PDA

View Full Version : 4430-4435 Revisited- A short story



Ian Mackenzie
11-12-2005, 06:51 PM
This thread is not a hack work or a diy attempt at re-engineering an already splendid design but a look at what may provide more listening pleasure.

In a sense the 4430-4435 appears a more simple concept, but beneath that simplicity lies a very sophisticated and elaborate design.

Go here for details:

http://www.audioheritage.org/html/profiles/jbl/4430-35.htm

So we are not going to mess around with the design but look at the more subtle and perhaps the not so obvious. As the saying goes, if it it ain't broke don't fix it.

Reading over a few posts by members who have bi amped the 4430-4435 it would seem that while they felt there were some gains they also felt that something was lost.

As discussed here:

http://www.audioheritage.org/vbulletin/showthread.php?t=6542&highlight=4430+crossover


I also found this when biamping the 4344-4345 and the findings can be read elsewhere in the 4343-4344 upgrade thread.

Its a case of not what you are doing but how you do it. The 4430-35 has stood the test of time but with the associated equipment things have moved on a bit.

The intention is to build a purpose built active crossover and amplifier and arrange for a member to trial it for a time and post their observations.

If you are interested post or send me a pm.

Ian

yggdrasil
11-13-2005, 05:34 AM
You are so full of energy.:applaud:

frank23
11-13-2005, 09:52 AM
such a good idea!

but I don't even have the time / energy to build cabinets for the 2235's I have had lying around [freshly reconed] for the past 3 years so I would not be worthy of such an effort to be made

it would be very nice to have a good active compensation setup for the 2344 as the active compensation in the M552 crossover is not good at all, not even for the 238x series for which it was meant

frank

Ian Mackenzie
11-13-2005, 12:51 PM
Frank,

The 4430-35 has the compensation built into the passive crossover.

The compensation remains in the circuit when bi amping.

You are correct that other compensations may not be correct.

If you search the forums there are several threads that discuss the 4430-35 crossover in detail..incl voltage drives.

Actually....I have an idea.

My brother has a spare pair of 2235H's and I know where I might find the 2344 horns. I might build an exact 44430 clone over the holidays however i do not propose that this will become a diy 4430-35 thread.:D

Prior to my current system I spent 18 months testing & using the 2344 horn and it's passive crossover, so I have some idea of it characteristics. The important thing is not to mess with the JBL design.

Ian

Zilch
11-13-2005, 02:58 PM
If we had the M552 crossover schematic, a simple mod to the "CD pre-emphasis" might make it play 2344 nicely.

[Think NL200t3....]

All I have is the block diagram in the manual.... :(

Ian Mackenzie
11-13-2005, 03:31 PM
Perhaps but that is not the intent here.

If one takes the time to read the profile and the White Paper it becomes obvious why they did what they did. No doubt there were other options but the one taken did everything they wanted to do at the time.

I will look at flattening the ripple (as discussed by David Smith) where possible at the lower end on the horns response.

I suspect you might have missed that with your rta .

Robh3606
11-13-2005, 04:03 PM
The M552/553 are tailored for professional use. This includes the CD curves for the horn families. As far as the schematic you just have to know where to look.


http://www.jblproservice.com/pdf/M-Series%20Electronics/M552.pdf

http://www.jblproservice.com/pdf/M-Series%20Electronics/M553.pdf



Rob:D

Zilch
11-13-2005, 05:11 PM
Thanks, Rob.

I'll scoot on over there now.... ;)

HEY! There's that donut power thingy I was talkin' bout.... :p

morbo!
11-13-2005, 05:26 PM
HEY! There's that donut power thingy I was talkin' bout.... :p

lololol:applaud:
:rotfl:
i now will officially call it the donut power thingy from this day forth

Ian Mackenzie
11-13-2005, 10:23 PM
For technical reasons boosting the driver HF response is inappropriate and has undesireable trade offs.

Please read:


"This thread is not a hack work or a diy attempt at re-engineering an already splendid design but a look at what may provide more listening pleasure.

In a sense the 4430-4435 appears a more simple concept, but beneath that simplicity lies a very sophisticated and elaborate design.

So we are not going to mess around with the design but look at the more subtle and perhaps the not so obvious. As the saying goes, if it it ain't broke don't fix it. "

The forums are full of hair brain modifications, I do not propose to involve myself in that pointless pursuit. Besides I think it only adds insult to injury of the original designer..David Smith

I will be working on this below the line from this point.

Ian

Zilch
11-13-2005, 10:37 PM
I also read:

"Its a case of not what you are doing but how you do it. The 4430-35 has stood the test of time but with the associated equipment things have moved on a bit.

The intention is to build a purpose built active crossover and amplifier and arrange for a member to trial it for a time and post their observations."



As one who has tried both active and passive 4430 crossovers per JBL specifications, and also having built N3134 from scratch, there is likely no more stalwart proponent of the original design here than myself.

Frank mentioned M552. I said I'd look, with an idea in mind. If anything comes of it, we'll report it elsewhere.

I'm NOT a subversive.... :p

Ian Mackenzie
11-14-2005, 02:57 AM
I also read:

"Its a case of not what you are doing but how you do it. The 4430-35 has stood the test of time but with the associated equipment things have moved on a bit.

The intention is to build a purpose built active crossover and amplifier and arrange for a member to trial it for a time and post their observations."



As one who has tried both active and passive 4430 crossovers per JBL specifications, and also having built N3134 from scratch, there is likely no more stalwart proponent of the original design here than myself.

Frank mentioned M552. I said I'd look, with an idea in mind. If anything comes of it, we'll report it elsewhere.

I'm NOT a subversive.... :p

You really need to invest in a Clio or what ever over a simple RTA but I guess you know that already. Otherwise its snakes and ladders perpetually.

I saw an advert for a cheap Clio if you are interested.

Earl K
11-14-2005, 06:44 AM
A solution for both was to use a series resonant bypass network.* 1 microfarad in series with about .08 milihenries would resonate around 15K and give about 2 dB more output for the highest frequencies.* It also gave a little less around 5k where the 2425/biradial combo was a little hot.* A variable resistor in series the resonant leg gave a nice “upper treble” adjustment so we now had a pair of controls with really useful control centers and range.


- I inquired about the inclusion of the series .02 mH inductor ages ago ( obviously the value was adjusted after the 7.5 ohm conjugate was added ). I felt the answer I received at the time didn't reflect its' true purpose . The above info vindicates that view .
- Is this UHF "resonance" part of the "magic" that people speak of when referring to this product ?
- This series resonant circuit approach (LCR ) was rarely implemented in other 2 way networks as a "UHF boost". Perhaps these other networks would benefit from the addition of this "magic" bullet.

<> EarlK

Robh3606
11-14-2005, 08:43 AM
"Is this UHF "resonance" part of the "magic" that people speak of when referring to this product ?"

I a word no. At least not for me anyway. I have run the set-up using an outboard EQ to bring up the last octave and the results sounded the same. What makes the "magic" is how unhorn like the 2344 sounds. With a 1" dome up top you can get a very nice balance off axis. With most other horns the balance changes with the DI index. Listening to the 4344's you can easilly hear when you are out of the window. The XPL 200's the 2344 are more closely matched in this respect and the 2344 is actually better. That said if you want the most clarity and detail from 10K and up the XPL's and 4344 with the 077/2405 wins out. There may be some ripple lower in the passband, which honestly never bothered me, but the imaging and overall smoothness in the room more than makes up for any issues. IMHO

Rob:)

BooBoo Magoo
11-14-2005, 09:55 AM
4430 and 4435 are sooooo last century! :applaud:

spkrman57
11-14-2005, 11:01 AM
4430 and 4435 are sooooo last century! :applaud:


Soon I will get to try the 4430's in my own home and with several different amps to try on it. I expect my McIntosh MC-240 to be the best match with the 4430's as far as what I like to hear from a system.

Ron

Titanium Dome
11-14-2005, 11:29 AM
4430 and 4435 are sooooo last century!


Most of my life occurred in the last century!

Ron

So did mine, so far, but I hope to get at least 40-50 years in this century.

Either way, our age is beginning to show. Same with the old speakers: their age is showing.

BooBoo Magoo
11-14-2005, 11:32 AM
Did you win a pair of XPL200's on eBay yesterday?

Titanium Dome
11-14-2005, 11:42 AM
Did you win a pair of XPL200's on eBay yesterday?

Nope. I was gorging myself at a Hawai'ian brunch with my sweetie, then we went out by the ocean to take some pictures of her in her new, pretty blue Island dress. The auction ended while I doted on her.

Some bloke beat me out by $10 at the last minute, but the memories of my time with her will outlast the temporary loss of the speakers. (The food, however, has already left the building.)

spkrman57
11-14-2005, 11:59 AM
Either way, our age is beginning to show. Same with the old speakers: their age is showing.[/QUOTE]


Todays music is more electronic black box than some of the oldies we grew up with.

So if the speakers are old, and my music is old, and my ears are also older,
then maybe that is why I like it that way.

Most of my music I prefer on a nicer system has natural instraments and vocalists who sing so that I can understand what they are singing!

Ron

Zilch
11-14-2005, 01:17 PM
Well, I see neither 0.02 mH nor 0.08 mH. It's the 0.04 mH L1, under discussion, no?

Lancer
11-14-2005, 01:34 PM
- I inquired about the inclusion of the series .02 mH inductor ages ago ( obviously the value was adjusted after the 7.5 ohm conjugate was added ). I felt the answer I received at the time didn't reflect its' true purpose . The above info vindicates that view .
- Is this UHF "resonance" part of the "magic" that people speak of when referring to this product ?
- This series resonant circuit approach (LCR ) was rarely implemented in other 2 way networks as a "UHF boost". Perhaps these other networks would benefit from the addition of this "magic" bullet.

<> EarlKGiskard beat this to death. All the pertinent schematics, both passive and active were posted. All the voltage drives showing what was going on were posted. The only thing missing was the raw curve(s) from the EDS. The EDS can't be found. Maybe David kept a copy. :dont-know At this point it doesn't really matter because if you're going to go about designing a new filter you're going to run all the curves you need anyway.

Personally I'd forget about the L-Pads, start with a simple 6 dB/octave high pass to counteract the natural 6 dB/octave low pass of the Horn/Driver combo, charge couple the whole mess and use fixed attenuators. That would be where I'd start. Might work, might not.

Or, simply charge couple the 4430 network and move on.

Earl K
11-14-2005, 01:59 PM
Hi,

Well, I see neither 0.02 mH nor 0.08 mH. It's the 0.04 mH L1, under discussion, no?

Ron, You're right . It's the 0.04 mH that's in question . That's what I get for posting before my third cup .


Giskard beat this to death. ,,,,

- You're right , this topic has been beaten to death / though I don't believe it was ever articulated ( until the appearance of that posted article ) that a couple extra db of gain ( in the UHF area ) was manufactured by the implementation of a "resonating circuit" . Perhaps I missed that statement of fact, somewhere along the way .

:p

Zilch
11-14-2005, 03:27 PM
You're right , this topic has been beaten to death / though I don't believe it was ever articulated ( until the appearance of that posted article ) that a couple extra db of gain ( in the UHF area ) was manufactured by the implementation of a "resonanting circuit" . See also N200B and 311xA HF, where:


"It should be understood that the HF boost actually delivers more power to the compression diriver at the high frequencies. Care should be exercised in high power applications to see that the thermal limits of the driver are not exceeded."

[That's not to imply I understood it, 'cause I didn't.... :p ]

Ian Mackenzie
11-14-2005, 03:32 PM
Earl...others,

Aside from measuring it after making it, it stands out like dogs balls if you look at the Q of the filter. Gain is a product of all bandpass filters depending on the Q.

I recall mentioning the gain of this filter in project May, it was like talking to cold fish...John West rejects!

Now back to the issue and the facts :

Below are some curves if the response and impediance of the 2344 and the driver.

Note the bumps in the impedience (blue curve) and the corresponding ripple in the response.

Its embedded into the horn (2 db of ripple) , and the crossover adds some ripple (1 db) as mentioned by David Smith.

The issue is the ripple and the associated impediance bumps. When I built up 3134 crossover a couple of years ago I noticed the ripple but did not know what caused. I always felt the midrange could be improved in anotherwise excellent horn system.

David hints at a potential solution which I am looking at and this is the reason I started the thread. As Lancer says everything has already been posted about the crossover previously and gum slapping over that is a waste if bandwidth!

The quick and dirty way would be to eq it with a graphic and introduce its own inherent issues but I there is perhaps a more elegant solution.

When I have look at the practicality of the solution and post some details.

Ian

Ian Mackenzie
11-14-2005, 03:54 PM
I am looking a preliminary investigation of viable options.

1. A current source amplifier filter
2. Typical voltage source filter correction
3. Feedback error correction

Ian Mackenzie
11-14-2005, 03:55 PM
It might take a little time to investigate this properly however as a concept it could be quite promising.

Ian

Ian Mackenzie
12-29-2005, 03:01 AM
I will be having a look at this over my break and I would like to see the SPAM and Thread Jacking posts deleted unless those who lust so much for SPAM and Thread Jacking insist on them remaining. In the case of the latter a new thread will be posted at the appropriate time with a hyperlink to the project proper at an alternate ISP.

If Moderator can perform this function it would prove useful and be appreciated by myself and to those who putting the effort into this project.

I will also ask that we are looking for contributions of direct relevence to the design of smoothing the lower part of the horn response with means other than liberal use of a RTA/Graphic or parametric eq.

Someone who is a dab hand at passive contour filter design and who has sufficiently accurate measurement equipment and design software (spice / Soundeasy/ Calsod) might like to look at modifying the response by these means or as hinted earlier by considering a negative impedance of an amplifier.

Earl K
05-18-2006, 11:36 AM
This thread is not a hack work or a diy attempt at re-engineering an already splendid design but a look at what may provide more listening pleasure.

- I'll submit that a modiified N3100mkII network is the best candidate to fill these requirements ( see my comments over in Bernard Wolfs thread on 275nd / 3100 compatibility ). ( I'll include a link, later )
- This mkII topolgy is what I'll use if I ever combine my surplus le14h woofs with my 2426/2344 horns.
- The network also holds good promise in getting the 2344 combo to perform down at 800 hz. ( I was able to achieve a minus 5 db down point at 800 hz without much tinkering of my main LC elements ) .


I will also ask that we are looking for contributions of direct relevence to the design of smoothing the lower part of the horn response with means other than liberal use of a RTA/Graphic or parametric eq.
- That's exactly what I did by reassigning 2 out of 3 LCR notch-filters that are in the mkII network . ( The third had to be assigned to keep a 9600 hz peak in check due to the extra UHF response made available through the mkII network ).
- The network I built was optimized for 15db of gain reduction through the midband. This measured out acoustically as well as electrically . This means that right now it's optimized for a 95 db matchup ( as in a 4435 or a ME150H or a 2234 single ). Matching the 4430 will require a rework of the lpad / that will therefore entail a reworking of all other LC values ( exluding the notch filters ) .


Someone who is a dab hand at passive contour filter design and who has sufficiently accurate measurement equipment and design software (spice / Soundeasy/ Calsod) might like to look at modifying the response by these means or as hinted earlier by considering a negative impedance of an amplifier.

- My contouring was accomplished without referring to Widgets' ( above ) CLIO study of the driver/horn combo. The peaks were clearly identifiable as being in need of attenuation when using a 2.5 db per line resolution on an old DSP8000 ( Behringer RTA ) . I happened on this thread a few days back while conducting a search for an old quote of mine . ( Still can't find all my own postings ;) )

- From my perspective ; an adapted N3100mkII network should be considered near the top of the list ( if not the top ) when one is using that driver/horn combo .

- I'm leaving this study as a work in progress / therefore / nothing is completely optimized for anyone else usage .


I will be having a look at this over my break and I would like to see the SPAM and Thread Jacking posts deleted unless those who lust so much for SPAM and Thread Jacking insist on them remaining.

I imagine all things considered, that this posting of mine may initiate more spamming. Sorry about that ! ;)

:)

Zilch
05-18-2006, 01:33 PM
It's interesting to trace the evolution of JBL crossover design.

I think of that newer cascaded notch filter topology as a passive implementation of multiple parametric EQ cuts.

NL200t3 provides the basic 6 dB/octave CD compensation Lancer mentions above.

NL3100 "shapes" it with one notch, and MKII reworks it with three.

[A gross over-simplification, of course.]

Another one I studied was SRX712 for 2431H. 'Course that was before somebody screwed up Page 3 of the pdf and chopped off the four-notch compensation cascade at the top:

http://www.jblproservice.com/pdf/SRX700%20Series/SRX712M.pdf

Here's a scan from a printed copy of Rev.B version 004 I worked with. They're now apparently on version 006, so they've been tweaking it somewhat themselves. ;)

[These are Bo's newest monitors, as I recall....]

Ian Mackenzie
05-18-2006, 03:16 PM
Wow,

Excellent stuff

Ian

Zilch
05-20-2006, 11:47 AM
-This mkII topolgy is what I'll use if I ever combine my surplus le14h woofs with my 2426/2344 horns.

- The network also holds good promise in getting the 2344 combo to perform down at 800 hz. ( I was able to achieve a minus 5 db down point at 800 hz without much tinkering of my main LC elements ).

-Matching the 4430 will require a rework of the lpad / that will therefore entail a reworking of all other LC values ( exluding the notch filters ).

- From my perspective ; an adapted N3100mkII network should be considered near the top of the list ( if not the top ) when one is using that driver/horn combo .

You spec it ... I'll build it.... :thmbsup:

4313B
05-20-2006, 12:21 PM
- The network also holds good promise in getting the 2344 combo to perform down at 800 hz.The 2425/2344 is already stressed at 1 kHz with the 2235H so you want to go higher not lower.
1.2 kHz with an LE14H-1 or LE14H-3 and perhaps 1.35 kHz with a 1200FE, 2206H or 2204H.
Like Greg has stated before, the 1200FE is really nice.

Earl K
05-20-2006, 01:57 PM
The 2425/2344 is already stressed at 1 kHz with the 2235H so you want to go higher not lower. 1.2 kHz with an LE14H-1 or LE14H-3 and perhaps 1.35 kHz with a 1200FE, 2206H or 2204H.

- Yes, you're right. I've even seen fairly loose "wave-cresting" of moderate level Pink Noise in that FR area / on that horn/driver/xover combo.
- I'll hope that those people who want to make a 4430 knockoff will follow your advice. I suspect they may sometimes play things at a louder level than I do. And that'll really initiate a lot of unnecessary distortion within that specific FR area.
- Maybe Zilch would do up a CLIO study and show us the distortion ( using the lower Xover point )

( Now ; )
- The actual likelyhood of the described project ( being realized ) is pretty remote in the big scheme of things / though / if I managed to get to that point and I didn't like what I heard ( ie; distortion caused by the lower xover point ) / then I'd abandon the horn/driver combo and shelve the project ( till I could get a bigger horn or longer throat adapter built to accomodate a larger diaphragmed driver - which is my actual preference ) .


Like Greg has stated before, the 1200FE is really nice.

- Yes, I suspect this would be yet another aquaplased woofer, that I would very much like .



:)

Ian Mackenzie
05-20-2006, 07:28 PM
Hi Earl,

Unfortunately I no longer have my 2344's and they are now with Cyclotronguy so I can't induge myself.

Kent (Cyclotronguy) may wish look at your work / study and pursue it. Other than that Rob might be interested in this alternative arrangment.

regards

Ian

spkrman57
01-07-2007, 04:00 PM
Maybe 1.2khz to 1.6khz crossover possibilities.

Problematic to me would be 2226 efficiency might be more than 2426/2344A could keep up with.

I just got a pair of 2344A's and will try in near future, after my other projects get further ahead first.

Ron





The 2425/2344 is already stressed at 1 kHz with the 2235H so you want to go higher not lower.
1.2 kHz with an LE14H-1 or LE14H-3 and perhaps 1.35 kHz with a 1200FE, 2206H or 2204H.
Like Greg has stated before, the 1200FE is really nice.

Robh3606
01-07-2007, 06:18 PM
Problematic to me would be 2226 efficiency might be more than 2426/2344A could keep up with.

I just got a pair of 2344A's and will try in near future, after my other projects get further ahead first.

Ron

Hello Ron

The only issue you may have is the energy in the last octave. Figure the 96 Db sensitivity of the 4435 is real close to the limit where you can attenuate the midband response and still have useful output from 10-20K. As long as you don't go to far beyond that you should OK. The 2425 can easily keep up SPL wise.

Rob:)

spkrman57
01-08-2007, 08:07 AM
Also, my version will be 16 ohm as I have already the 2226J and 2426J drivers to use.

Just need to build the cabinet and crossover for it!

In a few weeks I should have some progress to report on this adventure!;)

Ron

Ian Mackenzie
01-12-2009, 02:19 PM
The 2425/2344 is already stressed at 1 kHz with the 2235H so you want to go higher not lower.
1.2 kHz with an LE14H-1 or LE14H-3 and perhaps 1.35 kHz with a 1200FE, 2206H or 2204H.
Like Greg has stated before, the 1200FE is really nice.

Zactly.

Pity about the warranty though.

One of the things you need to be mindful of is that the waters get muddied with the issue of commercial agenda's where a member pops up ocassionally with a bright idea and tells you everything you want to hear and not want you need to know. ie distortion measurements and what a manufacturer recommends as a minimum crossover point.

A lot of vendors are selling Ribbons this way. The nature of the distortion at the frequencies involved is such that they think it sounds good until compared to a quality dome tweeter( ref Zaph Audio measurements)

So when I have cleared some current projects (in a few months) I will run some measurements to invesigate what is more important..distortion or damping resonances or a smooth response or diaphragm excurcision. Throw into the mix the exponential reactance.. ....are they all related?

It might be easier just to use another horn and a lot a people go that route.

On the face of it moving the crossover point upwards is the easy way out and also helps directivity at the lower end of the horn. Perhaps some drivers perform better than others? It would be useful to explore this.

I also propose to attempt running some compression drivers including a 2426, Tad 2001 a driver used by Earl geddes the from a current source amplifer to iron out some of those links.

This will dove tail with the two way project I have going elsewhere which is at box stage at the moment.

I always believe diy should be fun and not about expensive jewellery.

If you ask a manufacturer how many units he sells with the most expensive parts (jewellery) he will tell you very few but the illusion and hype attracts the customers and it requires the least effort to line his pockets.

Diy is not an exclusive club where you have to be up sold to be a member. So leave the jewellery to the toffey nosed infantile egomaniacs.

The reality is besides the law of diminishing returns a properly engineered product is never the most expensive model on the shelf but requires the most effort to get to market. Its no surprise it pays the wages. The same applies to diy in that performance /pleasure does not come from wild expenditure but something that has been carefully thought out.

4313B
01-12-2009, 02:26 PM
Pity about the warranty though.It's absurd.

Evidently it makes perfect sense to some. "The five year warranty only covers complete systems. We are no longer in the component business."
And evidently it doesn't make any sense to others. "Further evidence of the downward spiral."

stephane RAME
01-13-2009, 11:33 AM
AES - volume 31 -June 1983

http://www.pispeakers.com/AES_v31_n6_p408.pdf

Ian Mackenzie
01-14-2009, 04:45 AM
It's absurd.

Evidently it makes perfect sense to some. "The five year warranty only covers complete systems. We are no longer in the component business."
And evidently it doesn't make any sense to others. "Further evidence of the downward spiral."

I still think the LE14-4 or the 1200FE make a beautiful sound even on a bad day.

There are still a few 2344s kicking around in cyber space.

I have lots of plans and not enough time.

I guess if I bought a pair and you ran them for 90 days they would have to be good!!

Let me think about it.

Ian:)

This week has been difficult, too many 11 hour days so if I appear intolerant forgive me

Ian Mackenzie
01-14-2009, 04:48 AM
Hi Stephanie,

Yes the whoe directivity thing versus loading the driver has caused a degree of insanity for those who like to keep horns compact.

The K.B.Keele paper is also very interesting (CE)

Ian

stephane RAME
01-14-2009, 10:27 AM
Hi Ian,
I had a pair of 4430 for 14 years with me, very good model, a little hard to drive, very good memories.
Best regards.
Stéphane :applaud:

kilgoretrout
01-15-2009, 02:25 PM
lololol:applaud:
:rotfl:
i now will officially call it the donut power thingy from this day forth

Anyone who can combine Towlie and Morbo into one id is a hero in my book. Maybe I should just get a little high.