"Quasi-optimal" crossover for high-efficiency loudspeaker system

[fpitas]

In the end, I think those who still think that systems like these (large woofer + horn) inevitably force you to renounce coherence and imaging for the sake of sheer dynamics, should do themselves a favour and give one of these "quasi-optimal" crossovers a go!

Marco

Agreed. I'm using the Le Cléac'h crossover (3rd order on woofer and tweeter) very successfully with a woofer/horn system. I was pleasantly surprised at the improvement from the previous standard LR4 setup; better sounding transients and generally more natural sounding.

 

[jmbee]

4th order L-R Low Pass, -6dB @ Fx (+)
4th order L-R High Pass, -6dB @ Fx (-)
Offset = 0.31*c/Fx

Marco

Hello Marco,

I think the official 4/4 is :

4th ordrer Butt Low Pass, -3dB @ Fx* 0,93 (+)
4th order L-R High Pass, -6dB @ Fx* 0,93 (-)
Offset = 0,31*c/Fc

crd

 

[marco_gea]

Yes! That was a copy-and-paste mistake! Thank you for catching it and rectifying. Unfortunately I can no longer edit my post :-(

Marco

 

[fpitas]

The quasi-optimal crossovers presented here have the Low and High Pass slopes equal, or have the higher slope on the Low filter. Do any exist with (for example) a 2nd order Low Pass and 3rd order High Pass?

 

[jmbee]

Hello,

I wrote a page with all kind of " quasi-optimal " i know, in French but all images needed could help.
Just pay attention to Bessel kind, sometimes Fc is given for -3dB, ( original excel JMLC )
sometimes in N*45° ( standard fan, and other softs ).
Here you will find a Butt 2 / Bess 3

crd

filtre "quasi-optimal" : modèles et variantes - Le blog de jimbee

 

[fpitas]

Thank you, jmbee.

 

[danvprod]

Thank you very much for this thread. It's helping me to answer a lot of questions on my JBL horn build. I've had similar problems setting up the crossover for my 2226J in BR box and 2445J in a StereoLab horn.

I started with the basic LR4/LR4 crossover but quickly found that wasn't going to work because of the physical offsets of the drivers. I'm currently using a LR2/LR2 but I planning to try some of the quasi-optimal crossovers here.

I have a question about the 'Offset' figures you represent here. This corresponds to the corresponding delay that crossover creates, and how far back the CD should be from the woofer's acoustic center?

So if I am aiming for a 1kHz x-over point and I use:

4th ordrer Butt Low Pass, -3dB @ Fx* 0,93 (+)
4th order L-R High Pass, -6dB @ Fx* 0,93 (-)
Offset = 0,31*c/Fc

For example, the offset is: .106 (meters?) so the horn driver should be 10.6 cm behind the woofer's acoustic center?

And in this case are my x-over points 930 Hz and 1070 Hz?

Right now my physical offset is measurement at about 0.3ms, or 9.144 CM, which would seem to be pretty close.

I am also having a hard time making credible phase measurements in-room. Is the best tool to do this HolmImpulse with a gated measurement? Should I be measuring in the center of my room at about 1m from both drivers and gating before the first reflection?

Thanks again for the help.

 

[fpitas]

I believe the offset is from the point the acoustic centers are aligned. In other words, with the acoustic centers aligned, the woofer then moves forward (less delay) by the specified offset. At least, that approach worked for me.

 

[danvprod]

Thank you, and also I may not be reading the nomenclature properly.
Is this correct:

4th ordrer Butt Low Pass, -3dB @ Fx* 0,93 (+) -- 4th order butterworth @ 930 Hz (Fx=1000), normal polarity (+)
4th order L-R High Pass, -6dB @ Fx* 0,93 (-) -- 4th order LR @ 930 Hz, wired in negative polarity (-)
Offset = 0,31*c/Fc == offset is .31 * 341m/s / Fc (how to calculate the corner frequency, is it 930 Hz here? ==.114m?

Am I correct?

 

[fpitas]

Are we talking about this one?

http://ekladata.com/5cdaEVWofJQbgNEEfIp5Hmx0y9o.jpg

 

[danvprod]

Are we talking about this one?

http://ekladata.com/5cdaEVWofJQbgNEEfIp5Hmx0y9o.jpg

Yes, that is the one. What is that spreadsheet tool you are using? Is it available for download? Sorry for being dense here...

 

[fpitas]

That's just the jpg from jimbee's blog.

filtre "quasi-optimal" : modèles et variantes - Le blog de jimbee

There's an Excel spreadsheet available at a link near the top of the blog that you can use to experiment.

 

[danvprod]

Cheers, TY Sir.

 

[fpitas]

BTW, the way I approached it was setting things the way the blog says, then tweaked the woofer delay to get the smoothest frequency response within the crossover region. So far the best sounding one for my horn-based system is this one:

http://ekladata.com/WvgOKEH3BTEPV5td8GnR2MdMpGA.jpg

although the classic Le Cleach setup is close:

http://ekladata.com/DcAvq8mSUN74Elx9qF6-cfzVKM4.jpg

 

[danvprod]

Francis -- I don't think I can do 3rd order Bessel filters in MiniDSP with advanced bi-quads. I can just cascade 2nd order filters. I can do Butterworth 3rd orders as they are included in the stock x-over options in the MiniDSP.

So I will try the classic Le Cleach.

The other question I had is about the acoustic centers. Where/how do I measure that from. My 2226J is surface mounted to the enclosure. The depth is 5.5" according to the spec sheet. Should I use that figure and then have the rear of my 2445J 3" behind that, or 8.5" from the front of the baffle?

 

[fpitas]

I got my acoustic centers aligned by using LR 4th order (acoustic) on woofer and horn, inverting the horn signal, then setting the woofer delay for the best null at my crossover frequency. That amount of delay has been a good place to start.

 

[danvprod]

Excellent -- I will try that. I'll post the measurement results here. Should get a chance to work on this this weekend.

 

[danvprod]

With the Le Cleach setup:

http://ekladata.com/DcAvq8mSUN74Elx9qF6-cfzVKM4.jpg

Here are my measurements, what do you think?

 

[danvprod]

Here are better measurements the speakers with the crossover. I ended up adding a .2s delay to the woofer to make up the rest of the offset in the miniDSP. The summing is pretty nice and flat now, I think.

 

[fpitas]

Looking good!