DingDing goes DIY: JBL 2404 mounted inside or outside 2360A and one or two 2123?

[DingDing]

First time speaker builder, but have done subs in the past. In the process of planning and have all parts needed, less wood, but that's easy. ;) The concept is to achieve extreme dynamics from 60 Hz up, as I've got ample headroom below 60 already.

Crossover, got 12 channels from a pair of miniDSP 4x10 HD's available. Time aligning drivers will be done digitally.

The idea is using at least two 2226 per speaker, they will be sealed and EQ'ed a little bit to give good extension down to about 60 Hz. Furthermore, there will be at least one 2123. So: 2404 or 2405 + 2446 in 2360A + 2123 + 2226.

Thinking a lot about how to incorporate 2404 (or 2405) together with 2360A. I can put the uhf driver inside the cabinet for the mids (separate enclosure inside the big cabinet) and woofers or mount it inside or outside the horn.

The 2360A's has a pair of holes and I could build a jig there to place the UHF driver in, either inside or outside the chassis. Planned XO is approx 6-10khz. Not quite sure yet. Kind of like this.

Inside:
67412

Outside:
67413

My take on late night light photoshopping random pictures from the web, lol. Is this a niet, niet? Why?

Two or more 2226's will yield very high sensitivity and a single 2123 will be the limiting factor. Could I use a pair in each or would that create all sorts of problems? What if I mount them super close like I see some PA-systems does with multiples? Can they cover the same range without issues? I was hoping I could have them go to 1.3-1.6kHz approx before 2446 takes the stage.

Front loaded horn for the one or two 2123 may be an option, but can't model with hornresp, so haven't looked into it much.

Any input? Please be gentle. :D

[Lee in Montreal]

Yes. You could run two 2226 and a single 2360. With proper eq, you won't really need a tweeter. Actually, if you like sizzling highs, then maybe you could use a different horn. Maybe a Yuchi 290 and couple it to a 2405.

BTW Those 2226 needs some serious bumping around 40Hz (with Q = around 1.0) to get that nice round and deep bass. Get at least 5cft per driver.

Keep this a two-way and you won't be disapointed. Your sub(s) will fill in foe the bottom octavle. It is what I do with my 2226 and 2360

Lee

PS Bass reflex tuned to 50Hz. Cross the 2226 at around 600-700Hz to the 2360. Use a 2445 as it was designed for the 2360.

[DingDing]

Yes. You could run two 2226 and a single 2360. With proper eq, you won't really need a tweeter. Actually, if you like sizzling highs, then maybe you could use a different horn. Maybe a Yuchi 290 and couple it to a 2405.

BTW Those 2226 needs some serious bumping around 40Hz (with Q = around 1.0) to get that nice round and deep bass. Get at least 5cft per driver.

Keep this a two-way and you won't be disapointed. Your sub(s) will fill in foe the bottom octavle. It is what I do with my 2226 and 2360

Lee

PS Bass reflex tuned to 40Hz. Cross the 2226 at around 600-700Hz to the 2360. Use a 2445 as it was designed for the 2360.

Thanks for the input, I have a 2445J too, but need another one to make a pair, why is 2445 better suited than 2446 for 2360A?

I'm running two 2226's with 2360 now. It's great. XO @ 600, just like you say. The 2404 is sitting preliminary for testing, and I think it benefits the highs immensely. Current XO @ 10khz at the -6dB point. Makes the breakup/ringing of Ti less prominent.

However, my experience with 4343 says that there is great merit to that 10" 2121 driver, hence why 2123 was bought for this project. I really want to see how that driver will fit into this scheme.

[Lee in Montreal]

How did you eq the 2360?

[DingDing]

I didn't investigate it a lot because it subjectively sounded hard without any EQ.

I did do a slight negative peak filter at 10 kHz with and without a positive high shelf above, but it didn't work well for me subjectively. I didn't measure though, did it all by ear. I gave it a week until I hooked up 2404 and thought the result was substantially better. I think it's the 10khz or later Titanium ringing my ears can't stand because the peak filter helped somewhat, but not to the extent 2404 did.

[1audiohack]

I certainly agree and love the magic of a 2123. A snare played full stick will nearly crack the enamel on your teeth and 2123 will replicate that better than any other cone driver I have ever had.

Being as you already have plenty of linear overlap between the 2226's and 2360's that a single 2123 will play loud enough.

Barry.

[Lee in Montreal]

I didn't investigate it a lot because it subjectively sounded hard without any EQ.

I did do a slight negative peak filter at 10 kHz with and without a positive high shelf above, but it didn't work well for me subjectively. I didn't measure though, did it all by ear. I gave it a week until I hooked up 2404 and thought the result was substantially better. I think it's the 10khz or later Titanium ringing my ears can't stand because the peak filter helped somewhat, but not to the extent 2404 did.

"slight" is not enough, hence your decision to add a tweeter. The correction is of around 12db. When I am back from riding Mississippi, I'll give you a screenshot of the correction I did. For more details, I suggest you look into "CD correction"

Start here http://www.audioheritage.org/vbulletin/showthread.php?34120-JBL-2360-new-install/page3

[loach71]

Nice pix!

DO you like the miniDSP? Any issues with them?

[ivica]

First time speaker builder, but have done subs in the past. In the process of planning and have all parts needed, less wood, but that's easy. ;) The concept is to achieve extreme dynamics from 60 Hz up, as I've got ample headroom below 60 already. Crossover, got 12 channels from a pair of miniDSP 4x10 HD's available. Time aligning drivers will be done digitally. The idea is using at least two 2226 per speaker, they will be sealed and EQ'ed a little bit to give good extension down to about 60 Hz. Furthermore, there will be at least one 2123. So: 2404 or 2405 + 2446 in 2360A + 2123 + 2226. Thinking a lot about how to incorporate 2404 (or 2405) together with 2360A. I can put the uhf driver inside the cabinet for the mids (separate enclosure inside the big cabinet) and woofers or mount it inside or outside the horn. The 2360A's has a pair of holes and I could build a jig there to place the UHF driver in, either inside or outside the chassis. Planned XO is approx 6-10khz. Not quite sure yet. Kind of like this. Two or more 2226's will yield very high sensitivity and a single 2123 will be the limiting factor. Could I use a pair in each or would that create all sorts of problems? What if I mount them super close like I see some PA-systems does with multiples? Can they cover the same range without issues? I was hoping I could have them go to 1.3-1.6kHz approx before 2446 takes the stage. Front loaded horn for the one or two 2123 may be an option, but can't model with hornresp, so haven't looked into it much. Any input? Please be gentle. :D Hi DingDing, Why not 2242 + 2226+ 2123+ 2446&2360 +2405..... regards ivica

[DingDing]

Guys, remember that you're not talking to a savvy person here. This is a big undertaking and will be a slow and fun learning process, so if I go against conventional wisdom etc it's just to find boundaries and experience them. A bad design may give poor audio but teach you a lot about what not to do. Not saying you should stand back and let me fail miserably, but if I do something out of the ordinary or against your better judgement it's not because I don't respect your opinions, but because I want to experiment and learn. Hope you guys remember that so you don't think I'm completely a stubborn nut job, haha.


I certainly agree and love the magic of a 2123. A snare played full stick will nearly crack the enamel on your teeth and 2123 will replicate that better than any other cone driver I have ever had.

Being as you already have plenty of linear overlap between the 2226's and 2360's that a single 2123 will play loud enough.

Barry.

Haha, 'crack the enamel on your teeth' is exactly what I'm after. ;D


"slight" is not enough, hence your decision to add a tweeter. The correction is of around 12db. When I am back from riding Mississippi, I'll give you a screenshot of the correction I did. For more details, I suggest you look into "CD correction"

Start here http://www.audioheritage.org/vbulletin/showthread.php?34120-JBL-2360-new-install/page3

Read the whole thread, sorry about modding your pix! You make beautiful boxes. :) Cool that you will share the settings. Not hell bent on using 2404, but as of right now the pendulum is tilting towards it. Still have a lot of work to do and didn't even measure as I said, so that has to be done. Do like the big sound that these horns project though. There has been thoughts about Truext Be too, any reason why you didn't try it? From the measurements on here it seems you get less hf output, but less ringing which must be good, especially when used in combination with a uhf driver?


Nice pix!

DO you like the miniDSP? Any issues with them?

LOL

miniDSP is what it is, a super flexible and afforable step stone into active crossover and DSP for the noob. Sonically it's good enough for me, it lacks a display, is very unstable below 25 Hz when you run the 96kHz plugin due to rounding errors in the lower frequencies. It's more stable, but not perfect when running 48 khz plugin. This is important when you're setting your high pass filters, because the unstable filters will give you an overshoot, especially grotesque with steeper filters. Build quality is a little cheap, the software can be quirky sometimes. The lower models are very prone to noise judging from endless threads around the web. The 4x10 HD is also a bit prone to noise, but I'm using passive attenuators line level. You lose a bit of headroom, but a black background is worth it. Furthermore, the output impedance is rather high with 560 ohms. One positive thing is that you can correct things in real time, with other alternatives you have to upload a file everytime you make a change. Having two is a bit of a problem, since you can't operate them simultanously from the same PC. The software doesn't understand that you're trying to communicate with another device etc, but this is not an issue for most people as they're only using one.

I have learned a lot by using miniDSP and REW. I would not be able to do things like this if it was not for this technology, so I'm very happy. That being said, I will eventually move over to a multi channel dac such as Antelope Orion 32 or Lynx Aurora 16 with a computer to make for even more control. There are options for EQ and filters using open source Equalizer APO under windows, or you can use proprietary software such as Audiolense XO to make FIR filters and use JRiver as a convolver. pos in here also made RePhase which can be used to make FIR filters manually. And he is giving it away for free.

All in all, I think miniDSP 4x10HD is very good for the price, but as far as build quality and attention to detail (such as scalability with several units) it falls through a bit. The instability is also a concern, but the option is that they limit filters to around 20 Hz, and we don't want that. It's better to measure an output and try to extrapolate the correct settings while doing that, as the actual filter you get is not what you see in the GUI. The biquad filters are also awesome, for example if your doing a linkwitz transform or other custom filters. They will be in my system until I get a more advanced setup though, which won't be anytime soon, so yes, I think it's a great device. The company is also very professional in dealing with you.

Woha, that turned out to be a ramble. Hope it makes sense... LOL.

[DingDing]

Hi DingDing, Why not 2242 + 2226+ 2123+ 2446&2360 +2405..... regards ivica

Hi ivica.

Seen the price of 2242 lately? :D

I'm using four LMS Ultra 5400 from about 40-60 Hz and down, sealed. They are monsters and super ugly, but sq-wise they are amongst if not the best long throw drivers in the world according to many. Very happy with how they perform and they are faaar more than I need already. I think I'm covered there.

[Lee in Montreal]

Hey DingDing

Try Google Translation on this one. Basically, the guy unplugged the 2123 for the lower medium range. Not needed. My suggestion is to keep it as sinple as possible. And only add a driver if you can't really do without. Pair of 2226 and 2360 shall be enough. ;-)

http://www.melaudia.net/brufa1205-01.php

http://www.melaudia.net/zfoto/brufa1205/IMG_0060-800x533.jpg

[DingDing]

^ Thank you for the resource, did read it. It may be a mistake to incorporate it, but can't resist trying seeing how easy it is to try as I have the drivers and miniDSP. In light of this the system will be modular with everything in separate enclosures so I can just remove or reorient different drivers depending on results. This way I can try many combinations.

Reading online people seem pretty divided between multi way and 2-way. My experience with 4430 vs 4343 (w/2121h & 2235h) puts the latter on top by quite a margin, and that has been some of the inspiration behind the desire to incorporate many drivers. May even try to do something Everest and 4435 does, which is to put the LPF lower on one 2226 to up the sensitivity as frequency drops (as I understand it).

[audiomagnate]

Yes. You could run two 2226 and a single 2360. With proper eq, you won't really need a tweeter. Actually, if you like sizzling highs, then maybe you could use a different horn. Maybe a Yuchi 290 and couple it to a 2405.

BTW Those 2226 needs some serious bumping around 40Hz (with Q = around 1.0) to get that nice round and deep bass. Get at least 5cft per driver.

Keep this a two-way and you won't be disapointed. Your sub(s) will fill in foe the bottom octavle. It is what I do with my 2226 and 2360

Lee

PS Bass reflex tuned to 50Hz. Cross the 2226 at around 600-700Hz to the 2360. Use a 2445 as it was designed for the 2360.

I disagree about being able to use EQ to negate the need of a tweeter. the dispersion just isn't there. It doesn't sound right at all to my ears.

[1audiohack]

The single pole low pass on a dual woofer system can be a good idea depending on several factors including the upper crossover point and the physical seperation of the drivers involved.

If you have each woofer on a seperate DSP channel that you can also manipulate the delay between them you can really have some tuning fun.

I think everyone should build a five way active system at least once but beware, if you don't have some measurement equipment and know how to use it, the probability for frustration seems to multiply by an order of magnitude with each element added.

Barry.

[DingDing]

The single pole low pass on a dual woofer system can be a good idea depending on several factors including the upper crossover point and the physical seperation of the drivers involved.

If you have each woofer on a seperate DSP channel that you can also manipulate the delay between them you can really have some tuning fun.

I think everyone should build a five way active system at least once but beware, if you don't have some measurement equipment and know how to use it, the probability for frustration seems to multiply by an order of magnitude with each element added.

Barry.

That's motivating to read :D

The goal is to have the drivers closely together so they connect acoustically, but at the same time there's the visual aspect of having some distance between them (plus symmetry). With Everest and 4435 the distance between the drivers is pretty big. IIRC one of the 4435 woofers are crossed over @ 1khz while the the other comes in @ 100 Hz. If 2123 is incorporated 2226 will never see 1khz besides while rolling off, maybe 300-400 at most, so it may be a better idea to just run both in the same passband due to the low XO in the first place?

Because of miniDSP it will be fairly easy to try. The drivers will have a separate channel and be time aligned using REW and a calibrated Dayton EMM-6 mic using a soundcard with loopback connection to get correct timing information. Already have that measurement setup and an UMIK-1. Have not done a lot of gated measurements, so that's a hurdle that needs to be won. If the weather is good I can do it outside to get a bigger window.

Oh yes, this is a big undertaking for someone with my lack of knowledge and it's the first speaker build. At the same time that's part of the challenge and what makes it educational and fun. It may very well be that the sound sucks and I lose some hair over it, but that's OK as long as the project has been educational.

In the middle of the 4343 restoration and will begin this project once that's done. Was hoping to post some Sketchup drawings and see what you guys think and when the speakers are built I could post measurements and get some guidance on what to do/try given response, phase and so on. If it goes belly up, my 6 cats will have two great houses to play in. :D

[DingDing]

Quick questions regarding distance from 2123 to 2226.

Assuming 343m/s speed of sound, and xo @ 400 hz => wavelength = 343/400 = 0.8575m.

It has to be within 1/4 the wavelength to connect acoustically, right? So that would be 21.5 cm from the acoustic centers of the drivers. That's fairly close.

Does it need to be 21.5cm from both drivers, or the average distance between them in relation to 2123? Can make a drawing if that question is unclear.

Is it correct to use the XO point as a basis for calculating maximum distance between drivers? As the drivers overlap after the xo the distance will obviously change with frequency and can and will create lobing, right?

Can delay in the digital domain do anything to this particular problem within some constraint? Two drivers 10m apart can't sound like one driver no matter how much delay (if you're in the middle) obviously, but if we're talking about 10-20cm off from the maximum distance, will it work well?

[loach71]

miniDSP is what it is, a super flexible and afforable step stone into active crossover and DSP for the noob. Sonically it's good enough for me, it lacks a display, is very unstable below 25 Hz when you run the 96kHz plugin due to rounding errors in the lower frequencies. It's more stable, but not perfect when running 48 khz plugin. This is important when you're setting your high pass filters, because the unstable filters will give you an overshoot, especially grotesque with steeper filters. Build quality is a little cheap, the software can be quirky sometimes. The lower models are very prone to noise judging from endless threads around the web. The 4x10 HD is also a bit prone to noise, but I'm using passive attenuators line level. You lose a bit of headroom, but a black background is worth it. Furthermore, the output impedance is rather high with 560 ohms. One positive thing is that you can correct things in real time, with other alternatives you have to upload a file everytime you make a change. Having two is a bit of a problem, since you can't operate them simultanously from the same PC. The software doesn't understand that you're trying to communicate with another device etc, but this is not an issue for most people as they're only using one.

I have learned a lot by using miniDSP and REW. I would not be able to do things like this if it was not for this technology, so I'm very happy. That being said, I will eventually move over to a multi channel dac such as Antelope Orion 32 or Lynx Aurora 16 with a computer to make for even more control. There are options for EQ and filters using open source Equalizer APO under windows, or you can use proprietary software such as Audiolense XO to make FIR filters and use JRiver as a convolver. pos in here also made RePhase which can be used to make FIR filters manually. And he is giving it away for free.

All in all, I think miniDSP 4x10HD is very good for the price, but as far as build quality and attention to detail (such as scalability with several units) it falls through a bit. The instability is also a concern, but the option is that they limit filters to around 20 Hz, and we don't want that. It's better to measure an output and try to extrapolate the correct settings while doing that, as the actual filter you get is not what you see in the GUI. The biquad filters are also awesome, for example if your doing a linkwitz transform or other custom filters. They will be in my system until I get a more advanced setup though, which won't be anytime soon, so yes, I think it's a great device. The company is also very professional in dealing with you.

Woha, that turned out to be a ramble. Hope it makes sense... LOL.[/QUOTE]

Thanks for the detailed reply!

[DingDing]

^ YW! :)

---

Just read Truextent's white paper on beryllium vs aluminum vs titanium vs ribbed titanium (http://www.audioheritage.org/vbulletin/showthread.php?30412-Truextent-White-paper-for-Large-Format-Diaphragms&p=305351#post305351) in large format compression drivers, and was quite pleased to see the performance of the aluminum diaphragms. My 2446 drivers are using ribbed titanium now and looking at the spectral decay in those measurments I think it's quite clear why they are a little hard sounding. The paper seem to favor Be in applications where you're going to boost the high end because of more output and smoother response, but the decay plots indicate that if you're going to use this driver between 1.5-8/10khz range you get very uniform decay with aluminum. This is great news as they are not that costly.

What aluminum diaphragm would you guys pick for 2446? Wanting to stay clear of ribbed ones judging from the paper.

Is Radian 1245-8 Diaphragm (http://www.parts-express.com/radian-1245-8-diaphragm-fits-most-jbl-15-and-2-8-ohm--294-726) worth giving a shot?

[1audiohack]

For aluminum, the genuine JBL article, the D16R2441.

Now that said, if you use aluminum, and if you are not going to run a super tweeter, the lack of that "extra energy" of a ribbed 2450 diaphragm will require you add even more EQ to the top end and in my experience this sounds even more strained than the "noisy" 2450 diaphragms do.

Are you using a protection capacitor on the 2446's? A 10uF on an 8 Ohm version or a 20uF on a 16 Ohm will get about 6dB out of the mid band and lessen the EQ requirements for the top octave while offering some protection in the process, cheap, easy, and to me it sounds better than 18dB of boost on top.

If you are for sure going to use 2404's or 2405's with them you could use D16R2440's (still available new) in them and they willl just shut off about 9.5kHz by themselves. That's really about where that long parellel wall slot in a 2360 starts to sound rough anyway. Integrating a 2405 (what I would use) with a 2360 really well is a challenge but this is the combination I liked the best. I have done all of the above by the way.

It seems no one tests the coated smooth Ti SL's (D16R2450SL) or the D8475Nd diaphragms. Ever wonder why?

Barry.

[DingDing]

For aluminum, the genuine JBL article, the D16R2441.

Now that said, if you use aluminum, and if you are not going to run a super tweeter, the lack of that "extra energy" of a ribbed 2450 diaphragm will require you add even more EQ to the top end and in my experience this sounds even more strained than the "noisy" 2450 diaphragms do.

Are you using a protection capacitor on the 2446's? A 10uF on an 8 Ohm version or a 20uF on a 16 Ohm will get about 6dB out of the mid band and lessen the EQ requirements for the top octave while offering some protection in the process, cheap, easy, and to me it sounds better than 18dB of boost on top.

If you are for sure going to use 2404's or 2405's with them you could use D16R2440's (still available new) in them and they willl just shut off about 9.5kHz by themselves. That's really about where that long parellel wall slot in a 2360 starts to sound rough anyway. Integrating a 2405 (what I would use) with a 2360 really well is a challenge but this is the combination I liked the best. I have done all of the above by the way.

It seems no one tests the coated smooth Ti SL's (D16R2450SL) or the D8475Nd diaphragms. Ever wonder why?

Barry.

Wow! Thank you! You sure know your stuff :D This is a new world for me and the rabbit hole goes deep too, you even match the numbers on the diaphragm to the driver and optimize the position. Yikes! :D Based on your response here I found another thread you were active in, namely this one: http://www.audioheritage.org/vbulletin/showthread.php?36992-2440-vs-2441-Diaphragm-in-2445-driver where you describe the matching and optimizing process.

It looks like the D16R2441 is preferred to the D16R2440 because of the versatility of more hf output. No need to have the driver roll of earlier and the added sensitivity of the latter is not all that important in a home environment I assume, especially when you can lowpass it digitally anyways? Another thing would be if 2360A is replaced with something else and you want some extra wiggle room for the xo. That's an argument to just get the Truext. Be too...

No protection, no kids, :) Not long since I got them and been running without a cap. Been reading up on cap or no cap, and a lot of folks have run without it for years and have had no problems. When my amps power off, I simply let the music play and they slowly fade out, so no dc, crackle or pop when turned on or off. Protecting against power outtage, maybe put a UPS on that amp.

Really tempted to give the D16R2441 a try. Seem like they hoover around $200 each new from various retailers, maybe possible to get a used pair in the marketplace here too.

2405 really "belongs" in 4343, but I can use it to see how it fares compared to 2404 and get a second pair if they're that much better. Some people say they prefer 2404 with 2360A, so hoping for that.

All this being new, guessing based on that other thread the D16R2450SL may not be popular because of the added mass of the aquaplas and as a result possibly poorer impulse response?

[1audiohack]

Hi;

I don't run caps for protection either, not even on my Be loaded drivers but on these large format constant directivity horns I find that they sound better to me with a single cap sized right for some mid band attenuation and less agressive EQ.

If you look at the directivity charts for the 2360 you will note the collapse in the vertical response to about 20-25 degrees above 10kHz. You can use a 2404 crossed in right about 9-10kHz to maintain the directivity but it is just hard to get a 100X100 degree super tweeter to play that loud and do you really need that energy bouncing off the ceiling? This is about the vertical dispersion pattern of a 2405 at 10K anyway and since you can't really use a 2404 in the center of the mouth of the 2360, I have had better luck integrating the 2405. Below the 2360 as you have pictured.

It is of interest to me that JBL used the D8R475 in the M9500 horn in a 2450 motor with no super tweeter, in a statement speaker system. They measure and sound very good too. I am not trying to sell you anything, it is just my opinion that those diapragms are dynamite for Only $210.ea USD.

Barry.

[audiomagnate]

Hi;

I don't run caps for protection either, not even on my Be loaded drivers but on these large format constant directivity horns I find that they sound better to me with a single cap sized right for some mid band attenuation and less agressive EQ.

If you look at the directivity charts for the 2360 you will note the collapse in the vertical response to about 20-25 degrees above 10kHz. You can use a 2404 crossed in right about 9-10kHz to maintain the directivity but it is just hard to get a 100X100 degree super tweeter to play that loud and do you really need that energy bouncing off the ceiling? This is about the vertical dispersion pattern of a 2405 at 10K anyway and since you can't really use a 2404 in the center of the mouth of the 2360, I have had better luck integrating the 2405. Below the 2360 as you have pictured.

It is of interest to me that JBL used the D8R475 in the M9500 horn in a 2450 motor with no super tweeter, in a statement speaker system. They measure and sound very good too. I am not trying to sell you anything, it is just my opinion that those diapragms are dynamite for Only $210.ea USD.

Barry.

Just a data point, I don't use protection caps because I have tested both ways and I swear they change the sound. I have had an active four way system without protection for years and never had a problem. I'm careful and no one else touches my system.

[DingDing]

Hi;

I don't run caps for protection either, not even on my Be loaded drivers but on these large format constant directivity horns I find that they sound better to me with a single cap sized right for some mid band attenuation and less agressive EQ.

If you look at the directivity charts for the 2360 you will note the collapse in the vertical response to about 20-25 degrees above 10kHz. You can use a 2404 crossed in right about 9-10kHz to maintain the directivity but it is just hard to get a 100X100 degree super tweeter to play that loud and do you really need that energy bouncing off the ceiling? This is about the vertical dispersion pattern of a 2405 at 10K anyway and since you can't really use a 2404 in the center of the mouth of the 2360, I have had better luck integrating the 2405. Below the 2360 as you have pictured.

It is of interest to me that JBL used the D8R475 in the M9500 horn in a 2450 motor with no super tweeter, in a statement speaker system. They measure and sound very good too. I am not trying to sell you anything, it is just my opinion that those diapragms are dynamite for Only $210.ea USD.

Barry.

Ok, will try with and without a cap then. Was not intending to use a cap even if Be is chosen because of what audiomagnate says in regards to SQ. It would suck big time if they were blown, but at the same time YOLO and it feels wrong to spend that much money to get the best diaphragm and limit the result with a cap.

Good points regarding the vertical dispersion and integration of the tweeters, had not thought about that.

The reason I got 2360A was that they came with the 2446's and was a bargain, about $170 each with the original 2506b brackets and everything. All in very good condition too, although the horns have some minor scratches and stuff which will be fixed before spraying. Point being that 2360 may be substituted for something else, as I was really after the 2446 drivers. This makes me inclined to jump on the Be diaphragms, because those will give more flexibility if this project fails and the drivers will be used in another application where more hf output is preferred. It seems like most people gravitate towards those diaphragms and everything else is just a temporary stop. Trying out a bunch of different ones may well become more expensive than just getting the best sooner rather than later?

D8R475 is interesting too, but I must admit that the most enticing to me right now is your recommendation for D16R2441 and ofc. Truextent's Be. D16R2441 because both 4343 (w/2420) and 4430 (w/2421A) use aluminum and I really like the sound from those. Be because I've yet to see one negative thing mentioned about them except for the price.

[ivica]

For aluminum, the genuine JBL article, the D16R2441.

Now that said, if you use aluminum, and if you are not going to run a super tweeter, the lack of that "extra energy" of a ribbed 2450 diaphragm will require you add even more EQ to the top end and in my experience this sounds even more strained than the "noisy" 2450 diaphragms do.

Are you using a protection capacitor on the 2446's? A 10uF on an 8 Ohm version or a 20uF on a 16 Ohm will get about 6dB out of the mid band and lessen the EQ requirements for the top octave while offering some protection in the process, cheap, easy, and to me it sounds better than 18dB of boost on top.

...

It seems no one tests the coated smooth Ti SL's (D16R2450SL) or the D8475Nd diaphragms. Ever wonder why?

Barry.

Hi Barry,

Very interesting experience and suggested way of using CD horns with 4" diaphragm JBL driver.
Especially as You prefer 2450 ribbed Ti vs 2441 aluminum (either used without UHF driver).
What about 2445 (non-ribbed) Ti diaphragm ? Is it somewhere "in-between" these two ?
Some of the measurements I have done either SL or Be (Trex) decline the FR response over 10kHz more
rapidly, so I believe that the large amount of EQ would be needed if wanted to reach 16kHz.
Ribbed diaphragm seems to operate up to 12~13kHz before break, I wonder what would be their behaviors
if being SL-coated.


regards
ivica

[dezmond]

I'm doing the same thing. I am gonna try 2 x 2404's under ea 2360 .

[DingDing]

I'm doing the same thing. I am gonna try 2 x 2404's under ea 2360 .

How will you mount them? I was thinking about making some sort of wooden jig that can be attached to the holes in the 2360 and then have 2404 mounted to that.

Basic concept

67695

This way the cables for 2404 can be hidden behind 2360. Have also been thinking about making the mounting piece for 2404 detachable from the 2360 mount so I can make new ones depending on the driver I end up with. Maybe also allow for rotation to adjust positioning through measurements from sweetspot.

[dezmond]

I will show a picture later tonight. I have them on each side of the 2506b bracket tilted down slightly. I am using jbl as4731 cabinets .

[dezmond]

Here some pictures -

[dezmond]

Not saying its the best way, just trying it like this.

[DingDing]

Thank you for the pictures. Do you like the result?

I will build that bracket to have 2404 mounted just below the mouth and in the same plane as 2360. Then delay 2404 to account for the distance between 2446 and 2404. Will post pictures of the bracket and may 3D model it if you're interested. What concerns me about having them behind the mouth is that a lot of the frequencies will be bounced off the rare of the 2360.

[Ian Mackenzie]

Very interesting thread

[Ian Mackenzie]

Quick questions regarding distance from 2123 to 2226.

Assuming 343m/s speed of sound, and xo @ 400 hz => wavelength = 343/400 = 0.8575m.

It has to be within 1/4 the wavelength to connect acoustically, right? So that would be 21.5 cm from the acoustic centers of the drivers. That's fairly close.

Does it need to be 21.5cm from both drivers, or the average distance between them in relation to 2123? Can make a drawing if that question is unclear.

Is it correct to use the XO point as a basis for calculating maximum distance between drivers? As the drivers overlap after the xo the distance will obviously change with frequency and can and will create lobing, right?

Can delay in the digital domain do anything to this particular problem within some constraint? Two drivers 10m apart can't sound like one driver no matter how much delay (if you're in the middle) obviously, but if we're talking about 10-20cm off from the maximum distance, will it work well?

Long time ago but basically yes but its not overly critical

Dispersion match at the crossover point has more impact than critical spacing.

The horn will nearly always loose pattern control at the low end.

So if its an 80 degree horn it might flare out to 100 near the bottom end. That can help if the driver load and distortion is not rising.

Use -6 db as the reference for the dispersion

In fact you can compromise other things by trying too hard on clustering drivers like diffraction and early reflections.

I could not get the 2344A close to the 2123 without reflections.

An easy but relative precise way of determining absolute phase (and lobes) in the crossover region is to reverse the phase on one driver once you have the slopes set up.

Unless you're using DSP filters there is going to be phase shift and delay in the low pass filter and phase shift delay around the low end of the horn.

So the voice coil centres are not likely to be the reference for absolute phase.

With an analyser measure the notch at the crossover and gradually move the voice coil centres until you get an even deep notch at your reference point (the mic position).

The move the mic in the vertical plane to work out the lobes.

LR 24 db filters really shine in terms of minimal lobes.

Its not possible without an optimiser like LEAP for passive networks but you can shoot for 36 db slopes in digital with DSP and the lobes are significantly reduced around crossover point (ref M2)

Is it important? Only if your horn has wide vertical dispersion in the proposed crossover point .

With a narrow vertical dispersion on the horn this can actual help minimise vertical lobes

Check out Pi Loudspeakers white papers if you are interested in the topic.

Wayne is helpful and a nice guy.