a 90°x40° horn with constant directivity. Their directivities match in the 1kHz to 2kHz range, so that's where crossover is done. And the CD horn needs top-octave compensation to counter the falling response due to mass rolloff.

In your case, the 350Hz tractrix will provide acoustic EQ because of collapsing directivity. You don't want the top-octave compensation in the crossover. Also, you probably want crossover to be about an octave lower than the 4Pi, probably somewhere in the 500Hz range, +/- 100Hz. You're using the same woofer as me, so it could be run higher, but a 350Hz tractrix horn probably will want a little bit lower crossover than that.

I've seen some guys with smaller tractrix horns use my crossover, or a modified form. A 650Hz tractrix can be crossed-over to a 15" woofer around 1kHz, to match the directivity at crossover. But most tractrix guys aren't worried about directivity, and run first-order crossovers and physically align the voice coils. Simply pad the tweeter to match the woofer, probably something like 10dB or so.

 

[djnagle]

Wayne, is there a way to really know if the voice coils are aligned properly? Or is it just a look as see type thing?

Also, can you explain this. "will provide acoustic EQ because of collapsing directivity"

 

[Cosmos]

4507 cab wiring corrects for polarity; you may have re-inverted it. Check the polarity with a 1.5V battery to make sure the cone moves out with (+) to red.

I have replaced the light solid core internal wiring with 12 ga. stranded. (You actually could hear a volume difference when I did just one side). When I did so, I checked cone movement with a battery and marked the external connections with + for forward cone movement.

For comp, you mean bypass R1 with a cap, I presume. Yes, but you've made an L-pad, attenuating ~10 dB, in addition to the HF comp.

Yes, I meant to bypass R1 with a cap to allow full power to pass above the HPF created by the cap while shelving or attenuating everything below that HPF with the L pad. I guess what I was asking was "Do the starting values look reasonable for a 2445J, figuring that the nominal impedance is something in the neighborhood of 12 ohms and I want about a 12 db attenuation below the HPF?"

I do not know how to calculate those values.. Really, I should install the values and measure with an RTA and adjust from there.

See Dickason for calculating parallel RC contour filters, or measure.

Perhaps I am just clueless, but I do not know who Dickason is nor where to find his or her calculators. Further, it would also be helpful if I knew what a parallel RC contour filter was. I can only assume it is a time or phase correction circuit.

To answer these questions, I have started doing some research to try to figure it out on my own... Engage the Zilch-o-warp device.. Perhaps, I can grasp some new knowledge along the way..

DCX has parametrics; use them for the comp.... :yes:

Actually, I already am.. What a nice sounding system already.. great promise.. even if the sweet spot is only 9.78" wide (enough space for my head if I don't tilt it)



BTW, I view this setup as a learning adventure. I do believe that I will end up with some form of a CD waveguide. However, there is alot to be said for this tractrix stuff. The image is incredible.. and what I have learned so far on active crossovers is just plain stunning to me.. the flexibility is addictive.

This is but one step in what I outlined in my "What is Relative Efficiency" thread...


Thanks!
Dan

 

[Cosmos]

The 4Pi crossover wouldn't work for this application because it is designed for a DI-matched speaker, matching a direct radiating 15" midwoofer to a 90°x40° horn with constant directivity. Their directivities match in the 1kHz to 2kHz range, so that's where crossover is done. And the CD horn needs top-octave compensation to counter the falling response due to mass rolloff.

In your case, the 350Hz tractrix will provide acoustic EQ because of collapsing directivity. You don't want the top-octave compensation in the crossover. Also, you probably want crossover to be about an octave lower than the 4Pi, probably somewhere in the 500Hz range, +/- 100Hz. You're using the same woofer as me, so it could be run higher, but a 350Hz tractrix horn probably will want a little bit lower crossover than that.

I've seen some guys with smaller tractrix horns use my crossover, or a modified form. A 650Hz tractrix can be crossed-over to a 15" woofer around 1kHz, to match the directivity at crossover. But most tractrix guys aren't worried about directivity, and run first-order crossovers and physically align the voice coils. Simply pad the tweeter to match the woofer, probably something like 10dB or so.

Wayne,

That explains a bunch of what I heard.. I did not mean to infer there was anything wrong with your crossovers.. It just wasn't the right circuit for this application. If there's any consolation, I also threw the Econowave crossover in and it didn't sound good either. Just plain the wrong design. I will be coming back to a 4pi design, both active and passive as well as comparing that to the Econowave crossover, waveguide and drivers... with the proper equipment. In my case, I thought the resultant boost above 8 or 10KHz would help counter the HF rolloff of the big 2445...

My goal is to sample these relatively modest expense systems that are aimed directly at high end sound... and in the end, pick the system(s) that best suits my tastes and learn along the way.. :thmbsp:

Your input and guidance is most appreciated! Perhaps others can also gain some knowledge along the way as I do..

 

[Cosmos]

The DCX2496 has a feature that will adjust for phase correction. You need a mic, but it's a great feature. It's what you need to be doing with your setup.

Thanks for your input. However, please enlighten me.. I have been playing with the phase correction features, but I don't see a way to measure the results with the DCX2496. Am I missing something???

 

[Zilch]

Perhaps I am just clueless, but I do not know who Dickason is nor where to find his or her calculators. Further, it would also be helpful if I knew what a parallel RC contour filter was. I can only assume it is a time or phase correction circuit.

Loudspeaker Design Cookbook, in the crossover chapter.

A response shaping (contour) filter, actually.... :yes:

 

[Cosmos]

Loudspeaker Design Cookbook, in the crossover chapter.

A response shaping (contour) filter, actually.... :yes:

Now I really feel clueless. I've had that book for several years...:withstpd:

 

[westend]

Contour networks from Google books . I had to look it up.

 

[Gibsonian]

Thanks for your input. However, please enlighten me.. I have been playing with the phase correction features, but I don't see a way to measure the results with the DCX2496. Am I missing something???

I am not sure if this feature captures all you want, but it is section 4.2.2 in your manual (Delay corretion/auto align). It is an auto correction feaure that corrects for phase and time alignment, how effective I cannot say for certain, as I do not have my DCX anymore, but did use this feature when I had it. It will detect out of phase and correct as well as time adjustments between drivers. You should run it if you haven't but you need a mic placed and best listening position to do it.

 

[Wayne Parham]

Wayne, is there a way to really know if the voice coils are aligned properly? Or is it just a look as see type thing?

You can do acoustic measurements to find the apparent position of the driver. It will shift with frequency. However, most people that run first-order crossovers simply physically align the voice coils, by eye, as you say.

Also, can you explain this. "will provide acoustic EQ because of collapsing directivity"

As the horn begins to beam, the energy is focused into a tighter and tighter pattern. Since beamwidth is narrower as frequency goes up, the on-axis response is equalized acoustically to favor high frequencies. The off-axis response falls proportionally, sort of giving its energy to the on-axis response.

Constant directivity horns do not have collapsing directivity, so they provide no acoustic EQ. That's why the crossover has electronic compensation for the top octave. It conjugates the mass rolloff of the driver.

As an aside, waveguides do not employ a diffraction slot, so they begin to beam in the top octave. Some compression drivers also have a peak from diaphragm resonance up high too. These sometimes combine to make a sharp peak in the ~16kHz range. You've seen that in many of the response curves published on the Econowave thread. Tractrix horns get it too, as they beam up high.

Horns that have a diffraction slot tend to smooth out this notch up high, because they maintain higher dispersion using their diffracton slot. I don't find that a good reason to use them, personally, as the disadvantages in hifi applications far outweigh that one trait, in my opinion. But it is something you can see. The point is, I think it's real important to use a smooth compression driver on a waveguide or tractrix horn. Some people use foam in the horn, others put a very small value cap (~0.47uF) in shunt, across the voice coil. There is a place on my crossover board (C1alt) for this shunt cap, if desired. In any case, I think it best to use a compression driver with a well damped diaphragm, even more so on waveguides and other horns that have no diffraction slot in the throat.

That explains a bunch of what I heard.. I did not mean to infer there was anything wrong with your crossovers.. It just wasn't the right circuit for this application. If there's any consolation, I also threw the Econowave crossover in and it didn't sound good either. Just plain the wrong design. I will be coming back to a 4pi design, both active and passive as well as comparing that to the Econowave crossover, waveguide and drivers... with the proper equipment. In my case, I thought the resultant boost above 8 or 10KHz would help counter the HF rolloff of the big 2445...

I knew that but thanks for saying anyway. Any crossover that provides top-octave compensation is going to sound shrill on a tractrix horn because mass rolloff is compensated twice - once acoustically by the collapsing DI of the tractrix horn, a second time electronically by the compensation network (R1/R2/C1).

Please understand that my crossovers are designed specifically for the speakers they're used in. People have modded them for other applications, with varying results. The Econowave project is based on my crossover, but that makes sense - the Econowave is virtually identical to the three Pi and four Pi models. The only thing is, the values of the core splitter filters should probably be adjusted to be specific for the drivers chosen, because those set the phase, which in turn sets the position of the forward lobe and vertical nulls. Those are often different for each driver, because they all have different electro-mechanico-acoustic properties. But some are similar enough to each other they can be swapped in without modification.

 

[Big Tuna]

Hey Dan, sounds like you got your hands full. I wouldn't begin to even understand what your getting at here. I myself have a few projects I will be working on sometime in the future. Where can one find this "loudspeaker design cookbook"? I have a few projects I will be working on and I think that will be very helpful with them.

 

[Cosmos]

Hey Dan, sounds like you got your hands full. I wouldn't begin to even understand what your getting at here. I myself have a few projects I will be working on sometime in the future. Where can one find this "loudspeaker design cookbook"? I have a few projects I will be working on and I think that will be very helpful with them.

Amazon has them here..

It is a good book.. one of these days I'll get past the foreword..

 

[Zilch]

Parts Express and Madisound sell "The Book."

Compare pricing.... :yes:

 

[Cosmos]

I am really diggin' this digital electronic crossover. This thing is so flexible... It has me looking at upgrading to the dbx Driverack PA+ or 260, which includes on the fly RTA and an auto EQ function, among many other things... Much more flexible than the older crossover only devices.

 

[Cosmos]

Today, I tried 500 Hz Edgarhorns in lieu of the 350 Hz horns. The 500 Hz horns sound like they have better HF extension, but they also are much thinner down low. Perhaps the perceived extension is because of the weaker lower mids... I also raised the crossover point to 800 and then 1000 Hz.. haven't found the sweet spot for the 500 Hz Edgarhorns yet...

So far, I'd pick the 350 Hz horns and either suffer the extension or use super tweeters...