PDA

View Full Version : 3-way JBL inspired build - 123A - 2420/2307 - supertweeter - DSP crossover



TimVG
08-03-2017, 08:30 AM
Hello all

I'll be logging my new project here.
Parts used will be the JBL 123A - JBL 2420 compression driver and 2307 horn, combined with a supertweeter. Possibly not JBL but a Visaton TL 16H. In which case I'd be crossing over at 1200hz and 7500hz through a DSP.
Still drawing out the cabinet.. Any tips or input are of course appreciated.

Best

Tim

Mr. Widget
08-03-2017, 08:36 PM
Investigate woofer/box tuning. I think if you use a box that is a bit larger than the L100 it would likely be an improvement over the stock L100. I would also bump the tweeter crossover up to 10KHz.


Widget

TimVG
08-04-2017, 12:28 AM
Investigate woofer/box tuning. I think if you use a box that is a bit larger than the L100 it would likely be an improvement over the stock L100. I would also bump the tweeter crossover up to 10KHz.


Widget


I figured as much. Still drawing things out, but my idea at the moment for cabinet, is a 50% (approximately) downscaled version of the 4343B model which should yield a volume of around 75l. I'm aiming for 65l~ net volume. Still have to model the driver with regards to tuning.

I'll definitely expirment with the crossovers as well. Since it's all DSP controlled it won't take long at all.

Best

Tim

TimVG
08-04-2017, 03:58 AM
So, I've simulated the 123A-1 in a 70l cabinet tuned to 40hz and sealed. I'm doubting the numbers, but if they are correct, that is a 5db bump centered around 60hz and I'm inclined to go sealed here. These will be placed relatively close to a backwall as well. Thoughts?


77748

TimVG
08-05-2017, 10:56 AM
Was able to source another pair of 2420/2307 in combination with a pair of 2405 tweeters :) .. So the Visaton tweeters are out and we're going full JBL:D

https://s1.postimg.org/cgar9nbrv/jbl.png (https://postimg.org/image/cgar9nbrv/)

Oh.. And the mint 123A-1 pair I was able to get ..

https://s1.postimg.org/vos8ii9x7/57_1.jpg (https://postimg.org/image/vos8ii9x7/)

TimVG
08-17-2017, 09:03 AM
Hello all

I'm currently cutting the panels for this project.
For the tweeter and horn cutouts: I'm finding conflicting info. Sizes range from 3 1/16" to 3 3/16" for the tweeter and 4" to 4 1/4" for the 2307 horn.. My question is if someone has the definite answer? I'd like to get it right the first time.

Best

Tim

Ian Mackenzie
08-17-2017, 01:56 PM
Hello all

I'm currently cutting the panels for this project.
For the tweeter and horn cutouts: I'm finding conflicting info. Sizes range from 3 1/16" to 3 3/16" for the tweeter and 4" to 4 1/4" for the 2307 horn.. My question is if someone has the definite answer? I'd like to get it right the first time.

Best

Tim

http://www.audioheritage.org/vbulletin/showthread.php?2250-2307-2311-2312-Bolt-Pattern

emilime75
08-17-2017, 04:06 PM
I've found certain JBL parts, with the same part numbers, can vary in dimensions. You'd be better off measuring your actual parts and cutting accordingly.

Ian Mackenzie
08-17-2017, 04:28 PM
I think you will find the template accurate.

The hole needs to be slightly larger than the diameter of the mouth.

More important are the location of the holes in the mounting flange

http://www.audioheritage.org/vbulletin/showthread.php?2250-2307-2311-2312-Bolt-Pattern&p=19608&viewfull=1#post19608

TimVG
08-18-2017, 05:32 AM
Thanks for all the feedback!

TimVG
08-18-2017, 11:53 AM
Got some assembling done.

https://s27.postimg.org/6nobs9tgf/20991300_10211874020940407_10198742_o.jpg (https://postimg.org/image/6nobs9tgf/)

https://s27.postimg.org/5mo335cgv/20960846_10211874870681650_1562308562_o.jpg (https://postimg.org/image/5mo335cgv/)


Next up is the baffle and back panels, sanding, making sure the whole thing is airtight.. TBC

TimVG
08-23-2017, 11:19 AM
I patinated the plywood enclosures with a mix of vinegar and steel wool.

https://s28.postimg.org/jtzo8cx7t/20170821_082501.jpg (https://postimg.org/image/jtzo8cx7t/)

Afterwards I put on a grey wash with a brush, and sanded in three steps from coarse to fine with the grey wash still semi wet. This gives a nicer overall look than just using a brush.

https://s28.postimg.org/5c72e3vah/20170822_081733.jpg (https://postimg.org/image/5c72e3vah/)

After letting the cabinets sit to dry out completely I once again did some sanding, to bring out the grain.

https://s28.postimg.org/9o9wrpnah/20170822_081817.jpg (https://postimg.org/image/9o9wrpnah/)

TimVG
08-24-2017, 10:58 AM
To finish off the cabinets, I'm gonna apply some varnish.

This gives a nice protective layer, and brings out the contrast between the wood and greywash.

https://s29.postimg.org/7x3ga55wz/20170824_190048.jpg (https://postimg.org/image/7x3ga55wz/)

TimVG
09-08-2017, 03:12 AM
Baffles finished in RAL 8019


https://s26.postimg.org/lkw5hszpx/20170908_095157.jpg (https://postimg.org/image/lkw5hszpx/)

https://s26.postimg.org/dgo1d2dat/20170908_095144.jpg (https://postimg.org/image/dgo1d2dat/)

Test fit

https://s26.postimg.org/dutyqems5/20170908_120102.jpg (https://postimg.org/image/dutyqems5/)

TimVG
09-11-2017, 09:14 AM
So far so good..

https://s26.postimg.org/5sveajcyd/20170911_163357.jpg (https://postimg.org/image/5sveajcyd/)

TimVG
09-16-2017, 10:09 AM
https://s26.postimg.org/fgc1jyf6d/20170914_191110.jpg (https://postimg.org/image/fgc1jyf6d/)

https://s26.postimg.org/rours1t91/20170916_185946.jpg (https://postimg.org/image/rours1t91/)

https://s26.postimg.org/ches7p1ed/20170916_185813.jpg (https://postimg.org/image/ches7p1ed/)

TimVG
09-18-2017, 01:52 PM
https://s26.postimg.org/f864fjzo5/20170917_162923.jpg (https://postimg.org/image/f864fjzo5/)

https://s26.postimg.org/42btelphx/20170917_180625.jpg (https://postimg.org/image/42btelphx/)

https://s26.postimg.org/rhtqjy991/20170918_104357.jpg (https://postimg.org/image/rhtqjy991/)

https://s26.postimg.org/3te8oocph/20170918_224400.jpg (https://postimg.org/image/3te8oocph/)

Ed Zeppeli
09-18-2017, 02:44 PM
They look fantastic. How are you setting up the Xover?

TimVG
09-19-2017, 05:36 AM
They look fantastic. How are you setting up the Xover?

Thanks Ed. 1250hz 12db/octave between the 123A-1 and the 2420/2307. And 8000hz between the former and the 2405. 24db on the 2420 and 12db on the 2405. I tried a bunch of different combinations. Most measured well but this one sounded most natural, I also like the wider dispersion of the 2405 so I'm crossing it slightly lower than the standard 9,5-10k. What made a big audible difference (but not so much in terms of measurements - probably since I don't have a high-end mic and anechoic chamber) was time aligning the drivers. I've never heard a big 43xx series, and while I've no doubt they sound good, a DSP controlled crossover can really make a difference here in my opinion.

TimVG
09-22-2017, 06:49 AM
Crossover update. 7500hz between 2420 and 2405 with a whopping 36db/octave gives the flattest on- and off-axis (both in the horizontal and vertical plane).
Changing the slope and/or moving the crossover to anything else seems to give not as good as a response. Perhaps some minor interference issues that are absent with this steep filter.

I will post some measurements soon enough.

ivica
09-24-2017, 07:27 AM
Crossover update. 7500hz between 2420 and 2405 with a whopping 36db/octave gives the flattest on- and off-axis (both in the horizontal and vertical plane).
Changing the slope and/or moving the crossover to anything else seems to give not as good as a response. Perhaps some minor interference issues that are absent with this steep filter.

I will post some measurements soon enough.

Hi TimVG,

I have some questions:
1. Did You apply time delay compensation between the drivers (concerning 2420&2307 delay, relative to the 2405 )?
2. Did You try LR 24db/oct LR slope filters while applying time delay compensation)?

regards
ivica

TimVG
09-24-2017, 08:48 AM
Hi ivica

Yes, the units are time alligned with respect to the position of the voice coils, but also with respect to ear height relative to the units (between 2307 and 2405).
I also tried the 24db filter but there is still some interference between the 2420 and 2405 that way. I assume this there is little to be done about this since the wavelength for 7500hz is around 4,5cm, and the distance between the centre of both drivers is far greater than that. So there is bound to be some interference.

While I'm technically not knowledgeable enough to give a definite answer, I assume the sharp dropoff (combined with the phase shift of a 36dB filter?) gives a smooth response on and off axis nonetheless. The crossover is inaudible (both are actually!) so I'm not worried I should be implementing a different filter.. although there's no harm in experimenting of course.

Best

Tim

bldozier
09-24-2017, 09:45 AM
the combination kind of looks delicate with out lpads on the baffle.

did you consider a vintage electronic network, maybe a difference in sound, id like to know which you like better?

TimVG
09-24-2017, 12:30 PM
No, I just have the electronic crossover - not planning on doing passive.

TimVG
09-25-2017, 05:17 AM
https://s26.postimg.org/7130mh1qt/image.png (https://postimg.org/image/7130mh1qt/)

Left and right speaker average response of a couple of measurements around the listening position. There are a couple of filters in place to dampen some resonances in the woofers. No EQ whatsoever on either the 2420 or 2405.

TimVG
09-25-2017, 12:57 PM
Fooled around a bit further with time alligning, gain, .. dark arts..

While the individual response of the speakers hasn't altered much, there is now only one filter in place on both woofers to take out some excess energy, the combined sweep of both speakers has improved, and it's definitely audible.

https://s26.postimg.org/ihy8mpul1/samen.jpg (https://postimg.org/image/ihy8mpul1/)

Couple of peaks still (keep in mind the 2dB resolution of the graph) and I will try to take care of them next weekend, and do some A/B testing to see which I like better. Thing are sounding great as-is.

ivica
09-25-2017, 01:14 PM
Fooled around a bit further with time alligning, gain, .. dark arts..

While the individual response of the speakers hasn't altered much, there is now only one filter in place on both woofers to take out some excess energy, the combined sweep of both speakers has improved, and it's definitely audible.

https://s26.postimg.org/ihy8mpul1/samen.jpg (https://postimg.org/image/ihy8mpul1/)

Couple of peaks still (keep in mind the 2dB resolution of the graph) and I will try to take care of them next weekend, and do some A/B testing to see which I like better. Thing are sounding great as-is.

Hi TimVG,

Unusual sharp roll-off of 2405 (over about 14kHz).

regards
ivica

Ed Zeppeli
09-25-2017, 01:22 PM
Just and FYI. You can upload a jpg right from your hard drive and it will be stored here. Good thing to do as some links/photo accounts go dead in time.

Cheers

TimVG
09-25-2017, 01:25 PM
Hi TimVG,

Unusual sharp roll-off of 2405 (over about 14kHz).

regards
ivica


I'm not worried about it. Probably a combination of the high energy loss of the omnidirectional microphone (UMM-6) I use, combined with the fact that the tweeters are about 15° off axis (vertical)

TimVG
10-27-2017, 01:59 AM
While I'm completely loving this system. I always try to get my hands on a spare part - just in case. Since (in terms of max SPL) the weakest link is the 123A.. Guess what I bought.

https://s1.postimg.org/1lh7s9phvf/20171027_105340.jpg (https://postimg.org/image/1lh7s9phvf/)

TimVG
11-01-2017, 01:43 PM
Today I installed the 2213nd-1 pair into my cabinets. The difference is not subtle and required a total recalibration of the system. Between 100hz and crossover (1250hz) the system has gained about 4db in sensitivy - Below 100hz the balance shifts and the 123A pair ends up around 2db louder at 50hz. With some recalibration and some bottom end EQ I ended up with more or less the same frequency response, with slightly more emphasis on the mid-range, which was ever so slightly lean before. The 2213nd sounds less colored to my ears and while tuning the system I was able to get away with a 1dB increase between 300 and 900hz (which flattened the response compared to the above register).

I may modify the cabinets in the future and turn them into ported cabinets as I feel this will compliment the new woofers, on the other hand, even with some EQ I'm not running out of headroom.

berga12
11-02-2017, 03:56 AM
Step by step...time by time....

you'll be placing a 15"/18" ported below your nice cabinets, and replace the 123A (despite is a nice woofer) with a 2202 :)

Or remove the 12" for a single 15" if you have enough liters (see 4333)

TimVG
11-02-2017, 04:54 AM
Step by step...time by time....

you'll be placing a 15"/18" ported below your nice cabinets, and replace the 123A (despite is a nice woofer) with a 2202 :)

Or remove the 12" for a single 15" if you have enough liters (see 4333)

Already replaced the 123A with the 2213nd! :) See my post above.

berga12
11-02-2017, 05:52 AM
I know, but it's just a matter of time... :) you'll change again!

2213 can't support in a proper way a compression driver and the slot tweeter...

you would need a 15" in addition.

TimVG
11-02-2017, 06:12 AM
Match in timbre is good between the compression driver and the 2213nd. I agree one or two 15" woofers would lift the system to a different level, but there is a size constraint. I still have two subwoofers in the back with each 2 x 13,5" drivers and 1kw per box :-)
I also have a 4430 replica in the works..

berga12
11-02-2017, 06:42 AM
wow nice!

open a 3D on the 4430, I would like to start also with this build, I've some doubt to find every Crossover component!

TimVG
12-27-2018, 06:44 AM
Latest measurements of my system. Purchased a Dayton DSP 408 to fool around with, which is a relatively cheap unit but with potential nonetheless.

JBL 2213nd (24dB octave low-pass @ 1000hz)
JBL 2420 on 2307/2308 horn/lens combination (18dB high-pass electrically @ 1000hz - modified voltage drive between 1khz and 2khz - no low-pass)
JBL 2405 (24dB high-pass @ 16khz)

SVS subwoofer (yet to finetune) takes over below 60hz.

All units time electronically time aligned to the vertical axis of the 2420 (So not aligned at the listening position, but rather as you would simply align the units physically).

Please note filters are electric and the acoustic slopes differ slightly.

Listening window (0° - 10° - 20° - 30° average) @ 1m at acoustic axis between 2213nd and 2308.

https://i.ibb.co/4trP8jx/0tot30.jpg (https://ibb.co/pwFJf3p)

L/R speaker average around listening position + L/R average with target curve.

https://i.ibb.co/VHmpDvx/av.jpg (https://ibb.co/s5HJQFC)
https://i.ibb.co/D4VW4Gq/targ.jpg (https://ibb.co/v1wq1st)

ivica
12-27-2018, 11:07 PM
Latest measurements of my system. Purchased a Dayton DSP 408 to fool around with, which is a relatively cheap unit but with potential nonetheless.

JBL 2213nd (24dB octave low-pass @ 1000hz)
JBL 2420 on 2307/2308 horn/lens combination (18dB high-pass electrically @ 1000hz - modified voltage drive between 1khz and 2khz - no low-pass)
JBL 2405 (24dB high-pass @ 16khz)

SVS subwoofer (yet to finetune) takes over below 60hz.

All units time electronically time aligned to the vertical axis of the 2420 (So not aligned at the listening position, but rather as you would simply align the units physically).

Please note filters are electric and the acoustic slopes differ slightly.

Listening window (0° - 10° - 20° - 30° average) @ 1m at acoustic axis between 2213nd and 2308.

https://i.ibb.co/4trP8jx/0tot30.jpg (https://ibb.co/pwFJf3p)

L/R speaker average around listening position + L/R average with target curve.

https://i.ibb.co/VHmpDvx/av.jpg (https://ibb.co/s5HJQFC)
https://i.ibb.co/D4VW4Gq/targ.jpg (https://ibb.co/v1wq1st)



Hi TimVG,

very good results!

regards
ivica

TimVG
12-28-2018, 03:43 AM
Hi TimVG,

very good results!

regards
ivica


Thanks!

Here is my latest result after about 10 hours of measuring and dialing in the system. It is an average of around 8 measurements from both speakers around my exact listening position.
2nd graph has a slightly (1dB) more accentuated downward tilted shelf filter in place.
Please note the small dip in the 1kHz-2kHz region does not show up in the nearfield listening window measurements, so I opted not to correct it.


https://i.ibb.co/nz1dvXR/target.jpg (https://ibb.co/86xvpyX)
https://i.ibb.co/RTh4ytf/target2.jpg (https://ibb.co/nM38C49)

TimVG
01-02-2019, 03:49 AM
Time gated measurements @1m. <700hz omitted for this reason.

Black = on axis
Green = listening window
Red = Early reflections

https://i.ibb.co/xz5X2nr/spinhr.jpg (https://ibb.co/L9pY0Ts)