I measured the time alignment of JBL 2123 and 2420/2344

I have put it on here somewhere, but my system consists of a 2235 bass (4570 cabinet I think it is), 2123 mid and 2420/2344 on the highs. All are actively crossed over using a M553 with CD compensation.

So, the kids were asleep, the wife was out, so there was time to finally redo some measurements. Earlier I had estimated that the high horn had to move 11cm forward to align the acoustic centre with that of the mid. I came at this by feeding them both pink noise and connecting them in opposite phase and then shifted them to get the biggest suckout. I tried with impuls measurements, but I could not get it to work in an understandable manner.

So I rethought the setup and measurement setup and I think I nailed it. Important was to feed both drivers with the same signal (600Hz and up in this case), and to create about the same impuls levels. I made the measurement at 1.5m away halfway between mid and high and using pink noise. I fed the signal from my laptop into my Marantz CD7 digitally by an USB> SPDIF convertor and the Shure dynamic mic was coupled to my laptops mic input. I use an active 4th order linkwitz-riley crossover, so there are no components other than speaker cable between the amps and the drivers.

The screenshot below shows a yellow line and a green line.

The yellow is where the mid and high are both aligned at the front of the mid speaker box. You can see that there are two impulses and that the impuls of the high lags behind that of the mid.

The green is where the high horn has been moved 10cm forward to align its acoustic centre with that of the mid. As I use an active 4th order linkwitz-riley setup this is what I need.

I can't say anything about the absolute "quality"of the impuls responses as 1) I am not too familiar with that and 2) I don't know what settings "professionals" use and 3) I measured just how they were positioned in my room, which could be better I think, no "free space" optimization etc. My only goal was to time align the two drivers and check for polarity correctness.

https://lh6.googleusercontent.com/-O...52520600Hz.jpg

These show btw that the JBL convention of making the black input terminal positive is according to spec with these drivers. Remember to check your JBL drivers for that as not all JBL drivers are made like this. There is a table somewhere stating what JBL driver is made in what way.

The mid is a JBL 2123H with an efficiency of about 100dB/Watt. The high is a JBL 2420 compression driver on a JBL 2344 horn. Bass is a 15" JBL 2235H. Normal crossover points are 4th order 180Hz and 1600Hz with a CD compensation on the high from about 3000Hz compensating for the horn fall-off towards 16kHz.

This is the setup now. The supports look cheap, and they are, but they were the most acoustic dead ones I could find (the aluminium ones were looking much nicer, but kept ringing when you tapped them) :-)

https://lh3.googleusercontent.com/-z...to%2525201.JPG

So... the more important question. Do you like the sound better? Time/phase aligning can have profound changes to the sound eliminating suck outs and smoothing the response. It can also give you response bumps if you had previously designed the network to compensate for these suck outs.

From my experiments and experience, time alignment has very little to do with imaging in spite of all of the marketing hype that stated it was essential. We don't hear about that so much these days, but some manufacturers still claim that it is important for imaging. I am not saying that it isn't a good thing, simply that it doesn't seem to correlate with imaging as much as patter control and a few other aspects I haven't been able to quantify.


Widget

Quote:

---

Originally Posted by frank23

I have put it on here somewhere, but my system consists of a 2235 bass (4570 cabinet I think it is), 2123 mid and 2420/2344 on the highs. All are actively crossed over using a M553 with CD compensation....So I rethought the setup and measurement setup and I think I nailed it....The screenshot below shows a yellow line and a green line.

https://lh6.googleusercontent.com/-O...anaf 600Hz.jpg

https://lh3.googleusercontent.com/-z...800/foto 1.JPG

---

Very nice job of aliening both voice coils! However if you measure the frequency response in the cross over spectrum in the horizontal angles, you will find a disparity between the (Radial or Bi-Radial?) high frequency horn "wave front" and the "simi-hemispherical wave front" of the low frequency driver.

This problem has been addressed in two-way horn designs using radial horns for both the high frequency & low frequency horns with simple 2nd order 500 Hz L /C cross over networks. Also [quote Mr. Widget...."importance of Pattern control"....]

Unfortunately this requires a huge chunk of real estate; but with less than 5 watts drive __ that huge wall of sound is awesome! herki[Quote/]

That is of course the most important question. Does it sound better? I must say that due to kids running around the house I have not had an evening of critical listening for imaging in the before and after situation. But the before situation (horn aligned with the mid front) at least sounded fine for casual listening.

Because I use the M553 active crossover, I have to have time alignment to have the crossover work correctly. Otherwise the whole Linkwitz Riley setup is not optimal. But indeed, I have introduced quite some reflection opportunities by moving the horn so much forward that some of the mid reflects on the back of it.

I think a response without suckouts is more important than time alignment, but in this 4th order linkwitz-riley setup they go hand in hand. Would I have had a lower order passive crossover I would optimize response and patterns before time alignment I think.

I should do some response measurements, but haven't figured out how to do them reliably.

Quote:

---

Originally Posted by Mr. Widget

So... the more important question. Do you like the sound better? Time/phase aligning can have profound changes to the sound eliminating suck outs and smoothing the response. It can also give you response bumps if you had previously designed the network to compensate for these suck outs.

From my experiments and experience, time alignment has very little to do with imaging in spite of all of the marketing hype that stated it was essential. We don't hear about that so much these days, but some manufacturers still claim that it is important for imaging. I am not saying that it isn't a good thing, simply that it doesn't seem to correlate with imaging as much as patter control and a few other aspects I haven't been able to quantify.


Widget

---

Can't forget the lobing effect of horizontal displacement. Maybe not an imaging issue but certainly it does no good to have a nice impluse response graph only to find that the xover topology you settled on throws the "sweet" spot to your feet.

That would sure be a question in a setup with a normal passive crossover. But I use the active JBL M553 crossover which uses a 4th order Linkwitz-Riley setup. And that setup needs time alignment to function correctly with the 0dB lobe straight ahead. When one driver is delayed in comparison to the other, the 0dB lobe goes in that drivers' direction, mostly up as the acoustic center of the high horn is normally behind that of the mid or woofer.

This figure from the Rane site illustrates this (right situations shows lobe when time corrected, left when not time corrected):
http://www.rane.com/n160fig7.gif

http://www.rane.com/note160.html

But there is much more written about this. In my setup with mid and high at ear height, and an active 4th order LR crossover, time-alignment is a good thing (apart from extra reflections that is...).

Of course "time alignment" by driver staggering is only aligned for one finite point in space so your listening position needs to be coincident with that as well.

As stated previously, it is all about the compromises. :)


Widget

Quote:

---

Originally Posted by frank23

That would sure be a question in a setup with a normal passive crossover. But I use the active JBL M553 crossover which uses a 4th order Linkwitz-Riley setup. And that setup needs time alignment to function correctly with the 0dB lobe straight ahead. When one driver is delayed in comparison to the other, the 0dB lobe goes in that drivers' direction, mostly up as the acoustic center of the high horn is normally behind that of the mid or woofer.

This figure from the Rane site illustrates this (right situations shows lobe when time corrected, left when not time corrected):
http://www.rane.com/n160fig7.gif

http://www.rane.com/note160.html

But there is much more written about this. In my setup with mid and high at ear height, and an active 4th order LR crossover, time-alignment is a good thing (apart from extra reflections that is...).

---

The Rane site presents magnificent engieering, but I worry about all the complexity required to implement the 4th order Linkwitz-Riley hardware and the effect these additional electrical components force on the original sound.

[Quote: Rane].... http://www.rane.com/note160.html ....

A Perfect Crossover

A perfect crossover, in essence, is no crossover at all. It would be one driver that could reproduce all frequencies equally well. Since we cannot have: that, second best would be multiple speakers, along the same axis, with sound being emitted from the same point, i.e., a coaxial speaker that has no time shift between drivers. This gets closer to being possible, but still is elusive....

Not really! Consider the original RCA DUO CONE LC-I Speaker with a 15 inch extremely rigid low frequency Cone __"the first RCA LC model without the Bumps on the 15 inch cone" __ which is capable of Excellent Linearity well beyond 1,000 Hz, designed to roll off 6 db /octave from the 1,000 Hz "mass break point" without benefit of the typical external first-order inductor.

This L F cone has a two inch voice coil surrounding a very linear low distortion coaxial 2 inch cone tweeter with excellent low distortion well beyond 15,000 Hz via a simple first-order paper 4 MFD Capacitor. With the 2 inch tweeter nested in the LF cone voice coil, the conical 15 inch cone provides a very useful wave guide for the 2 inch tweeter, completely free of any doppler modulation.

Essentially, the 2 inch HF cone is a coaxial extension of the massive 15 inch LF cone providing a perfect acoustical crossover. herki[Quote/]

Quote:

---

Originally Posted by Mr. Widget

Of course "time alignment" by driver staggering is only aligned for one "time alignment so your listening position needs to be coincident with that as well.... :) Widget

---

[Quote..."your listening position needs to be coincident with..."] Since recording engineers do work in a "finite space" facing their control console & monitors, there should be no problems achieving an ideal monitor location. I have heard these monitors in the New York City recording studios including the 1970's last "state of the art" RCA Studios redesigned by JBL Engineers. The JBL Monitors there were accurate & sweet. .

IMHO, the main problem we have at home is "to duplicate the acoustic environment in the Recording Mastering Engineer's space." You need to retrieve all the original room reverberation experienced by the mastering engineer in mixing those recordings.

There is an excellent LH Forum Thread on this subject you can search. herki[Quote/]

Hi Frank,

So you prefer the 2420/2344 combo over the 2435/PT ?

PS: if you decide to sell your PT1010s I might be interested :D