Page 16 of 29 FirstFirst ... 6141516171826 ... LastLast
Results 226 to 240 of 427

Thread: Quick & Dirty 4430-Inspired Two-Ways Part II

  1. #226
    RIP 2011 Zilch's Avatar
    Join Date
    Jun 2004
    Location
    Berkeley, CA
    Posts
    9,963
    O.K., first try with 275nd diaphragms on 2426H motors, here with 2342 horns, LE14A woofers.

    It does not appear that these will work as drop-in replacements in most systems without crossover modifications. I tried various horn/crossover combinations to get this with AM6212 HF and NL200t3 LF for a listen without EQ. They still need help around 5 kHz.

    Compare impedance curve on 2342 to Giskard's run with standard 2425H on 2344A here:

    http://audioheritage.org/vbulletin/s...age=27&p=53994

    I'm not seeing anything anomalous.

    Mellow and smooth, aquaplas coating on this thinner titanium diaphragm mitigates the characteristic harshness, even with full HF extension.

    Also tried them on 2344A with 4430's N3134 crossover (bottom). Can't dial them flat. I don't think it's an alignment issue, since both drivers behave similarly.

    More once we complete the 1" driver/diaphragm study here....
    Attached Images Attached Images     

  2. #227
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    Try doing an impedance sweep or check your fixed resister crossover values.

    There way too much attenuation at 4 khz but the other section seems to work okay.

    What was the last diaphragm impedance you used ..16 ohms?

  3. #228
    RIP 2011 Zilch's Avatar
    Join Date
    Jun 2004
    Location
    Berkeley, CA
    Posts
    9,963
    Thanks, Ian.

    I did the impedance sweep and added it to the post above.

    Also did an audio sweep to be sure there was no buzz/rub from installer error.

    I don't have a 16-Ohm version of the crossovers I'm using here, so it's not that.

    It looks like the 2426H driver characteristics are sufficiently altered with installation of D8R275ND as to require different compensation, is all.

    In a former life, I'd be dismayed.

    CLIO will tell us what to do....

  4. #229
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    Odd,

    Okay I get it.

    I found the curves below on my HD for comparison.

    Your orange impedance curve looks almost identifical to the 2426 + Dr82425.

    Maybe the rear chamber of the 275nd is larger...or what ever?
    Attached Images Attached Images  

  5. #230
    RIP 2011 Zilch's Avatar
    Join Date
    Jun 2004
    Location
    Berkeley, CA
    Posts
    9,963
    Quote Originally Posted by Ian Mackenzie
    Maybe the rear chamber of the 275nd is larger...or what ever?
    I don't know. I'm not certain the dome radius is the same, even, though I have no reason suspect otherwise. The phase plug may be different, as well:

    http://www.lansingheritage.org/image...500/page09.jpg

    Aquaplassed diaphragms seem to be a "Now you see it, now you don't" phenomenon. I notice that JBL's using it on the 3" diaphragm 435AL-1 (2431H type) driver in Project Array.

    Are there any stock 1-3/4" aquaplas-dusted diaphragms other than 275nd to try in 2425/6?

  6. #231
    RIP 2011 Zilch's Avatar
    Join Date
    Jun 2004
    Location
    Berkeley, CA
    Posts
    9,963

    Big hornz RULE!!

    [At least for today.... ]

    N3100 crossovers constructed to run the Mini-Everests, completed yesterday. Anybody know (or guess) the function of the initial LCR filter with the fuse in the HF? Why?

    I modeled it in Spice, and it does nothing that I can discern:

    http://manuals.harman.com/JBL/HOM/Te...S3100%20ts.pdf

    Here they are running the Everest horns with 2426H and B380's for LF. Getting the aiming correct for the "locked image" effect is more critical and difficult with these big horns than H3100. I'm not sure I've gotten it right yet. RTA shows rolled off high end, but it's not apparent in the way they sound: clean, crisp, bright with solid midrange. Crossed at 750 Hz.

    H3100's produce a similar curve with a bit more HF SPL (~2.5dB) and steeper rolloff (second RTA pic). I'd pad 'em down except they sound just fine as they are.
    Attached Images Attached Images      
    Last edited by Zilch; 11-28-2005 at 11:00 AM. Reason: Rewrote for better illustration....

  7. #232
    RIP 2011 Zilch's Avatar
    Join Date
    Jun 2004
    Location
    Berkeley, CA
    Posts
    9,963

    Big hornz RULE!! (continued):

    Similarly configured 2352 horns with factory refurbished 2435HPL drivers. We still don't know if the combination is "correct" from a design perspective (Optimized Aperture horn playin' later two-stage phase plug design neodymium drivers), but they certainly measure and sound good together. Crossover derived from AM6212/00 provides about the flattest HF curve I've ever achieved without AutoEQ here....

    CLIO is in the mail, folks, thanks to a lead from forum member Ian Mackenzie....
    Attached Images Attached Images   

  8. #233
    Senior Member grumpy's Avatar
    Join Date
    Oct 2003
    Location
    SoCal
    Posts
    5,742
    Anybody know (or guess) the function of the initial LCR filter with the fuse in the HF? Why?

    I modeled it in Spice, and it does nothing that I can discern:
    Run Spice looking at the network impedance instead of the output voltage. Should see
    a difference ~2.3KHz (lower) unless I dropped a digit somewhere.

    -grumpy

  9. #234
    RIP 2011 Zilch's Avatar
    Join Date
    Jun 2004
    Location
    Berkeley, CA
    Posts
    9,963
    Quote Originally Posted by grumpy
    Run Spice looking at the network impedance instead of the output voltage. Should see a difference ~2.3KHz (lower) unless I dropped a digit somewhere.
    The impedance, huh? I've never had occasion to run a network on WT2. This could be the opportunity.

    [Gotta figure out how to make Spice show impedance, too.... ]

    EDIT: Here's the WT2 impedance (red) and phase plot of HF network, driver, and horn, without that filter.

    ACK! (I think ). I'll scrounge up the parts to see the difference.
    Attached Images Attached Images  

  10. #235
    RIP 2011 Zilch's Avatar
    Join Date
    Jun 2004
    Location
    Berkeley, CA
    Posts
    9,963

    As promised: "A Tale of Seven Drivers"

    On 2342 (4425) horns using NL200t3 crossovers with LE14H-3 in Citation 7.4 LF:

    1) 2407H - Yikes! Oh, yeah. It's a TWEETER, silly! (-5 dB)

    2) 2416H-1 - Used in I don't know what, I just HAVE 'em.

    3) 2418H-1 - JBL's cheapest compression driver.

    4) 2426H - Ummm, y'all know this one.

    5) 2426H/275nd - With aquaplassed diaphragm.
    Attached Images Attached Images      

  11. #236
    RIP 2011 Zilch's Avatar
    Join Date
    Jun 2004
    Location
    Berkeley, CA
    Posts
    9,963
    Last two using the "-16" version of the crossover HF:

    6) 2426J - Same deal, plenty available.

    7) LE85 - With throat adapter.

    Which driver would YOU be playing with 'em 2342's right now?

    "This ain't no party,
    This ain't no disco,
    This ain't no foolin' around...."




    [Note: 2342 horns still available new from JBL Pro Parts. ]
    Attached Images Attached Images   

  12. #237
    Senior Member
    Join Date
    Apr 2003
    Location
    GTA, Ont.
    Posts
    5,109

    LCR(s) & fuses

    Hi Zilch

    I always enjoy this thread a lot / thanks for keeping it going

    Anybody know (or guess) the function of the initial LCR filter with the fuse in the HF? Why?
    - I'll hazard a guess that the fuse is present because the circuit designer was concerned about the safety of the overall circuit ( & source amp ) due to relatively high voltages developing across both the inductor & capacitor .

    - These developed voltages are the product of the LCR "Q" multiplied by the applied voltage . ( Perhaps he thought the capacitor might create a dead short and in turn present this short as a parallel load to the driving amplifier . )


    - The "Q" of the series LCR circuit will determine how much voltage is developed . For instance; if the applied voltage measures 2 volts across the inline resistor ( of the LCR ) and the resonant circuit has a "Q" of 40 , then the measured voltage across either the capacitor or the inductor should be 80 volts . Conversely, one can get an idea of the "Q" of that "series resonant trap" by measuring the voltage across that same resistor and then measuring the voltage across either the inductor or capacitor . Dividing E(resistor) into E(inductor) should equate to the "Q" of the resonant circuit .

    BTW:
    - This info is taken straight out of "Electronics Math" by Alan Andrews , a Sams Publication ( # 21152 ),,,pages 168 - 170/ I may have relayed some of this incorrectly since I haven't yet done any hands-on bench testing / to correlate real-world results to the theory .

    - I've included some LCR info for you to ponder ( the basis for the chart is Giskards' info lifted from the May project ) . You can see by the determined resonance values that both the N3100 and the N9800 ( there are some other examples ) each use a couple of series LCR(s) which are electrically located within octave fractions of each other. In both circuits they have some isolation from each other provided by the inline ( padding or series ) resistor ( 7.5 ohms & 16.5 ohms respectively for the N3100 & the N9800 ) .

    EDIT: The following statement would seem to be contrary to the additional information provided in the following posts
    - The speculation is that the higher "Q" LCR is always positioned closer to the source with the lower "Q" LCR(s) placed closer to the load ( when both occur in the same circuit ). I don't know enough about the pratical implementation of these "resonant traps" to take it much further / though I do use some kludged together "resonant traps" with my biamped Altec 288s .

    Cheers
    Attached Images Attached Images   

  13. #238
    RIP 2011 Zilch's Avatar
    Join Date
    Jun 2004
    Location
    Berkeley, CA
    Posts
    9,963
    Thanks Grumpy and Earl. So, we apparently have a notch filter @ 2.3 kHz. Being a resonant circuit, if the driver were to become disconnected, or some other failure occur, the impedance could drop to zero with the oscillation. Thus, we have doubled shunt resistors and a fuse to protect the amplifier against such an occurance. The fuse would blow with a sustained 22.5 watt dump to common.

    Note that S2600 uses a similar notch, with 20 uF instead of 30 uF. S3100MKII doesn't use it.

    I do see a little peak at 2.3 kHz in the RTA curve for H3100 above, which was being driven without the notch filter. I also ran a painfully long 500-point impedance curve on the driver and horn from 500 Hz to 10 kHz. If there's a major resonance peak at that frequency, it's not apparent. Perhaps there was more of a problem before they started coating the rear of the horns with Aquaplas:
    Attached Images Attached Images  

  14. #239
    Junior Member
    Join Date
    Feb 2005
    Location
    Toronto Canada
    Posts
    3

    S 3100 network

    I was looking at the 3100 network and found the front string of components a little curious. Hanging at the front of the network it can't effect the system response as long as the driving impedance is low (any modern amp and reasonable lead resistance.) Modeling the network will show NO difference to the volts at the driver terminals unless a few ohms are modeled into the amplifier source Z. The clue here is that the resistor (R5 parallel to R6) is 10 ohms. This makes it less a notch and more of a conjugate circuit as a conjugate generally has a resistor equal to the impedance value you are "flattening out to". I suspect that the overal system impedance curve went very high at 2300Hz (calculated center frequency) and the designer wanted to flatten it out some. The fuse doesn't protect the driver, if it blows the conjugate goes away but the driver still sees the rest of the network. Also note that, unless biamped, this circuit is at the input to the whole system and works to flatten the whole system impedance.

    A second thing to note is the notch filter at the back side of the circuit. (C9 L2) It is also tuned to about the same frequency (2500 Hz) and WILL put a dip into the driver response. What they don't tell you is that the inductor probably has a significant resistance built into it. Otherwise it would be very narrow in effect and damaging to the impedance curve. I would model it with a range of resistances (say 1 to 4 ohms) and see what the effect is on the response curve.

    Finally, the network is basically second order. The R3 C5 combination is an EQ feature common to all the constant directivity horn networks. R3 sets the midrange level and the C5 bypass lets the highs around it. Raising or lowering R3 would lower or raise midrange level. Reducing C5 would raise the bypass frequency and reduce the lower treble.

    Regards,
    David

  15. #240
    Member DRG's Avatar
    Join Date
    Sep 2005
    Posts
    37
    Thanks David!

    Also note that, unless biamped, this circuit is at the input to the whole system and works to flatten the whole system impedance.

    Thank you for explaining this. It seems G.T. has been doing it alot lately in his filters. "Lately" is relative, I think it's been close to 15 years now. :dont-know

    Here's a very recent example. Notice the lack of notch filters compared to other recent JBL designs.
    Attached Images Attached Images  
    People who make sound their business depend on JBL for their sound.

Thread Information

Users Browsing this Thread

There are currently 2 users browsing this thread. (0 members and 2 guests)

Similar Threads

  1. Quick & Dirty 4430-Inspired Two-Ways Part I
    By Zilch in forum Lansing Product DIY Forum
    Replies: 449
    Last Post: 03-05-2006, 05:41 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •