JBL Cabaret 4698B Speaker - BWC alert!

Discussion in 'Speakers' started by purr1n, Jan 30, 2019.

  1. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    47,263
    Dislikes Received:
    74
    Trophy Points:
    113
    Location:
    Antarctica
    Finally, here is a measurement with the microphone shoved in front of the mouth of one of the four huge ports. I wanted to see what frequency the ports were tuned at.

    JBL 4698B port tuned to approximately 44Hz.
    upload_2019-4-28_0-31-53.png
     
  2. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    47,263
    Dislikes Received:
    74
    Trophy Points:
    113
    Location:
    Antarctica
    So one thing that always puzzled me with JBL's approach with the 4698b was using the woofer without a crossover and letting it naturally roll-off. In reality, this almost always means some unevenness and an eventual peak around 1-2kHz (for an 18" woofer) just before roll-off (example shown below). How the heck did JBL pull this off? Since after all, we did not see anything major sticking out in the mids for the overall frequency response seen here: https://www.superbestaudiofriends.o...98b-speaker-bwc-alert.7412/page-2#post-254866

    Here is an example of the Eminence Omega Pro 18" driver's frequency response. I'd hesitate to run this without a crossover because I would be concerned that the 1.2kHz bump would show up in the overall frequency response.

    Eminence Omega Pro 18" FR
    [​IMG]

    Out of curiosity, I decided to take a measurement of the JBL E155, with the microphone shoved up right next to the driver like this to isolate the behavior of the woofer. I could have put the darn speaker into biamp mode to isolate just the woofer, but I was too lazy. There's no crossover in front of the woofer, so this should get me a pretty good idea of what the woofer is doing. The output of midrange and tweeter should be attenuated significantly.
    upload_2019-4-28_11-24-6.png

    JBL 4698b E155 woofer measurement via microphone shoved close. Note peak before rolloff marked in red.
    upload_2019-4-28_11-11-35.png

    As we can see, the E155 isn't different from any other woofer. It doesn't have any magical properties. There a 5db peak at 600db, and the typical associated shallow dip just before that peak at 400Hz. However, there is one very important distinction that makes the E155 work. The peak is much lower in frequency than is usually seen with many other big woofers. The 600Hz does need to be addressed, but how is this done without a network circuit?
     
    Last edited: Apr 28, 2019
    LetMeBeFrank, numbercube and spwath like this.
  3. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    47,263
    Dislikes Received:
    74
    Trophy Points:
    113
    Location:
    Antarctica
    The answer lies in the mids. I took a similar close mic'd measurement of the midrange driver.
    upload_2019-4-28_11-26-34.png

    JBL 4698b. E1110 midrange measurement with microphone shoved close. Note that the woofer output will still be in this plot, but highly attenuated. It would appear that the E110 might have a dip around 700Hz, just where the woofer has a peak. I would be an awesome coincidence huh, where the summed output of the woofer and midrange come out to nearly flat. Perhaps this is why the woofer can be run the way it is without a crossover network.

    JBL 4698b E1110 midrange frequency response via microphone shoved close.
    upload_2019-4-28_11-31-42.png

    To confirm, I found this JBL E110 frequency response graph on the Internet. Yup, it looks like that 700Hz dip is there and inherent to the driver. The E110 midrange, like any other driver in the world, also has issues with rising response and a 2.5-3kHz peak before rolloff, but this is addressed by the crossover circuit. The E110 has a coil for high-pass and a cap for low-pass. Note that we can still see the 2.5-3kHz peak in the ARTA measurement above, but highly attenuated by the coil.

    JBL E110 frequency response.
    upload_2019-4-28_11-34-11.png

    I am pretty sure the E155 and E110 drivers were designed way before the the 4698b speaker was conceived so this was likely a good coincidence. This is the kind of stuff I love, when engineers find elegant solutions, finding ways to make things work with less complexity rather than more complexity.
     
    Last edited: Apr 28, 2019
    Priidik, misformatt, sphinxvc and 7 others like this.
  4. msommers

    msommers High on Epipens

    Friend
    Joined:
    Nov 6, 2016
    Likes Received:
    2,227
    Dislikes Received:
    32
    Trophy Points:
    113
    Location:
    Calgary, Alberta
    Home Page:
    Marv what mic and software did you use for this? I'd be curious to try this at the speaker and listening position as well.
     
  5. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    47,263
    Dislikes Received:
    74
    Trophy Points:
    113
    Location:
    Antarctica
    One more thought: the benefits of not shoving a coil in front of a woofer driver are significant, especially in this case were the nominal impedance of the E155 is 4-ohms. A typical big-ass air core 1.4mH inductor (400Hz low pass for 4-ohms) using 18-gauge wire is going to have a DCR of about 0.65-ohms. Think how much energy loss there is with the inductor being 16% of the load.

    However, there is always a good argument for using crossovers or higher slope crossovers. We get the drivers working in their prime spots and we push their bad behavior down in amplitude. The cost, as explained above, is more "crap" in the way. More power robbing coils and to a lesser extent caps. It's about priorities, balance, and trade-offs. There's no free lunch anywhere. One problem of course is that it is difficult or impossible to measure subjectively observed dynamics.

    (One thing I won't do anymore is work with exotic material drivers that ring like hell. We need steep crossovers. This is really the realm of active networks, amps for each driver, maybe DSP, and wholly another topic.)

    The good news is that for those who want a little bit more accuracy, JBL has the 43* series.

    Here are some photos of the JBL 4345.
    upload_2019-4-28_12-10-20.png
    upload_2019-4-28_12-11-55.png

    The JBL 4345 is a 4-way speaker from the 70s: 18" woofer, 10" low-midrange, midrange-high horn (with fancy lens), and slot supertweeter. More crossover parts, but the approach is still minimalist. The polar response is better too. Overall efficiency is less than the 4698b, but we can argue that the 4698 cheats via the use of a 4-ohm woofer.

    JBL did some cool stuff for the 43* series, like biased capacitors for the networks.

    I'd love the build something like this. Some of the drivers are unobtainable now. However, it might be a neat idea to do a modern interpretation. Maybe I should quit my job and see if Schiit would be willing go in on it (to keep parts cost down).
     
    9suns, Ringingears, fraggler and 8 others like this.
  6. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    47,263
    Dislikes Received:
    74
    Trophy Points:
    113
    Location:
    Antarctica
    Dayton Audio UMM-6 USB microphone. ARTA software.
     
    Boops and msommers like this.
  7. Cspirou

    Cspirou They call me Sparky

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    4,628
    Dislikes Received:
    5
    Trophy Points:
    113
    Location:
    Northwest France
    I've mentioned this a few times but I am not sure if you are familiar with it

    http://www.troelsgravesen.dk/The-Loudspeaker.htm
     
    msommers likes this.
  8. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    47,263
    Dislikes Received:
    74
    Trophy Points:
    113
    Location:
    Antarctica
    I know, but when Troels comes up, I tend to close my eyes and plug my ears and go la la la la. He likes to pick the most expensive drivers possible and then takes a surgical approach to crossover design to make everything work. I'm a cheapskate who likes to identify high value drivers and minimize crossover parts count.

    Troels design needs to be priced at $15k or more to be commercially viable.
     
  9. dmckean44

    dmckean44 In a Sherwood S6040CP relationship

    Friend
    Joined:
    Nov 17, 2016
    Likes Received:
    1,209
    Dislikes Received:
    7
    Trophy Points:
    93
    Location:
    San Diego
    If you thought you could sell a lot of them, you could probably keep costs down by doing the injection molded speaker cabinet thing like with PA speakers.
     
  10. dBel84

    dBel84 Friend

    Friend
    Joined:
    Sep 27, 2015
    Likes Received:
    1,164
    Dislikes Received:
    2
    Trophy Points:
    93

Share This Page