Page 1 of 9 123 ... LastLast
Results 1 to 15 of 126

Thread: DEEPER BASS: D. B. KEELE/E-V's STEP-DOWN MODE

  1. #1
    Senior Member RMC's Avatar
    Join Date
    Feb 2017
    Location
    Canada
    Posts
    1,626

    DEEPER BASS: D. B. KEELE/E-V's STEP-DOWN MODE



    DEEPER BASS, INTRO

    I've decided to start a new thread with the information I gathered on Keele/E-V's "Step-Down Mode" and to re-group everything here for convenience purposes for those who may be interested or need it. That info regarding a derivative of the B6 LF alignment may well be relevant even with some JBL drivers (e.g. Lorr Kramer and Greg Timbers did model a classic sixth-order alignment with the 2245H woofer in their famous article "New Lows in home-built subwoofers", 1983, available in the library of this Web site).

    Consequently, for better understanding purposes of what will folow, I'm re-posting here to begin with what I previously wrote in two posts in another thread regarding "Step-Down Mode". The latter is sometimes referred to as "approximate B6 alignment " by Electro-voice (THE major user of step-down mode), or "pseudo B6/Quasi B6 (QB6)" by D.B. Keele. Pretty much interchangeable wording.

    For those less familiar with the matter, the B6 LF alignment is a filter-assisted vented-box speaker design. This may be the answer for folks who want deeper bass but just can't afford the space taken by a larger enclosure, or when your life partner turns thumb-down on that...

    To complement the two re-posts, I'm adding a third contribution here now (scroll down) with regards to the temptation/feasibility of using a regular equalizer to do the job.

    Don't forget while reading, Faux = peak-boost frequency of the auxiliary filter needed.

    A fourth contribution is in the making, which may well be 4th and 5 th, considering the first draft already written, plus the amount of documents still on my desk...

    Richard

  2. #2
    Senior Member RMC's Avatar
    Join Date
    Feb 2017
    Location
    Canada
    Posts
    1,626
    DEEPER BASS, PART 1, ELECTRO-VOICE

    Here is how Electro-Voice describes, in simpler terms than D. B. Keele, the "Step-Down" operation (lower tuning than normal + EQ) to get lower bass from a smaller box:

    "... the "step-down" mode, which approximates a B6 Thiele alignment. Step down is a good way to extend system low-frequency response by increasing amplifier power at certain frequencies instead of enclosure size.

    In step down, the enclosure is tuned to a lower-than-normal frequency. This increases system output at the new tuning frequency and reduces output slightly in the region of original tuning. The smoothly falling response which results can be equalized to provide a new system 3-db-down point that is about 0.7 that of the original. To achieve a similar response extention without equalization would require an enclosure at least twice the size.

    In the step-down mode, system input is increased at the lower end of the operating range, by a maximum of 6 db close to the new tuning frequency. Note that this boost does not affect system instantaneous peak output, which is related only to the speaker's linear displacement ability and effective diaphragm area."

    The frequency where an increase is applied (EQ) is called a "peak-boost frequency". According to E-V, other frequencies sufficiently close, within + or - 5%, will provide no significant performance compromise.

    With regards to Subpassband speaker protection (the other side of the coin) here is what E-V says:

    "Below the enclosure tuning frequency, cone excursion increases rapidly. Since acoustic output is also falling rapidly, there is no utility in driving the system with signals much below the tuning frequency. (...). ... high output systems, especially subwoofer systems, should be protected by a high-pass filter with a 3-db-down corner frequency of about 0.8 the enclosure tuning frequency... Below the corner frequency, a rolloff of 12 db per octave is usually sufficient."

    So, an amplitude boost followed by a bandwidth cut, that E-V calls "boost-and-cut equalization". "The equalization has no effect on the large-signal output but does have the effect of separating the small- and large-signal curves except at the very top of the frequency range shown."

    Quoted from: Electro-Voice, Pro Sound Facts, #7, oct. 1984, pages 3 and 5

    See Figure 3 in the XEQ-2 crossover/equalizer data sheet to look at examples of increases applied at various frequencies (+ 6 db boost with a filter Q of 2 and high-pass filter of 12 db/oct.).

    E-V did not mention, in this context, the LF room gain possibility based on speaker placement.

    More interesting stuff to come from E-V and D. B. Keele in subsequent posts.

    Richard

  3. #3
    Senior Member RMC's Avatar
    Join Date
    Feb 2017
    Location
    Canada
    Posts
    1,626
    DEEPER BASS, PART 2, D.B. KEELE

    As indicated in an earlier post, here is the second part of the "Step-Down Mode" information, this time from D. B. Keele(1).

    Keele started with different driver types, modeled them in conventional LF alignments (e.g. B4, Qb3, C4, closed box) and also simulated new pseudo-sixth-order alignments to compare results with those. He adds that all the new alignments have coincident F3, Fb and maximum boost frequencies (p. 354).

    Though Keele talks about a "modest amount of lift" (+ 6 db at the peak boost frequency in the bass range, 1.07 times F3), doing this requires four times the amount of power. What would otherwise be driven with 25 watts now requires 100, and what would otherwise have taken 100 now requires 400 Watts. Not really modest in my view, however today the cost of Watts can be pretty cheap.

    Keele also says "It must be stressed that these alignments represent only discrete selections from a continuum of possible alignments." (p. 355) So there's a lot more than meets the eye (e.g. E-V's).

    Keele says its not driver Fs that's important but Fs divided by Qt which indicates how low a driver will go. Adding for a desired F3 in the 25-50 hz range, Fs/Qt can be from about 80 to 160 hz. Drivers with that ratio would be likely candidates for B6 alignments. BTW for example the Fs/Qt ratio for JBL's 1400 ND stands at 106, therefore within that range.

    Talking about a system aligned for a B4 response (F3= Fb=40hz) converted to a pseudo-B6 alignment (F3=Fb=28hz, boost @ 30hz ) Keele mentions "The net effect on the system response is one-half octave extention of low-frequency response with only about 3 db less maximum acoustic power output capability in the passband." (p.356) Three db is a notable quantity and represents half the power. The lower tuning and EQ boost will return the response back to a roughly flat condition according to Keele. That conversion "... somewhat reduces the maximum acoustic output capabilities of the system in the 35-70 hz range, but greatly increases the maximum output below 35 hz" (p. 357)

    In his concluding summary, Keele mentions using these "... alignments offers real advantages in providing maximum useful low-frequency acoustic output while minimizing driver diaphragm excursion both in and out of the systems operating frequency range." (p. 357).

    In another quasi-sixth-order (QB6) box he simulated (Fb=Faux=F3= 26 hz), the 6 db peak boost from auxiliary circuit happens at 1.07 times Fb = 28 hz (p. 358). (Faux being the frequency of the auxiliary peak boost filter, as I understand).

    Note that in the simulations Keele did, when the new Fb was at 26 hz the 6 db boost was applied at 28 hz, for Fb 28 hz it was applied at 30 hz and for Fb 43 hz it was applied at 46 hz. Not exactly at Fb...

    Also interesting to note that for a 4 cu. ft. box Keele used about 20% overvolume to compensate for box losses, while using 30% overvolume to offset box losses for a 7.5 cu. ft. box. More overvolume for larger boxes and less for smaller ones appears to agree with Leach (2) who suggests to use lower QL number for larger boxes and higher QL number for smaller boxes. That means larger boxes are more lossy than small ones. In the computer design process, entry of a lower QL will lead to a larger volume (Vb), and entry of a higher QL number leads to a smaller box. Unfortunately, Leach did not mention which higher or lower QL to use, but he did mention the initial design process with QL 7 is correct, and QL 7 good for 2 to 3 cu. ft. boxes. I would probably use QL 5 for larger boxes and QL 10 for smaller ones.

    I have some issues with Keele's article and also with some E-V Step-Down Mode items which I'll cover in a future post.

    Richard

    (1) A New Set of Sixth-Order Vented-Box Loudspeaker System Alignments, JAES, 06-1975, P. 354. This can be found on his Website, dbkeele.com, AES papers, item # 8.

    (2) W. Marshall Leach Jr., Vented-Box Loudspeaker Design with a Given Driver, 2001, P. 1, available on Net.

  4. #4
    Senior Member RMC's Avatar
    Join Date
    Feb 2017
    Location
    Canada
    Posts
    1,626
    DEEPER BASS, PART 3, EQ

    STEPPING-DOWN WITH A REGULAR EQUALIZER? (For Didier and Rob in Europe, and other members)

    For cost reasons a number of people will be tempted to use their present equalizer instead of making or purchasing a dedicated boost & cut filter. Hopefully, it will work correctly knowing what to watch for.

    In simple terms, the filter Q number of the 6 db peak boost filter (Q 2) referred to in E-V's litterature for step-down EQ, refers to the shape of its response (symmetrical bell shape). The important parameter of that filter's response is the ratio of its height to its width (called "Q") which determines bandwidth. Low Q means wider shape and high Q means narrower.

    The ideal situation is if you have access to Parametric EQ where you can define bandwidth, frequency and amplitude. Failing that, graphic EQ MAY still work if you have a slider at the right center frequency (or pretty close), among other conditions. 1/3 octave models offer much more center frequencies than 2/3 octave models, but instinctively the more sliders on an EQ the narrower the filters have to be (higher Q). That in order to minimize overlapping/interaction problems with other filters next to the one(s) used.

    For graphic EQ, the next thing you need to know is the applicable filter Q number with regards to the sliders. Many EQ manuals don't elaborate much on this: "Constant Q filters, accuracy 3%" as do my 2/3 oct. Pro models. But I also have an older octave band Hi-Fi EQ and the manual says: "Filter Q 2.5 nominal, + or - 10% ." The manufacturer describes his filters as being medium Q. It has sliders at 30, 60, 120hz, etc. which don't offer a lot of practical center frequencies for Step-Down operation, however the 30 hz slider could be a good prospect for SOME typical Step-down applications.

    A little higher than needed Q number indicating filter width being slightly narrower than perfect fit (Q 2). Such filter will have a little less "reach" on both sides of the underdamped filter's peak boost frequency. Assuming better filter tolerance than 10%, which does happen with audio gear specs being better than stated (E-V requires 5% tolerance here), that would still likely mean: a small dip in LF response above Faux (peak boost frequency of auxiliary filter), plus a little earlier drop in LF response below Faux, therefore a bit less deeper bass. Although not perfect, that case may still be acceptable for some. A slight negative ripple on both sides of that filter would imply a LF response less roughly flat as it could be after EQ is applied.

    A boosting filter with a narrower impact on VLF isn't necessarily a bad idea since it reduces the risk of excessive cone travel if there's no VLF High-Pass Filter on the EQ, E-Xover or Amp. {This is where the issue of the step-down required high-pass filter of 12 db/oct. below Faux is relevant, more so if none is built-in the gear used}. On the negative side, since the boosting filter won't impact as low as ideal, one won't get the full step-down low-bass benefit (but then that should help a little on cone excursion in the absence of the high-pass filter).

    In the event that a high-pass filter is present in the gear used, its corner frequency must be just below Faux action for best results, since above this it kind of defeats the purpose of boost equalization, and if its too low it doesn't help much cone travel/box unloading. Again, when a driver is working at or close to Fb the cone travel beware is less an issue, its more below box tuning region that caution must be used...

    BTW just to make things clear, in the real thing (E-V's boost/cut EQ) the SAME filter does both tasks of VLF reduction/protection AND 6 db peak boost since it is an underdamped (peaky) 12 db/oct. high-pass filter. It "hits two birds with one stone" as they say. In the equalizer thing you need TWO filters to do the job correctly: one for 6 db peak boost (i.e. eq slider) and one for reduction/protection from VLF. Since one of my Pro EQ has a variable (20-200 hz) low-cut (high-pass) filter of 12 db/oct. I may try this EQ thing one day...

    Another post will follow in the near future about some issues I have with Keele/E-V step-down process.

    Richard

  5. #5
    Administrator Robh3606's Avatar
    Join Date
    Apr 2003
    Location
    Rocinante
    Posts
    8,170
    Here is what a BX-63A hooked up to CLIO looks like. Attached are two plots that have the low frequency output measured at 63Hz, 90Hz and 125Hz and a second which has the high pass with the 125Hz curve. Seems relevant as it was EQ supplied for the B380 and 460. There is also an option the JBL 5235 active crossover around 30hz or so very similar.


    Rob
    Attached Images Attached Images    
    "I could be arguing in my spare time"

  6. #6
    Senior Member RMC's Avatar
    Join Date
    Feb 2017
    Location
    Canada
    Posts
    1,626
    Hi Robh,

    Quite Interesting input, and thanks for the attachements. I have not found a detailed data sheet for BX 63/A, other than JBL Pro Service Technical Manual which doesn't mention what I needed to know.

    "Seems relevant as it was EQ supplied for the B380 and 460" You're absolutely right its relevant! JBL's BX63 was in fact used in the Kramer/Timbers article I referred to in post # 1 (introduction), where I mentioned " (e.g. Lorr Kramer and Greg Timbers did model a classic sixth-order alignment with the 2245H woofer ....)"

    I should have expanded more on JBL's own "Step-Down" and the BX 63 in a JBL related Web site, MEA CULPA. My only excuse: I find JBL's "step-down" application somewhat limited VS Keele/E-V's pretty large application.

    As for "There is also an option the JBL 5235 active crossover around 30hz or so very similar." you ARE right on target. BTW Recently I pulled-out my JBL file to review some older info to contribute to "Favorite JBL speakers" Thread (where I voted only for HLA Series! that should put some "ambience" in the 43 series club), and realized I had for a long time in my file (kept at arm's reach just in case it gets noisy) the 5235 data sheet 8/89 I had not seen for years (lucky me!).

    In the product Features on front page, it mentions "Swithchable subsonic high-pass functions", not telling the whole story re capabilities. However, on the back side Specifications in the High pass filtering section, there you get it: 20hz, +6db, filter Q 2; 30hz, +6db, filter Q 2; 40hz, +6db, Filter Q 2. The other filtering options (other than crossover cards) are more for "subsonic" removal. All the filters for EQ boost & subsonic are given at 12db/oct. slope. So, JBL is in the game too.

    I wrote "That info regarding a derivative of the B6 LF alignment may well be relevant even with some JBL drivers" That meant Keele's/E-V's stuff may also be used with some JBL drivers (ref. Keele's Fs/Qt), which opens more doors and applications in my view.

    Since the BX 63 and/or 5235 are limited to certain frequencies (e.g. 20, 30, 40 hz) and Keele says "It must be stressed that these alignments represent only discrete selections from a continuum of possible alignments." then the possibilities become numerous...

    Finally, In the E-V XEQ- 2/3 data sheets one can find the formulas to calculate their own custom filters. I'm not pretending that one manufacturer's stuff is better than the other, but simply that one gives more detailed info to the speaker builder on this subject and provides more flexibility. Best regards,

    Richard

  7. #7
    Senior Member
    Join Date
    Apr 2014
    Location
    Europa Union
    Posts
    130
    In 5235
    It is possible to change the filter tuning frequency but some résistors have to be changed on the main PCB.

    Does someone know the volume and port tuning of a B380, just to understand how this fit to the natural response curve ?

  8. #8
    Senior Member Lee in Montreal's Avatar
    Join Date
    Sep 2010
    Location
    Montréal
    Posts
    2,487
    Kef Kube - Q transformer
    Small box. Deep bass. Beware of over excursion and clipping amp.






  9. #9
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    Hi Richard and others

    Attached is a spice simulation of a circuit l located over on diyaudio.com for the boost- cut equalisation
    It was posted by an engineer. I have not evaluated this myself but l understand is what is referred to as an under - damped high pass filter.

    The effect of varying the Q of the filter at the resonant frequency is the Peak or boost, below this frequency is the high pass filter. This is accomplished with a potentiometer

    The frequency of the filter can be adjusted with another potentiometer

    The practical application of this is straight forward for the diyaudio folks.

    I understand that a short form kit was available from the engineer who submitted the above spice model

    It was a while ago so l will see if this kit is still available?

    This application would be an in out black box pre a power amplifier much like the EV box used for domestic models.

    The Jbl BX 63 had as l recall the filter as described above in addition to a crossover function accompanied with a passive RC time constant to maintain purity of the audio signal for satellite loudspeaker.

    The EV documents do refer to 4x the power and l note Dayton Audio through Parts Express have plate amps with dsp equalisation that can be programmed for boost- cut equalisation. The Dayton product kills two birds with one stone and may be an opportunity for anyone wants to try out the step down.

    Below l found a collection of documents on the subject of Ev step down in production of PA loudspeakers

    There are graphs that illustrate the impact of the step down.
    I draw no conclusions from these links other than a Ev did apply the idea in commercial systems

    In the last link there is a collection of vintage approaches to bass equalisation and if you scroll down the EV domestic step down product. Interesting

    These are only links so open the doc and it’s quite easy to read.

    http://pdf.textfiles.com/manuals/STA...0EV/TL3512.pdf

    https://electrovoice.com/binary/TL44...Data_Sheet.pdf

    https://www.scribd.com/mobile/doc/14...-Speaker-Plans

    http://www.roger-russell.com/equalizers/equalizers.htm
    Attached Images Attached Images  

  10. #10
    Senior Member RMC's Avatar
    Join Date
    Feb 2017
    Location
    Canada
    Posts
    1,626

    A precision

    A clarification note. In part 1, E-V, Post # 2, I must add the following in order to avoid any confusion.

    The paragraph where I quoted E-V concerning "With regards to Subpassband speaker protection...", this should be seen like a "by the way" note on out of LF passband filtering and how E-V suggests to determine such filter's frequency on low-frequency enclosures in general (i.e. 0.8 times Fb or box tuning frequency). This is NOT part of the Step-Down process, because the way the 6 db boost (and cut) filter is designed it already includes subpassband filtering at 12 db/oct., In addition to equalization. No requirement for redundancy then. I should have put that paragraph at the end of my E-V text as a BTW note.

    In a spec sheet's section on subpassband speaker protection, E-V mentions "The step-down equalization described in the Step-Down section provides the required protection."

    However, In part 3, Stepping-down with a regular equalizer, post # 4, the need for a subpassband filter remains relevant...

    Richard

  11. #11
    Administrator Robh3606's Avatar
    Join Date
    Apr 2003
    Location
    Rocinante
    Posts
    8,170
    Does someone know the volume and port tuning of a B380, just to understand how this fit to the natural response curve ?
    Hello

    The box was 4.5 cubic feet and the tuning was 26Hz I have plotted both the response for both the equalized version and the box with no EQ at an input of 150 watts. You can clearly see the effect of the EQ in the displacement curves. They are great for music but looking at the displacement it's no wonder they can run into trouble used as subs in a modern HT system.

    Rob
    Attached Images Attached Images   
    "I could be arguing in my spare time"

  12. #12
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    Hi Rob,

    Do you use your 2235's as B380's or if not would you consider it?

  13. #13
    Senior Member RMC's Avatar
    Join Date
    Feb 2017
    Location
    Canada
    Posts
    1,626
    Hi Robh,

    Thanks for the simulation in post # 11. A few things aren't clear to me.

    What are the boosting filter parameters you applied (frequency, amplitude, and Q or bandwidth) ?
    Any LF subpassband filtering applied ?

    On the bottom graph you have TWO response curves (orange and red). But on the top graph I see THREE cone excursion curves ! (two orange at Fb and one red). The orange curves are typical shape of cone travel in a vented box. However, the red excursion curve, looking like a pair of boobs, isn't typical for a vented box...

    Finally, I notice on the red excursion curve that the right hand bump is at almost 8mm @35 hz, Xmax being 8.5 mm, and the left hand bump, around 20 hz, about 7+ mm, for the same input power but with even lower frequency! Effect of subpassband filtering ?

    Richard

  14. #14
    Administrator Robh3606's Avatar
    Join Date
    Apr 2003
    Location
    Rocinante
    Posts
    8,170
    Hi Rob,

    Do you use your 2235's as B380's or if not would you consider it?
    Hello Ian

    Well actually I do use them as B380's. Remember my old LE-14 subs under my HT mains?? They have been replaced by a pair of B380's. I originally got a pair of 136Hs to run as subs in my HT but every once in a while I would slap a mass ring which is a bit alarming. So I set them up as stereo music subs and got a pair of 2266's and set them up as HT subs. Use the DSP in a crown amp to simulate a BX-63 so same idea in a smaller box higher x max and power handling.

    http://www.audioheritage.org/vbullet...highlight=bx63

    Rob
    "I could be arguing in my spare time"

  15. #15
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    http://www.audioheritage.org/vbullet...l=1#post336627

    Hi Rob,

    Excellent. Your real thorough.

    The WT2 with BassBox Pro is a great combo.

    That must have been after the last time I was at yours.

    Great thread on how to do a diy sub right with some real practical experience.

    Glad you bought up the mass ring slapping with the 2235s. I recall Mr Widget talking about this.
    Had not thought about it but something to consider with any 2235 based systems.

    The amp is a winner with dsp filters and limiter. Nice.

    Small box too.

    PS I still have the 2245s in my 4345’s doing double duty for music and HT in the new music room. Never had to turn on the 4465 yet! Might mess with a crown and EQ per the bx63 for grins.

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Similar Threads

  1. Early 80's Don Keele & Mark Gander / JBL horn plans
    By Mike Caldwell in forum Lansing Product DIY Forum
    Replies: 4
    Last Post: 04-30-2014, 07:18 PM
  2. D.B.Keele.Jr's
    By stephane RAME in forum Lansing Product General Information
    Replies: 0
    Last Post: 04-12-2010, 11:52 AM
  3. Step Response
    By Robh3606 in forum General Audio Discussion
    Replies: 17
    Last Post: 02-09-2009, 05:14 PM
  4. Resurround Step-by-step
    By boputnam in forum Lansing Product DIY Forum
    Replies: 43
    Last Post: 01-21-2008, 09:23 PM
  5. Resurround Step-by-step
    By boputnam in forum Lansing Product Technical Help
    Replies: 31
    Last Post: 12-25-2005, 03:26 PM

Posting Permissions

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