Page 1 of 2 12 LastLast
Results 1 to 15 of 16

Thread: Active filters for S9900

  1. #1
    Junior Member
    Join Date
    Aug 2022
    Location
    Florida
    Posts
    3

    Active filters for S9900

    New member here. Has anyone had success designing an active crossover for the S9900? I remember hearing that Mr. Timbers uses them with an active configuration himself. I’d like to try it out.

    Thanks!

    J

  2. #2
    Administrator Robh3606's Avatar
    Join Date
    Apr 2003
    Location
    Rocinante
    Posts
    8,170
    Quote Originally Posted by jabbejokker View Post
    New member here. Has anyone had success designing an active crossover for the S9900? I remember hearing that Mr. Timbers uses them with an active configuration himself. I’d like to try it out.

    Thanks!

    J
    Not that I am aware of. The voltage drives are posted so you can use them as a starting point. I go through the compression driver EQ in my Passive Monitor thread. That could help you understand what the notch filters are doing.


    https://www.audioheritage.org/vbulle...24086-K2-s9900


    https://www.audioheritage.org/vbulle...assive+monitor




    Rob
    "I could be arguing in my spare time"

  3. #3
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942

    Question

    While the S9900 can be bi amped Harman never really supported or promoted bi amp of any of these statement systems.

    Greg Timbers is an advocate of bi amping systems he was involved in developing at JBL. Greg told me he had a hard time getting Harman to provide any support information for bi amping. I think part of that is Harman not wanting to service customer enquires on setting up bi amp systems and Harman not wanting to rock the boat in the hifi consumer market where consumer amplifier manufacturers could react adversely to a push for bi amping passive loudspeakers.

    Manufacturers like Elac and Dynaudio have walked a careful line with full active loudspeaker which have everything built into the loudspeaker. This is more of a convenience / connectivity benefit to consumers. Harman have a toe in the water with the 4305P self powered monitor.

    The other jbl active monitors are a different market segment from Harman’s perspective seen as broadcast, home studio and pro recording.

    The entry into full active consumer systems has been a learning curve for both manufacturers and the consumer. Some manufacturers are pursuing analogue active active systems while others like Dynaudio are streamlining their full active systems with more refined dsp engineering.

    From my own perspective if your living in a condo and want a compact lifestyle music system an active loudspeaker is the way to go. A small active system punches far above its size and you don’t need to accommodate a whole bunch on hifi equipment.

    For the traditional hifi consumer (a dying breed according to some) it’s sacrilege to loose control of what power amplifier drives the loudspeaker. More and more hi quality integrated amplifiers with digital connectivity are entering the market. This is creating preferences for an amplifier for modern times. The hifi customer gets to keep his/her trusted loudspeakers.

    What some might be referred to as the audio amateur segment which is below the line in terms of marketing visibility continues to explore bi amp and full active amplification of pre determined loudspeakers and project loudspeakers. This group have been coined the thinking cap customer by Nelson Pass who has nurtured diy audio for decades.

    The barrier to a performance improvement over the level of a pre determined passive loudspeaker is as the OP points out an understanding of how to execute bi amping a particular loudspeaker correctly. It’s not plug and play despite marketing claims by dsp equipment manufacturers. The acoustic performance of a loudspeaker designed for critical listening is a careful juggling act of numerous parameters which most audio enthusiasts are oblivious too. Transforming a passive loudspeaker to an equivalent or superior bi amped or Tri amp loudspeaker with external frequency dividing network and power amplifiers requires considerable technical insights, measurements and empirical adjustments. You can simply plug in an after market active frequency dividing network with one size fits all crossover slopes and crossover frequencies and it will go. But it’s unlikely to perform as well as the pre determined passive loudspeaker in all facets of subjective and measured performance.

    Prior to the pandemic l invested in considerable research and development of tailored electronic frequency dividing networks with Greg Timbers and Nelson Pass. We were able to engineer, design and construct tailored analogue electronic frequency dividing networks for pre determined loudspeakers that performed extremely well. However, the segment that would benefit most is perhaps unlikely to embrace the cost of such specialised high quality electronics. This segment has been showered by low cost mass consumer and low quality offerings that have conditioned the user into the belief that it should be a cheap and easy thing to do.

    So l am sitting on the fence and will probably supply such high quality equipment and the expertise to users who are seriously invested in this specialised area. That said l have seen Accuphase active crossover cards selling for US$500 a set on EBay.

    Ian

  4. #4
    Administrator Robh3606's Avatar
    Join Date
    Apr 2003
    Location
    Rocinante
    Posts
    8,170
    Yeah not for the faint of heart, Take a look at page 12! Control Operation.

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

  5. #5
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    See below.

    Technically
    In this instance an internal switch bypasses the LF passive network and the passive HP filter but the passive EQ remains in circuit. On perusal of the voltage curves the woofer not only has a crossover filter but system optimisation applied to the woofer. The optimisation accounts for how the loudspeaker will sound in actual use in a variety of listening rooms and account to for complex electro mechanical characteristics of the woofer low frequency response where the dcr of inductors impact on the low frequency response. The absence of the inductor dcr effectively lowers the QTS’ of the woofer seen by the bass reflex tuning. So what you win on damping you loose on the enhanced LF extension which is often audible. The net woofer sensitivity is also reduced by the passive network.

    To account for these factors we provide some focused tweaking specific the woofer. It’s not trivial and adds to the complexity and cost of implementation. Unless these intricate details are taken care of your better off staying passive. If done thoroughly and to a very high standard the results can be stellar. Precise level matching of the electronic crossover left and right low and high frequency outputs is required.

    The results
    This results in a remarkable improvement in pin point sound staging, sharper more visceral bass dynamics, spine tingling dynamic headroom and simplify a more effortless lifelike sound reproduction. It brings you closer to the emotional impact of the original live recording.

    Specific acoustic treatments of your listening room can then be investigated to manage reverberation and other audible problems your ears are sensitive too.

    This is not snake oil or marketing hype. It’s engineering a systematic approach to the mitigation of a number of product compromises that collectively impact on your listening experience.

    The implementation
    Implementation of an electronic frequency dividing network would require measurement of the volt drives at the driver terminals and or simulations. Furthermore an inbuilt LF low frequency shelf filter to fine tune the low frequency performance. To ensure you obtain the absolute pinnacle of straight wire transparency we use simple but elegant electronics circuits without the limitations of conventional mass production opamps. These circuits have over 100 x the value of key performance parameters such residual noise and class A sink and source current swing found in mass market audio op amps. This is to ensure the audio signal is never influenced by inherent design compromises in mass produced audio equipment.

    An analogy is once you own a really sharp very high quality carving knife for thanksgiving your never going to go back to a cheap knife! Not ever. What you don’t know won’t hurt you but once you have it you can’t live with out it. It’s a sad fact but true!

    The fun part
    The power amplifiers can be selected and optimised specifically for the low and high frequency duties. For example select a high performance digital amplifier such as a PS Audio power amp for powering the woofers. You then can focus on your own personal choice of high frequency power amplifiers to match to the rest of your system. You can literally sculpt the finer subjective characteristics of the overall system to suit your preferences.

    Ian
    Attached Images Attached Images   

  6. #6
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    Some of the design work in an electronic frequency dividing network for a pre determined loudspeaker system.

    See images below.

    Some of the design work in an electronic frequency dividing network for a pre determined loudspeaker system.

    See images below.


    1& 3. The top and bottom image is a modelled acoustic simulation of the Tri amp Everest DD67000. The in room response is well extended with the LF shelf filter we designed. The helper woofer signal is derived from cascading an additional low pass filter from the main woofer which is carefully designed so both woofers sum flat in the pass band.

    Greg Timbers designed the Everest DD67000 for tri amp operation should the user wish to explore tri amp mode of operation. Greg is at the top of his field when it comes to unique solutions for loudspeaker systems. Greg also has outstanding communication skills which enables other members of the team to pick up the solution and implement it correctly. Even so bringing this project together was a lot of painstaking hard work. The tri amp project was quite successful and a significant subjective advancement on the passive Everest DD67000.

    2. The second image is one of several prototype pcbs designed by Nelson Pass following detailed conversations with myself. The most recent iteration (image not shown) has a transversal design layout which mirrors how the user visualises the functional configuration of the crossover chassis The daughter boards are larger and are a consolidation of several of the smaller daughter boards with on board switching. This concept allows the user flexibility to either expand from bi amp to tri amp or quad amp mode while also offering flexibility to use different woofers and compression drivers subject to availability and performance criteria. The crossover slopes and crossover points are optimised ready for use with some tweaking room in the final implementation. Yes figuring that out did do my head in.

    3. In the third image down shows an actual measurement (not a simulation) the hump in the voltage to the woofer in a bass reflex system. Large woofers are often difficult to model accurately from T/L parameters and impedance measurements because the models rely on assumptions. So when we make real measurements we often find clues to why something sounds different. This problem exists with many pre determined bass reflex loudspeaker systems because the swings in the complex motional impedance cause a voltage drop across the inductor.

    4. The fourth image is a simulation of the vertical polar response of a two way loudspeaker in crossover region. This requires a number of off axis measurements of each driver. This representation allow selection of the best crossover topology for a particular woofer and horn.

    My understanding is that Greg has since come up with other approaches to tri amp of the Everest.
    Attached Images Attached Images      

  7. #7
    Junior Member
    Join Date
    Aug 2022
    Location
    Florida
    Posts
    3
    Thank you for your responses Ian! Very informative!

    It's safe to say that I don't have the ability or time to undertake this on my own. I heard in an interview that Mr. timbers actually owns a pair of S9900 that are driven Actively. My hope was that a schematic would be available somewhere and that an engineering friend or myself could undertake the construction of such a filter. I wonder if he or anyone else would be willing to share their hard work.

  8. #8
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    What l would say is you need to be committed and invested in such an undertaking.

    There are no short cuts and it’s not plausible in the diy sense unfortunately. The notion picking up a schematic and pulling a rabbit out of a hat isn’t feasible. If it was l would be out there publishing a book or selling kits. It just isn’t going to happen. People have attempted this in the past and you end up spending absurd amounts of time resolving questions, problems and dealing with peoples frustrations. Back in the 70’s Hafler and Dynaco supplied kits. But they were relatively simple and you switched it on and it went or it made smoke.

    The problem as you outline is what is referred to as the capability gap. The gap covers multiple disciplines including loudspeaker design, audio electronics and acoustics. That doesn’t cover how to install and use it. Even that can be very challenging. It’s like assembling IKEA.

    We all dream of a owning something and the idea pops into our heads we can make it. The enthusiasm often precedes the pragmatism and you fall in a heap.

    There are a few YouTube Clips of what’s involved in building a 500HP street car V8. Once you’ve watched an expert engine mechanic do it you would never dream of doing it yourself.

    When we design a product we understand that and we design and engineer it with a number of back end pre programmed core functions and just the right level of control for the user. We source the parts, test the parts, do Cad design, build prototypes, do beta trials to test and learn and finally package the product. There are also certain standards that need to be adhered to including CE compliance.

    An alternative is the dsp route but there is no free lunch there either. You still have the Gap. Someone has to interpret the jbl published information and define it technically. Then figure out the compatibility of dsp filters and then program, test and measure the results. That can be a very challenging undertaking if you don’t know where you are going with it all.

    I think Harman have dsp crossovers in some of their luxury consumer synthesis processors for the synthesis systems but they are really really expensive because the assumption is your not even remotely concerned about the price tag.

    Apparently DEQX is launching a new range of products soon. When you buy it you buy a level of technical support which is really important. They start at $5000.00.

    Sorry if this is a wet blanket. My point is not to be condescending. Your better of knowing the low down.

    If you want you can moderate your ideas and play around with some aftermarket active crossover. But with such an excellent loudspeaker why mess with it and end up with less than you have in its factory offering?.
    Last edited by Ian Mackenzie; 08-17-2022 at 09:17 PM. Reason: Typos

  9. #9
    Junior Member
    Join Date
    Aug 2022
    Location
    Florida
    Posts
    3
    Quote Originally Posted by Ian Mackenzie View Post
    What l would say is you need to be committed and invested in such an undertaking.

    There are no short cuts and it’s not plausible in the diy sense unfortunately. The notion picking up a schematic and pulling a rabbit out of a hat isn’t feasible. If it was l would be out there publishing a book or selling kits. It just isn’t going to happen. People have attempted this in the past and you end up spending absurd amounts of time resolving questions, problems and dealing with peoples frustrations. Back in the 70’s Hafler and Dynaco supplied kits. But they were relatively simple and you switched it on and it went or it made smoke.

    The problem as you outline is what is referred to as the capability gap. The gap covers multiple disciplines including loudspeaker design, audio electronics and acoustics. That doesn’t cover how to install and use it. Even that can be very challenging. It’s like assembling IKEA.

    We all dream of a owning something and the idea pops into our heads we can make it. The enthusiasm often precedes the pragmatism and you fall in a heap.

    There are a few YouTube Clips of what’s involved in building a 500HP street car V8. Once you’ve watched an expert engine mechanic do it you would never dream of doing it yourself.

    When we design a product we understand that and we design and engineer it with a number of back end pre programmed core functions and just the right level of control for the user. We source the parts, test the parts, do Cad design, build prototypes, do beta trials to test and learn and finally package the product. There are also certain standards that need to be adhered to including CE compliance.

    An alternative is the dsp route but there is no free lunch there either. You still have the Gap. Someone has to interpret the jbl published information and define it technically. Then figure out the compatibility of dsp filters and then program, test and measure the results. That can be a very challenging undertaking if you don’t know where you are going with it all.

    I think Harman have dsp crossovers in some of their luxury consumer synthesis processors for the synthesis systems but they are really really expensive because the assumption is your not even remotely concerned about the price tag.

    Apparently DEQX is launching a new range of products soon. When you buy it you buy a level of technical support which is really important. They start at $5000.00.

    Sorry if this is a wet blanket. My point is not to be condescending. Your better of knowing the low down.

    If you want you can moderate your ideas and play around with some aftermarket active crossover. But with such an excellent loudspeaker why mess with it and end up with less than you have in its factory offering?.
    I completely agree. Only if someone from the original development team had a bespoke solution would I consider undertaking this. Sincerely, thank you again for taking the time.

    -J

  10. #10
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942

    Thumbs down

    Quote Originally Posted by Ian Mackenzie View Post
    When we design a product we understand that and we design and engineer it with a number of back end pre programmed core functions and just the right level of control for the user. We source the parts, test the parts, do Cad design, build prototypes, do beta trials to test and learn and finally package the product. There are also certain standards that need to be adhered to including CE compliance.

    An alternative is the dsp route but there is no free lunch there either. You still have the Gap. Someone has to interpret the jbl published information and define it technically. Then figure out the compatibility of dsp filters and then program, test and measure the results. That can be a very challenging undertaking if you don’t know where you are going with it all.

    I think Harman have dsp crossovers in some of their luxury consumer synthesis processors for the synthesis systems but they are really really expensive because the assumption is your not even remotely concerned about the price tag.

    Apparently DEQX is launching a new range of products soon. When you buy it you buy a level of technical support which is really important. They start at $5000.00.

    If you want you can moderate your ideas and play around with some aftermarket active crossover. But with such an excellent loudspeaker why mess with it and end up with less than you have in its factory offering?.
    No worries.

    You will read online about people who have cloned the M2. Several people put a lot of thought into working out dsp filter solution(s). However it was an attempt at re engineering the JBL provided dsp solution for 3rd party dsp hardware. Some people did use the Harman dsp hardware and the provided dsp filter.

    The challenge they faced was implementing and installing both the Harman and the 3rd party hardware and 3rd party dsp filters. It’s all over a number of online forums. Out of this came unforeseen problems that required further intervention. Things like the frequency response not being quite right and loud fan noise from the Harman recommended equipment which speaks to it not being recommended for residential home use. Your left hanging there on your own to figure it out yourself.

    The question to ask is when you own the original product do you really want to indulge in that kind of nagging anxiety in attempting to get it all to work? You shouldn’t need to. Do you want to be farting around for hours, days, weeks or months with a tiny LCD screen trying to program a dsp processor with a really complicated dsp menu written by an under paid nerd in Mumbai?

    What l was working on with Greg Timbers (now retired) and Nelson Pass was a complete end to end solution including set up instructions, online technical support, telephone support and above all something that was tested and that did what it was supposed to do.

    Unfortunately even if people are provided all the help in the world they can still not follow the advice and get it wrong. That’s also a very real risk.

    Without the whole package it’s still up to you to make it happen properly.

    This type of product offer is not diy situation or a Salvation Army or a Red Cross exercise or philanthropy.

    Unfortunately the pandemic and some other issues have put that sizeable undertaking on hold. That kind of thing is very specialised and needs a sunny day scenario happening to let it work properly. Not to mention the time involved to support it. Some industries are suffering the pandemic hangover and are running twelve months behind while other start ups are going into administration. When l do offer it it will be to special order.

  11. #11
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    I modelled the S9900 bass reflex system over the weekend to look at the impact of bi amp mode on the bass response of this system.

    In summary bi amp operation will require a degree of finesse on the part of the user by way of careful equalisation in the bass response to ensure the same overall balance. As already discussed earlier the bi amp voltage drive curves are specified are unusual and are not going to be found in an off the shelf active crossover. I would recommend accessing a technician to set up your dsp active crossover and measure the system to ensure it’s functioning correctly. Otherwise it might sound odd or weird and you won’t be happy.

    As it stands the S9900 is tuned to 36 hertz in a 3.4 cu ft3 (96 litres).

    Using available data on the 1500AL-1 woofer used in the S9900 l constructed a model of the T/L parameters. See attachments. I then simulated the optimal bass reflex enclosure volume and the port tuning for the woofer. Using online Alignment Tables l was able to validate the driver T/L model by correlation of the S9900 tuning and enclosure volume with tuning from the tables. The results are really close. See attachments.

    My driver model used JBL published data and data from the prior 1500AL driver. My 1500AL-1 model has a QTS of 0.26, QES of 0.27, a VAS of 264 litres and an FS of 30 hertz.


    To obtain a true model of the bass reflex system however requires incorporating the dc resistance of the passive crossover filters. In the S9900 there are two inductors in the woofer low pass filter with a combined dc resistance of 0.9 ohms. Allowing for loudspeaker cables, terminations and the output impedance of the amplifier l added an additional 0.2 ohms of resistance. The combined dc resistance is 1.1 ohms.

    That’s significant. However, it would seem that the JBL engineers had foreseen this and increased the enclosure volume by about 30 litres to obtain optimal tuning of the bass reflex system. Technically the added dc resistance creates was is referred to as the modified QTS’ which is 0.31. The unmodified QTS is 0.26 according to my driver modelling. This was confirmed by optimal enclosure volume and tuning frequency using the online Alignment Tables which are 98 litres and 37 hertz. The results correlate with the S9900 enclosure volume of 96 litres and tuning frequency FB of 36 hertz. This is close enough for modelling purposes.


    When bi amp mode is used the impact on mid bass response is that it starts to fall away at a higher frequency. For comparison purposes the bi amp mode has an F3 is 54 hertz and the F6 point is 45 hertz. In passive mode the F3 is 45 hertz and the F6 is 36 hertz.

    So what do all these simulations mean?

    If you are a S9900 owner the thing to appreciate is that the S9900 had been set up and tuned specifically for passive operation. If you are considering bi amp operation of this particular system be aware that this will alter the bass frequency response and some equalisation will be required if you you want the system dialled in exactly as it should be. On some program material this may be noticeable. The bass may sound lean and less extended. Without being aware of this you would be wondering why does the bass sound different? If you have a active dsp crossover at your disposal and a technician who can spend time with you to dial it in that would be the way to go. The issue is this system is not a plug and play bi amp loudspeaker. The bi amp voltage drives are specific and will also need be addressed. An analogue active crossover could be designed to do the job in a Spice Simulation. I just have got around to it yet.

    Incidentally in tri amp mode of the DD67000 the same issue arises. Greg Timbers recommended a bass shelf at 100 hertz with +4 db of boost. Greg said the bass could sound lean otherwise.

    Other than that some users have added a sub to cover the 1st octave with the S9900.

    If you refer to the published technical information on the S9900 be aware that JBL measurements are under specific acoustic conditions and may not correlate with the simulations l have modelled here.

    Image 1 Bi amp mode F3
    Image 2 Bi amp mode F6
    Image 3 Passive mode F3
    Image 4 Passive mode F6

    Edited due to errors in initial simulations and text re written. The new simulations attached are taken directly from T/L QB3 alignment tables for clarity. Unfortunately l was unable to overlay the curves. There’s no fudging of the modelled results which show the QT radio button selected for both the bi amp case and the passive case.
    Attached Images Attached Images     

  12. #12
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    Simulation data attached.
    Attached Images Attached Images    

  13. #13
    Administrator Robh3606's Avatar
    Join Date
    Apr 2003
    Location
    Rocinante
    Posts
    8,170
    Hello Ian

    What software are you using for the simulations??

    Rob
    "I could be arguing in my spare time"

  14. #14
    Senior Member
    Join Date
    Apr 2003
    Location
    Australia
    Posts
    7,942
    Quote Originally Posted by Robh3606 View Post
    Hello Ian

    What software are you using for the simulations??

    Rob

    http://www.mh-audio.nl/Calculators/BRA.html

    I also use an App called Speaker Box Lite.

  15. #15
    Member sebackman's Avatar
    Join Date
    Feb 2004
    Location
    Europe
    Posts
    675
    Great info in this thread. Good work Ian!

    For us using DSP it would be very interesting seeing what actual settings Mr Timbers are using with his 9900's. Not only the XO's points but also the slopes, delays, Eq points and potential phase shifts.

    As he no doubt have had time and knowledge to achieve results the rest of us would have a difficulty to reach, we could all learn from his choices.

    Is there a way to find out?

    Kind regards
    //Rob
    The solution to the problem changes the problem.
    -And always remember that all of your equipment was made by the lowest bidder

Thread Information

Users Browsing this Thread

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

Similar Threads

  1. Altec Santiago 878B filters
    By Hans in forum Lansing Product Technical Help
    Replies: 3
    Last Post: 04-20-2011, 02:56 PM
  2. About 3-Pole Bump Filters ?
    By Earl K in forum General Audio Discussion
    Replies: 1
    Last Post: 03-06-2006, 01:57 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
  •