new RD-75 project. Comments???

Collapse
This topic is closed.
X
X
 
  • Time
  • Show
Clear All
new posts
  • Victor
    Senior Member
    • Apr 2002
    • 338

    new RD-75 project. Comments???

    I am in the planning stages for my new dipole 4-channels HT system. All channels will be based on RD-75 driver. Here is what I have in mind.

    For my two front channels I would like to build a 3-way system, with each side consisting of:

    1. The RD-75 in its own baffle. This will cover 300Hz and up range.
    I am going fabricate a one-sided trapezoidal baffle with a sloping 6-inch diameter half-round (found in http://www.tapeease.com/2rounds.htm ) on one side of it. I will route the opening for the RD-75 driver close to the vertical edge of the baffle. This vertical edge will be treated with a 1.5-inch half-round. The top will be about 8 inches and the bottom about 14 inches.

    The straight front will blend into a 6-inch half-round on the slopping side. The other end of this 6-inch diameter half-round will be connected to the slopping straight trapezoidal section. This section will be connected to the back of the front section. Resulting in a pizza-piece-like cross-section. All in all, the baffle should have no edges.

    2. A midrange array consisting of 4 Scan Speak 10-inch drivers, the 25W/8565 in Phoenix baffle (http://www.linkwitzlab.com/m_panel.htm). That is, the baffle will be identical to the Phoenix design just a bit longer to accommodate the 4 drivers. The drivers will be separated by 1 inch. So the total length might be about 50 inches. This baffle will have its own 1-inch foundation plate. This array will cover 70-300 Hz range.

    3. The bass will consist of 2 Phoenix W-woofers. I intend to use the implementation as shown in Woofer3 in Linkwitz’s site (http://www.linkwitzlab.com/woofer3.htm on the very bottom of the page) utilizing the Peerless 830500 drivers. The two cabinets will be setup on top of each other. Naturally they will be positioned facing in the opposite direction. The woofers will cover up to 70 Hz.

    I intend to position the Phoenix midrange baffle on top of the woofers. So it will be a 3-unit structure of about 78 inches (14+14 inches for the woofer cabinets and 50 inches for the midrange) in a total height. The separate baffle housing the ribbon will be placed next to it.

    For the two surround channels, I am thinking of pairing the H-frame woofer, as shown in http://www.linkwitzlab.com/proto.htm#PW1 crossed at 150-Hz with RD-75 ribbon housed in the baffle as discussed above. The RD-75 will work down to 150 Hz and since surround channels do not do all that much work, I hope this combination will work out.

    I intend to use an active crossover with all channels. Well, I would greatly appreciate any comments on this system.

    Regards,
    Victor
  • John Holmes
    Moderator Emeritus
    • Aug 2000
    • 2703

    #2
    Welcome to the Guide Victor!

    I have nothing in this area to offer you for, my knowledge in this area is well...depressing to say the least.:B However, I'm sure the pros in this arena will be along shortly to help you out.

    Again, welcome.




    "I came here, to chew bubble gum and kickass. And I'm all out of bubble gum!!!" My DVD's
    "I have come here, to chew bubblegum and kickass. And I'm all out of bubblegum!!!"

    Comment

    • ThomasW
      Moderator Emeritus
      • Aug 2000
      • 10934

      #3
      Victor

      That's a pretty complicated design. Sort of reminds me of a Chinese menu. Taking one item from every column anyones ever used to make a RD hybrid.

      How about a little background, what speakers have you built prior to these?




      theAudioWorx
      Klone-Audio

      IB subwoofer FAQ page


      "Complicated equipment and light reflectors and various other items of hardware are enough, to my mind, to prevent the birdie from coming out." ...... Henri Cartier-Bresson

      Comment

      • Victor
        Senior Member
        • Apr 2002
        • 338

        #4
        Thomas,

        I am more of an electronics guy. That is to say that my ‘war chest’ of projects contains an even dozen of power amps, pre-amps, D/A converters, active crossovers, etc. that I designed and build. I currently have a HT consisting of 4 Carver Amazing Mark IV speakers. Active crossovers are used throughout. Each channel gets 600 Watts into woofers and 300 Watts to the ribbon. All my amps are discrete transistor A-B class designs with secondary feedback structure reminiscent of Bryston topology. The amps do better then 0.05% THD at all frequencies and power levels. My D/A converter has undergone a few revisions over the years with the latest addition of jitter suppression circuitry based on a use of Asynchronous Sample Rate Conversion technique. I can describe my D/A converters and pre-amp in more detail if you are interested.

        As for the speakers, my most difficult project to date was a servo-controlled subwoofer. I borrowed an idea from an article in Wireless World magazine, changed a few things and ended up with a small sealed 15-inch cube with two 10-inch drivers in it facing in the opposite directions. The drivers were by Volt with dual voice coils. So I used one of the coils to get a motional feedback signal. This feedback signal gets filtered and shaped and then fed back into the amp. The result is near 16-100 Hz performance (-3 dB) with about 2% THD at near 100 dB levels.

        Well, having said all that I must confess that I am not much of a woodworker. However, a close friend of mine is a master cabinetmaker with significant speaker building experience (horns and Lowthers) and it will take both of us to build the project as I described in my first post. Between his ability to work the wood and my experience with electronics and measurement, I think we can do it.

        I appreciate your sentiment that what I propose is complex and looks like a Chinese menu. This is true. However, I luck the ability to do a fundamental speaker design. So my intention was to ‘stand on the shoulders of others’ and borrow from their experiences. I would like to stick with proven designs and for that reason the work done by Linkwitz appeal so much to me. Also his math looks full proof to me just as well.

        Well, naturally before I jump into the a complex project I would like to hear from more experienced people in hope of discovering the ‘holes’ that inevitably exist in any plan.

        Regards,
        Victor

        Comment

        • ThomasW
          Moderator Emeritus
          • Aug 2000
          • 10934

          #5
          Victor

          Ok great.

          Somethings to ponder. I've been corresponding with Dan Wiggins of AdireAudio, he's a pretty savy designer. They're working on a design using the RD50. I ask him about the RD75 vs the RD50 here's what he said.

          Thomas,

          Yeah, the RD75s do have some drawbacks; the RD50s doesn't seem to drop off the high end nearly as much as the RD75s. We're also looking to cross over to a pair of AV8s around 800 Hz, to keep things happy all around...

          Another thing is the dispersion in room. It really takes a different voicing to make them sound decent in a room; you have a much faster off-axis attenuation (vertical) than you do with domes or cones. means you have to bias the on-axis to be a bit hotter.

          Also, there will be a MONSTER center lobe, vertically speaking, from about 4 kHz and up. You get VERY little vertical dispersion above this frequency.

          The power response is definitely skewed compared to cone/dome systems, and that's the reason so many DIY systems based on the RD units sound dull. They measure flat on-axis, but in-room, where the power response needs to be considered as well, they end up sounding flat.

          Dan Wiggins
          Adire Audio
          Jon and I did measurements of the RD75 in a simple test baffle. Details are HERE These were done with a calibrated B&K 4133, using a Clio system.

          re: your system, the RD75 baffle design seems promising, complicated but promising. Do you happen to have a CAD drawing of it?

          My only real concern is the groups of midwoofers and subwoofers. It seems like a ton of work and money to make the Linkwitz design for both sets of drivers.

          Are you familar with the Lambda 15"s, purpose build for dipole use?

          I can't find a link on Nick's site but here's a post about them





          theAudioWorx
          Klone-Audio

          IB subwoofer FAQ page


          "Complicated equipment and light reflectors and various other items of hardware are enough, to my mind, to prevent the birdie from coming out." ...... Henri Cartier-Bresson

          Comment

          • Victor
            Senior Member
            • Apr 2002
            • 338

            #6
            Thomas,

            Fundamentally, I can use RD50, I suppose. However, it would be nice to know exactly how ‘high’ can RD50 go as compared to RD75. I was intending to use a judicial amount of high spectrum boost in the crossover to bring RD75 to be reasonably flat at around 15 kHz. Given that I just crossed the 40 mark, it is unlikely I can hear anything past 16.5 kHz anyway. Also it seems that by crossing at 800 Hz, as Dan Wiggins suggests, the main advantage of seamless upper midrange of RD-50 (75) is lost. That is the reason I am thinking more in terms of 300 Hz crossover to midrange.

            Yes, I have read about Lambda’s 15-inch drivers. However, nobody I know has so far utilized them. The specs do look good, however. The attraction of Lambda drivers is in a possibility of a 2-way system design. The problem is that the Lambdas will never play down to subsonic region. In a 2-way dipole system the Lambdas may do about 30 Hz, so the need for the subwoofer will be apparent. That brings us back to the 3-way system considerations, which may not be any less complex then what I originally proposed. What do you think?

            I do think that Dan Wiggins has a good point regarding the very uneven power response of RD-50 in the room. But it seems that this is the price I may have to pay for the use of a line source. Unfortunately one cannot predict the results here as every room is different. I will strive to achieve the flattest response in my room. I may use the Berringer unit to EQ the room modes in the end.

            Thomas, I did see the pictures of the baffle you made. It looks great! Those Eton drivers you used are rather inexpensive relative to what I have in mind, but you used 8 of them when I am thinking of going with 4. I agree with you that it seems a lot of work and money to build my system, but it seems that a 3-way system is inevitable in a dipole approach. I do not know of any other way to build a 20-Hz capable dipole woofer other then Linkwitz’s W-woofer. And I will need 2 of them per side in the front to achieve realistic SPL levels.

            The less expensive alternative, as I can see it, is to go with sealed sub-midrange combination. So, do you think my system has merit or I am asking for trouble here?

            Regards,
            Victor

            P.S. I have a sketch of my proposed baffle. It is not much, but it gives a general idea. I would post it here if I could figure out how to do that. Perhaps you can instruct me as to the process involved in posting a picture.

            Comment

            • Victor
              Senior Member
              • Apr 2002
              • 338

              #7
              John Holms,

              Thank you. I came to this forum kind of by accident and I am looking forward to learning a great deal from the experiences of the people who post here.

              Victor

              Comment

              • ThomasW
                Moderator Emeritus
                • Aug 2000
                • 10934

                #8
                Victor

                The link is just a test baffle. There's an asymetrical baffle waiting to be filled with drivers, but I've been sidetracked with other projects.

                If you look at the nearfield plot of the Etons you'll see they go down 80hz with no problems and have a very gentle roll off below that. Also mounted as you see them in the test baffle they are much more efficient than the RD75s. So I think that instead of using 10"s and 12"s you might get better performance with 8"s and a good sub (see comments about this idea below)

                I to was originally planning on a relatively low XO point 250Hz. With the conclusion of the intial tests, Jon suggested raising the XO point to 500Hz or so. He's also concerned that the planars maybe overstressed trying to reach higher SPLs at 250Hz.

                We'll be running tests on the black trapazoidal baffle in mid May.

                What confuses me is that our testing and limited listening seems to indicated that the RD75 aren't really world class drivers, but more likely a fairly good planar midrange. This conclusion runs contrary to most info I've read and what's been posted by owners. I have run accross a couple of people that validated our conclusions.

                Understand I'm not often swayed by the opinions of others. But in this case the info is baffling.

                I've looked for alternative planars but other than the Metaxas DIY ESL's, I've been unable to find anything suitable.

                Anyway back to your design....

                I don't think you'll find any drivers that have adequate output at subsonic frequencies that are practical for use in any open baffle. That is unless your willing to create a immense array, with very large drive units. Given your electronics background it would be possible to force a design to operate that way with enough power and EQ.

                I'm thinking that a true standalone sub is the best reasonable solution. Now if you have space for a real IB sub that's a whole different ballgame

                I'm unaware of any Behringer products that provide boost below 20Hz. The Symetrix 551/552 will provide boost down to 10Hz. But again with your background developing a custom unit shouldn't be an issue.

                To have a picture show in a post, the file must be uploaded to a web server. Then a link is placed in a post that puts in a call to the server where the image is stored. If you click on the 'edit' for this post you'll see how the image below is linked



                Thomas




                theAudioWorx
                Klone-Audio

                IB subwoofer FAQ page


                "Complicated equipment and light reflectors and various other items of hardware are enough, to my mind, to prevent the birdie from coming out." ...... Henri Cartier-Bresson

                Comment

                • Victor
                  Senior Member
                  • Apr 2002
                  • 338

                  #9
                  Thomas,

                  It will be very interesting to see the test results of your new baffle. I spent quite sometime reading the work of John Whittaker and my baffle is a derivative of what he did. Perhaps I can e-mail you the sketch that I have and then you might be able to post it, as I do not have an access to a web server. Anyway, my baffle will have no sharp edges at all and it will still be an asymmetric trapezoid. The dimensions of my baffle are dictated by the directive from Mr. Grebner. He felt that the total frontal area for the RD-75 baffle should be about 15x72 inches. I preserve the area but change the shape. The 6-inch half round I mentioned in the previous post is there in an effort to minimize the diffraction effects. The overall shape is as a variation on a John Whittaker's “compound baffle” approach. See his site for more http://www.snippets.org/alsr/

                  Your contention that RD-75 is really a midrange driver is well taken. Yes, I have heard that a few people mentioning that a super-tweeter might be necessary to ‘cure’ the high-frequency roll-off. Personally I think that in this case the 'cure' might be worst then the ‘disease’. In my experience with Carver 60-inch ribbons the high-frequency roll-off was never a problem once I went with the active crossover. The response of the Carvers can be made flat to 16 kHz with high spectrum boost. I would imagine the same is true for RD-75 since the technology is similar. In my ‘book’ 16 kHz is all I will ever need. But then again there are other opinions. I just think that it will be a very difficult undertaking to blend the super-tweeter in.

                  As for my project here, perhaps I should take your suggestion and consider a “true standalone sub”. What can it be? I think a single driver sealed box is definitely a possibility. I am not sure how would the dipole ribbon and the dipole midrange blend with a sealed sub. Do you think the combination will work?

                  I did consider the 8-inch drivers for the midrange, but then I re-read Linkwitz site again. From the material on his site it would seem that he thinks that 8-inch drivers are not adequate in general to go as low as 100 Hz in a system where large volume displacement is needed. So that is the reason why I was considering 10-inch units. The SS 25W/8565 is particularly suitable because I am thinking about using Phoenix baffle. I am sure that 4 drivers will be sufficient for midrange, although I can go with 6 and create an array of the same height as RD-75. Do you think a stand along shallow H-frame midrange tower 12.5 inches wide, with six 10-inch woofers is a way to go?

                  You mentioned an IB approach. In fact there is that possibility in my basement. I have a utility room situated in the center of my basement. My HT is located in the open area in the corner of the basement with utility room within 3 feet from one of the front speakers. In principle I can create an IB array utilizing this utility room, as it is big enough, but it would have to be a mono channel. So then I can end up with a setup where the frequencies below 70 Hz will get summed and channeled to this IB array. I do not know if this is advisable to do. As I see it there are at least 2 problems here, - the mono bass and the location of IB drivers. Actually I can place the IB drivers approximately 3 feet in front of the speaker and 3 feet away from the same speaker. The IB array will fire at 90 degrees relative to the firing position of the front speaker. Well, what do think of this?

                  Regards,
                  Victor

                  Comment

                  • ThomasW
                    Moderator Emeritus
                    • Aug 2000
                    • 10934

                    #10
                    Victor

                    I too have discussed variables with John Whittaker, he's an interesting fellow.

                    IB bass is IMO the best there available. I assume you've seen my IB subs, if not they are linked from the Klone-Audio link below

                    I have a large ESL array matched with the 12 Shiva IB. And a pair of Legacy Audio Whisper klones matched with a smaller 4 Shiva IB.

                    To take the workload of the ESL panels the big IB uses an mono 100Hz XO for HT and is lowered to 75-80Hz for critical listening of music. There is no smearing of the stereo image with the mono sub at either XO frequency.

                    Yes the 6-10's will certainly move more air when compared to my 8-8"s. But the 8-8"s are more efficient that the planars. So more displacement isn't necessary. Their output at 100Hz is fine. They have a roll off below 100Hz that's very similar to the SS 25W/8565 according the the FR plot on the Vifa site. But....

                    One of the main issues with any hybrid, is matching the sonic signature of the cones to that of the planar. This situation is exacerbated as the XO points go higher.

                    The original reason for choosing the Eton kevlars is that they are 'reasonably' transparent. However in recent months we've been building smaller 2/3-way systems using metal cone drivers. These offer even improved transparency and should IMO be given serious consideration.

                    I see no major issues with the "H" baffle for the midbass drivers. I'd use an asymetrical baffle to take advantage of it's reinforcement characteristics.

                    Thomas




                    theAudioWorx
                    Klone-Audio

                    IB subwoofer FAQ page


                    "Complicated equipment and light reflectors and various other items of hardware are enough, to my mind, to prevent the birdie from coming out." ...... Henri Cartier-Bresson

                    Comment

                    • JonMarsh
                      Mad Max Moderator
                      • Aug 2000
                      • 15259

                      #11
                      Interesting discussion, Victor and Thomas! I'll be very interested to hear about the final outcome of your project, Victor.


                      I'll toss out a few comments- you may have already thought of all of these, but just in case....

                      Your baffle concept for the RD75 or RD50 sounds fine- I think it will work well.


                      My concerns about how low the RD50 or RD75 can be run stem from issues regarding diaphragm displacement required at lower frequencies, and also the desirability to have as seemless a transistion between the midbass and ribbon transducer. The former affects distortion at higher output levels, of course, and the overall sense of ease for the finished speaker.

                      The latter is influenced greatly by the degree of clean driver overlap in the crossover region, which in general is fairly critical to getting a good transistion between the different speaker ranges.

                      Personally, I think it could be worthwhile to investigate crossover points using an active crossover, but my gut feel says I'd be happiest with a 400 Hz or higher crossover. It's quite possible to get dynamic drivers with pistonic operation up to 800 Hz, at 8" cone size; higher, of course, takes more exotic construction.

                      I find it interesting that there are many opinions about why you shouldn't crossover in one frequency range or another, and these are given as primary axioms guiding design choices for a particular speaker system.

                      I will concede that a poorly designed crossover may be more immediately audible in some frequncy ranges than another. It's also true, that due to driver physical size and spacing, some crosover frequency ranges (such as high treble) will present much greater integration challenges.

                      However, I'd suggest that any poorly implemented crossover is to be avoided. Though it's affect may not be as immediately identifiable to some listeners, the problems will nonetheless remain, and be quite audible with the right program material.

                      A "classic" example, in my opinion, is subwoofer integration- it's remarkable how many sub's have problems operating linearly in the octave above 100 Hz, yet folks will try to use them with a 100 Hz crossover point with "classic" symmetrical crossovers, and wonder why the integration and balance can't ever be gotten "right".

                      A simple guideline I'd suggest for anyone implementing a multi-way speaker with standard crossover topologies is to select drivers which they can operate with one octave of clean overlap above and below the crossover region. This would mean, for example, in a subwoofer setup, having mains which go clean and flat down to 50Hs, and a sub that goes up clean to 200 Hz, if you're using a 100 Hz crossover. Adjust the criteria appropriately, as you change the frequency.

                      A minimum guidline would be at least a total overlap of not less than one octave- this is less preferred, but in the previous example, would require mains flat to 75 Hz, and a sub clean up to 150 Hz.

                      I can't tell you how many times I've heard abysmal sounding sub/satelite systems where the sub starts croaking about 125 Hz (or even lower), and the sats can hardly make it down to 90 or 100 Hz. This is a recipe for problems.

                      With ThomasW's project, our plan is to conduct a series of measurements on and off axis to investigate the power response of the RD75 with the notch filer installed. I appreciate the logic behind Dan Wiggins comments about needing to run the ribbon's "hot" on axis, to balance the room power response, but I'm not sure I really feel comfortable buying into it; I think the on and off axis response of a system should be carefully balanced, and making it "hot" for the on axis (early arrival) sound, in order to balance the reverberent (room power response) has always been problematical in my experience. I'd suggest that it would be better to change the driver operating ranges or use a line source with similar midrange and LF lateral dispersion, so that they match up. The biggest single problem of the RD ribbons is the set back cavity of the front panel; this accounts for the 5-6 kHz cavity resonance which must be tamed with a notch filter, and the limited lateral power response above 5 kHz.

                      I have to wonder if a more pleasing overall balance on and off axis couldn't be achieved by getting off the RD before the cavity resonsnce and dispersion issues crop up, and using some other line source tweeter (with a steep crossover) with similar dispersion as the RD ribbons have below 5 kHz. The question becomes, what upper range ribbon to use?

                      Best regards,

                      Jon




                      Earth First!
                      _______________________________
                      We'll screw up the other planets later....
                      the AudioWorx
                      Natalie P
                      M8ta
                      Modula Neo DCC
                      Modula MT XE
                      Modula Xtreme
                      Isiris
                      Wavecor Ardent

                      SMJ
                      Minerva Monitor
                      Calliope
                      Ardent D

                      In Development...
                      Isiris Mk II updates- in final test stage!
                      Obi-Wan
                      Saint-Saëns Symphonique/AKA SMJ-40
                      Modula PWB
                      Calliope CC Supreme
                      Natalie P Ultra
                      Natalie P Supreme
                      Janus BP1 Sub


                      Resistance is not futile, it is Volts divided by Amperes...
                      Just ask Mr. Ohm....

                      Comment

                      • Victor
                        Senior Member
                        • Apr 2002
                        • 338

                        #12
                        Jon,

                        Thank you for your vote of confidence in my baffle concept. I do think that once the sharp discontinuities are minimized or all together eliminated, the diffraction related problems would be minimized. I also think that this baffle is not all that problematic to fabricate, so I intend to start on it a week or two at the most once I get the half-rounds from tapeease.com

                        I understand your comments regarding the best crossover point for the RD-75 and I fully intend to do some experimentation to determine the ‘best’ crossover frequency. I have on hand 2 Bryston 10B crossover boxes. They will enable me to create a stereo L-R 4th order variable crossover. So we shall see…

                        It would seem that crossing the RD-75 ‘before’ the 6 kHz resonance is a very good idea. However in my previous post I did express skepticism toward the integration of the super-tweeter. However, have you ever considered the Newform Research ribbons? They seem to do a lot better in the top end then RD-75 and they come modular, so that you can integrate one or more elements. I also seem to recall that they are monopolar in contraction. Anyway, perhaps it might be something to look at.

                        Yes the poor crossover implementation does often lead to a rather sad result. You see that with my “Chinese menu “approach, I can liberally ‘borrow’ from the crossover design that Z. Linkwitz had done for the AA Beethoven Grand speaker. This speaker uses the same 12.5-inch wide shallow H-frame baffle that I intend to use for my midrange. It also uses 4 SS 25W/8565 drivers at the same crossover frequencies (70-300) as I intend to use them. Furthermore, the Phoenix W-woofers crossover implementation is very well described on Linkwitz’s site. Finally I have very good understanding of what to do with RD-75 active crossover, given my experience with Carver 60-inch ribbon. In the end, this is a rather complex piece of electronics, but I am confident I can do it.

                        However, at this time, I am more in need of validating the proposed system architecture. In many ways I need the answers to the following:

                        1) Is it a good idea to stack the two W-frame woofers and the 50-inch long midrange section on top of each other?
                        2) Will I have problems with phase while attempting to position the RD-75 baffle next to this woofer-midrange tower?
                        3) Is the system design too complex and also expensive for the kind of ‘sonic’ return it will offer?
                        4) Should a good single driver sealed sub be just as ‘good’ and therefore a better and more economical solution as it is pared with the dipole midrange and dipole ribbon combination?
                        5) Is there anything else I can do to make the system less expensive and utilize the same possible performance?

                        Al those questions drive me positively nuts, but I must have a better picture in my head before I jump headfirst into this. My friend the cabinetmaker, says that audio is a progressive disease….

                        Regards,
                        Victor

                        Comment

                        • ThomasW
                          Moderator Emeritus
                          • Aug 2000
                          • 10934

                          #13
                          Victor

                          I'll let you and Jon continue the technical aspects of this discussion. He like you is an engineer. I'll supplement the discussion where approriate.

                          The NewForm ribbon also has a soft top end to. I've discussed this with John Meyer, he tends to downplay it. Since his single drivers can't be returned after purchase, I'm leary of buying for testing purposes.

                          Some leaf/ribbon/planar tweeters to consider are:

                          1) Phillips RT8P
                          2) ATD
                          3) ESg2-3
                          4) Raven 2




                          theAudioWorx
                          Klone-Audio

                          IB subwoofer FAQ page


                          "Complicated equipment and light reflectors and various other items of hardware are enough, to my mind, to prevent the birdie from coming out." ...... Henri Cartier-Bresson

                          Comment

                          • JonMarsh
                            Mad Max Moderator
                            • Aug 2000
                            • 15259

                            #14
                            1) Is it a good idea to stack the two W-frame woofers and the 50-inch long midrange section on top of each other?
                            2) Will I have problems with phase while attempting to position the RD-75 baffle next to this woofer-midrange tower?
                            3) Is the system design too complex and also expensive for the kind of ‘sonic’ return it will offer?
                            4) Should a good single driver sealed sub be just as ‘good’ and therefore a better and more economical solution as it is pared with the dipole midrange and dipole ribbon combination?
                            5) Is there anything else I can do to make the system less expensive and utilize the same possible performance?
                            I'm kind of swamped today with work stuff, camped out in a hotel in Baltimore on the East Coast, trying to catch up with stuff I couldn't do while traveling on business last week. I'll respond to a degree now, but probably update more later in the week, if time permits. (I hate traveling two weeks away from home in a row :x )

                            1) How tall will your woofer sections be- sounds like this will be well over six feet tall. How high are your ceilings?

                            2) This is starting to sound like one of the big Genesis systems in construction. Thomas and I figured we would want to "integrate" the ribbon and midwoofer line array. I can't think of a reason (off hand) why placing the ribbon baffle and the midwoofer/woofer array in such proximity would create problems- remember, direct lateral cancellation *should* minimize any acoustical interaction.

                            3) This is really the crux of the matter, isn't it? I had similar questions when starting my bi-amplified X1 Klones, and they're a lot simpler in some ways than the system you propose. Here are some inputs based on my experience in the last year- they may be quite relevant to your goals, or you may feel otherwise.

                            First, late me state that I've developed some marked "bias" regarding what electronics in their various configurations do to the overall system performance. I've developed a bit of a bent towards minimalist approaches that do an absolute minimum "damage" to the signal- I prefer a system that has absolutely breathtaking reproduction of voice and piano, acoustic guitar and standup bass, even if the deep bass and extreme treble left a little to be desired in extension.

                            If not careful, when developing an elaborate system one can build up a system which, in the end (IMO), is what I would call very impressive "mid-fi". An example would be what I was listening to a lot in 2000- my X1's, driven with it's custom electronics crossover, using OPA627's buffered by AD815's, incorporating baffle step EQ and LF EQ in the crossover, with Aragon Palladium's on the upper modules, and 8008's on the bass cabinets. Source component was my Sony SCD777ES SACD player. Preamp was a modified Marchand PR41 passive unit, using Shallco rotary switch attenutators, which I still use.

                            Contrast this with the current living room setup:
                            A pair of my MkII M8 bookshelf speakers (2nd development set of the M8 series, using Eton 8-800 9" midwoofers, Accuton C23 tweeters), an Ayre V-5 NLFB amplifier, a custom built and modified MP-DAC (built and modified by me; this is another NLFB design, using a Crystal 4390 with an "obsessive" regulation and isolation scheme, a passive output filter, and discrete NLFB output amplifier), used either with the SCD777ES as transport, but more recently a heavily modified Philips CD753 as a transport.

                            A MkIV set of the two way is being built, using HiVi M8a drivers and Vifa XT25; these will ultimately be used in my secondary system, or as surrounds, but in the interim may be combined with a Sumiko Delilah NLFB crossover and subs I'm building with Blueprint 1203s.

                            Another system currently in development, which, pending the outcome, may become my main home system, is a dipole design, not a line source, but using two HiVi 8's for the midwoofers (~250-1.25 kHz), the Vifa XT tweeter, and so far the top candidate for the bass (down to about 75 Hz) is the Titanic 10" woofer (either two or four per speaker). The midwoofer to tweeter crossover is the same high slope design I'm using in the bookshelf we've been playing with (article pending).

                            I plan to crossover from dipole bass to a more conventional sub below 75 Hz, owing to the losses which occur from cancellation with very LF dipoles.

                            This system is being developed for a passive crossover, as I wanted to have something that was semi full range running off a single amp, for electronics evaluations. Now, it's not going to be "cheap" to build, but it really shouldn't be too hideous- excepting that the efficiecny won't be all that great.

                            What I really *don't* want to do with it is use a lot of opamps in the signal path, in a complex electronic crossover and equalizer, so I'm trying to select the drivers and acoustical layout to permit using passive networks with high quality components.

                            I'm also going to be rebuilding my Aragon's (including the Palladiums) pretty soon with a new high power NLFB circuit I've been working on, so having some voltage and current drive available shouldn't be a problem. The Palladiums are balanced monoblocks which output about 125 watts class A (at 8 ohms), but into 8 ohms will exceed 600 watts before clipping, and over 1 kW at 2 ohms.

                            In case you haven't noticed, I've become a convert (if reluctant at first) to non loop feedback designs. I've designed some pretty good conventional circuits in my time, including some custom power amps for a pro company that won in some shootouts against some highly regarded monoblocks in $10K each category. A lot of people are skeptical about this, but go listen to an Ayre VX1, or V5 or V6 sometime. (there's a good review of the V6 in the current "The Perfect Vision"). Ayre's designer used to work for me in the 70's, and we've kept in touch over the years, and I've done some work for them, and given some inputs since.

                            The point of this meandering monolog, is that I believe an "elaborate" system, to be truly successful, shouldn't take anything "away" or add anything "bad" to the reproduced music that a simpler, smaller "purist" system will accomplish.

                            For many, this requires some recalibration of expectations. But that's really not so hard to accomplish. For starters, give an Ayre D1 DVD/CD player and one of their power amps and preamps a listen, hooked up to first class "simple" speaker, like an Avalon Eclipse or Acurus, or some of the better Joeseph Audio models, or the new Quad ESL's.

                            Any extension in bass, trebel, and dynamics (loud to soft, microdynamics in the presence of louder signals), through a larger, more complex speaker system, should not be at the expense of the fundamental naturalness and purity a system like this can deliver.

                            Not a simple goal, is it?


                            4) Well, as I point out above, I think combining a dipole midbass (75 Hz to 1 kHz) with a conventional (well designed, well placed) subwoofer system may work well; I think it's worth investigating. The ultimate would probably be a well designed IB sub, but in many rooms that's not possible, so more conventional solutions would have to be tried.

                            5) That's a really tough one to answer. That's why we're working on several projects, with varied approaches. Tom and I are like painters or other artists; we don't stop painting and creating just because we've got one wall full... which is a little crazy, but it's the journey as much as the end result. The latter is always evolving. We don't do DIY because it's cheap, but because it's a voyage of discovery.

                            As an example, we've built subs using the Stryke HE 15 driver. We built both a Stryke "cube", with the multiple passive radiators, and we built a carefully conceived ported enclosure, based on the concept of the Aerial SW12. There was a lot of controversy, surprisingly, surrounding our publication of the results and our feelings about the relative merits of each; but virtually NONE of the people voicing the "strongest" opinions had taken the time to build and evaluate both themselves.... :roll:

                            In other words, you won't really know what is down that road until you travel it yourself. Most times the journey changes you, your beliefs, and your perceptions, and that's a good thing.

                            Keep us informed, Victor of what you decide to do, and how it turns out!

                            Best regards,

                            Jon




                            Earth First!
                            _______________________________
                            We'll screw up the other planets later....
                            the AudioWorx
                            Natalie P
                            M8ta
                            Modula Neo DCC
                            Modula MT XE
                            Modula Xtreme
                            Isiris
                            Wavecor Ardent

                            SMJ
                            Minerva Monitor
                            Calliope
                            Ardent D

                            In Development...
                            Isiris Mk II updates- in final test stage!
                            Obi-Wan
                            Saint-Saëns Symphonique/AKA SMJ-40
                            Modula PWB
                            Calliope CC Supreme
                            Natalie P Ultra
                            Natalie P Supreme
                            Janus BP1 Sub


                            Resistance is not futile, it is Volts divided by Amperes...
                            Just ask Mr. Ohm....

                            Comment

                            • Victor
                              Senior Member
                              • Apr 2002
                              • 338

                              #15
                              Thomas,

                              Your remarks are always very helpful. I looked at the links for the proposed planar tweeters and it seems that the Raven is perhaps the best choice. Naturally you would never know until you install one and test it. I tell you, I’ll be extremely interested to see how the integration of the super-tweeter turns out. In my proposed baffle design the front area is rather flat, so the addition of another small driver is very possible. And, who knows, if yours (and Jon’s) efforts are successful, perhaps I may follow in your footsteps.

                              Regards,
                              Victor

                              Comment

                              • Victor
                                Senior Member
                                • Apr 2002
                                • 338

                                #16
                                Jon,

                                I certainly appreciate your involvement in this discussion. It must be difficult when you are on the road. Naturally if you have the time I’ll look forward to any more insights you may have. As for my concerns with this project, well, the proposed W-frame woofer is only 14 inches tall and my proposed midrange array will most likely be about 50 inches in length. So the overall height of the 2 woofer and midrange array contraption stacked on top each other will be about 78 inches, - the same height as the RD-75 baffle. Interestingly enough the ceilings in my basement are only about 90 inches. However the place where those towers will most like end up has ventilation ducts so the ceilings are only 80 inches there. The towers will just fit in.

                                Yes, my proposed system does sound like the monstrous Geneses and that scares me somewhat. I also thought about integrating the midrange and the RD-75 together on a single baffle, but then decided against it after I read John Whittaker's findings. The baffle must resemble the surface that would ‘promote’ a laminar flow so that the diffraction artifact are minimized. So I felt that it would be easier to accomplish with separate baffles.

                                I agree, Jon, who knows how it will sound! Also the minimalist approach is a fine philosophy, and I do follow it when it comes to the electronics I have built over the years. When I just started building audio electronics I was mostly using op-amps like AD711, but then after sometime I graduated to OP275. I then discovered AD797 and AD811. In my latest outboard D/A converter I use a lot of Analog Devices parts. I must say that the Sony player you have is still perhaps the best player around, with those pulse-current DACs. I use a very heavily modified Denon 2650 CD Player. It offers 16x oversampling and push-pull DAC configuration per channel. The D/A chips are venerable but still excellent AD1862K, 20-bit ladder types. Since I got this payer in the early ‘90s, I completely re-done its analog sections and the power supply. This player now sports fully balanced signal processing chain with discrete passive anti-imagine filters (4 of them!) and fully discrete BJT based buffers. Pretty crazy! The power supply is done with Sultzer type discrete regulators (1 mV of noise at the most). I use AD797 with power pass transistors. Overall, my CD Player shows near 18-bit performance at all frequencies of interest. Anyway, my latest pre-amps no longer use IC op-amps as I developed my own discrete op-amp. I use low-noise BJTs, the BC413 and BC415 in a fully complementary topology.

                                I, however, never found the LNFB circuits to be better then a more standard approach. I think it is a mater of choice as the THD numbers really get worst with LNFB, but not significantly to worry about it. The power amps you describe are pretty impressive especially the 125W class A monoblocks. They must be quite large. I have a limited experience with class A designs. I did build a Pass A75 once. The monster ended up to be about 50 pounds per channel. I since turned my attention to class A-B amplifiers, once I discovered Bryston topological ideas and the Ampzilla amplifier. I believe that A-B class is the engineering compromise that best suits my understanding of what the power amp should be. I since built a dozen monoblocs with power ranging from 400 Watts down to 150 Watts (8-Ohms numbers). All my amps are happy with loads down to 2 Ohms, with near perfect power doubling. I use the variation of the ingenious Bryston output stage. This stage is responsible for the extremely low THD at the high power levels.

                                All in all, I must say that the projects that you and Thomas are working on or have completed (I did spent sometime looking the Klone site) are truly impressive.

                                Going back to my project, I sense from yours and Thomas’s responses that my proposed architecture consisting of a stand alone baffle with RD-75 in it, the two W-frame Linkwitz woofers with total of four 12-inch Peerless 830500 drivers and a four driver midrange (10-inch SS 25W/8565) array in a Phoenix baffle may work after all. This is encouraging. So I think I will proceed with putting together the RD-75 baffle to begin with.

                                Yes, the system is very complex, but the woodwork is manageable and the crossover electronics is certainly doable. I think it is the integration of this system with the existing room that will be ultimately the ‘crux’ of the matter. I have looked at the IB possibility in my room, and I may end up actually doing it in the end. You see the open baffle system gets to be very inefficient at lower frequencies. Linkwitz’s “F-equal” analysis shows that at around 20-30 Hz you need 4 drivers in an open baffle for 1 driver in a sealed box. So I may after all run out of “oomph” around 20-30 Hz or so. Well, if that happens I will not hesitate to install an IB mono array (or the manifold), cross it at 40 Hz while letting the dipole woofer play full range.

                                As for the cost of my system, it looks like I’ll be spending at around $3000 in drivers for a stereo system (2 front channels). No wonder comparable commercially available designs, you mention the big Genesis, cost the as much as ZR6 Corvette.

                                One other thing I would like to get your opinion on. It is regarding my surround channels. Here is the quote from my original post:

                                “For the two surround channels, I am thinking of pairing the H-frame woofer, as shown in http://www.linkwitzlab.com/proto.htm#PW1 crossed at 150-Hz with RD-75 ribbon housed in the baffle as discussed above. The RD-75 will work down to 150 Hz and since surround channels do not do all that much work, I hope this combination will work out.”

                                Certainly it can wait, but I would like to ‘hear’ you thoughts on this along with whatever other opinions you may have on my proposed system.

                                Regards,
                                Victor

                                Comment

                                • JonMarsh
                                  Mad Max Moderator
                                  • Aug 2000
                                  • 15259

                                  #17
                                  Hi Victor,

                                  You know, I was relatively cool to my friend's first NLFB designs; personally, I think there were some issues with the amount of current drive he had available for the IR MOSFETs he was using at that time. That's why I talked him into the Magnatec FETs from UK for the V1, and trying extended beta bipolars for the 5 and 6 series amps.

                                  If you haven't heard an Ayre, or a Theta or BAT NLFB amp, you owe it to yourself to do so someday. There are things going on that static THD doesn't begin to describe, nor does the conventional thinking about TIM. Try it, you'll like it. But most of my friends who have heard a V-5 or V-6 have now either bought one, or are planning on it as soon as they can scrape up the funds.

                                  The Palladium's are an example of Aragon's cleverness to expand their product line while re-cycling a lot of the same mechanicals. Same heatsink and basic power boards as a stereo BB, with a somewhat beefed up power supply, but all the hardware is just devoted to a full balanced differential amp. Too bad they don't sound better than they do.... :? ThomasW and I have a lot of Aragon amps we've bought at nice prices used; I've been working on a new NLFB front end for them, and some power supply mods. I've got a pair of Palladiums, an 8008X3, an 8002, and an 8008ST to retrofit. ThomasW has more, if you can believe that... :LOL:

                                  BTW, my preferred CD player is NOT the Sony SCD777ES, though I agree with you that it's one of the better units, particularly in the price range it listed for, as well as what it sold for before being discontinued. It's sat on the bottom of my equpiment rack, unconnected, for quite a few months now. I've had an XA7ES for years that I'm getting ready to experiment a bit with now - just had the laser assembly replaced. Don't want to experiment yet with it! (. But my reference is the MP DAC- I'm building two more for friends, now. Didn't want to get a cottage industry going, though- I seem to always back into these things.

                                  While you're listening to different opamps, give the AD825 a listen, too. I've built a discrete version of the 797 topology; it's pretty cool, measures great, but it just doesn't sound quite right, or as good as the NLFB stuff I've been working on a couple of years. Wish I could tell you why.... ops:

                                  Are you intending to run your surrounds full range? If you're rolling off the bottom, to handle deep bass through LFE, I can't help but wondering if going with a panel more like your midbass, with some EQ for the low end (like Linkwitz did with the Dvorak or Bethoven when run without the woofers) mightn't give you more satisfying results. I'm still skeptical about a 150 Hz crossover for the RD75's. But, let's say that's OK- then you need a woofer clean to 300 or so.

                                  I'm looking at 11" woofers with 16 mm Xmaxwhich due to voice coil inductance have a 6 dB/octave roll off above 80 Hz, and a first cone mode around 800 Hz. I haven't done testing on them yet- just got them before my trip. A lot of questions to answer still.


                                  The Raven looks to be the best of the available ribbon tweeters- the only thing "against" it is the price- ThomasW and I talked on the phone yesterday about that, don't know that even he would be willing to have a two foot high ribbon stack (to get a good vertical listening window) for which you'd pay almost $5k just for the tweeters for both sides.

                                  The "Arvo Parrot" - no, let's call it by the "real name", the "Arvo Pärt", will cost between $1020 and $1540 for a stereo pair, depending on if I decide to use 4 instead of 2 woofers per side. (doesn't inlcude bass below 75, unless with EQ). Four will work "for sure", as Bullwinkle J. Moose would say, but then they'd be as big and heavy as Thomas's Whisper Klones, and I was really shooting for something more like the size of the Vivaldi or Dvorak main panels. Hopefully I'll know in a few weeks which way I'll have to go.

                                  What kind of timetable do you have for your project? Keep us informed of how it works out!


                                  Best regards,

                                  Jon




                                  Earth First!
                                  _______________________________
                                  We'll screw up the other planets later....
                                  the AudioWorx
                                  Natalie P
                                  M8ta
                                  Modula Neo DCC
                                  Modula MT XE
                                  Modula Xtreme
                                  Isiris
                                  Wavecor Ardent

                                  SMJ
                                  Minerva Monitor
                                  Calliope
                                  Ardent D

                                  In Development...
                                  Isiris Mk II updates- in final test stage!
                                  Obi-Wan
                                  Saint-Saëns Symphonique/AKA SMJ-40
                                  Modula PWB
                                  Calliope CC Supreme
                                  Natalie P Ultra
                                  Natalie P Supreme
                                  Janus BP1 Sub


                                  Resistance is not futile, it is Volts divided by Amperes...
                                  Just ask Mr. Ohm....

                                  Comment

                                  • Victor
                                    Senior Member
                                    • Apr 2002
                                    • 338

                                    #18
                                    Jon,

                                    You are making a very persuasive case for the LNFB circuits. Perhaps I should look at them again. It would be interesting for me to find an article describing a project where this concept is used. I must confess that I have never had a chance to see an Ayre, or a Theta or BAT product. I’ll make a point of finding one. However, if I remember correctly, BAT uses tubes, and I would have conceptual problems looking at tube-based product, but I’ll try anyway. I am familiar somewhat with Aragon products. The last time I saw one, I remember thinking that it can drive anything, - nice amps indeed.

                                    So you build a discrete AD797 op-amp, - that is very impressive! I remember looking at the published topology of 797 and it would seem to me that the architecture relies on the transistor matching to a greatest degree. Honestly, I cannot imagine building this thing out of discrete components, as I do not see how the Johnson noise figures of 797 can ever be attained with discrete transistors. My op-amps are a lot simpler then that. Also next time I am shopping for IC’s I’ll get some AD825 too.

                                    You know, your suggestion regarding my surrounds channels is good one. I think that you are right and the H-frame woofer will have difficulties pulling 150 Hz and the RD-75 will definitely struggle at such low frequency. I think I can build the same midrange array for the surround channels as the one used in the front, but I can EQ them a little different. I think even with a narrow 12.5-inch baffle the response of SS 25W/8565 10-inch drivers can be extended down to about 40-45 Hz and that will be sufficient for the rear channels, providing that I will have a 0.1 bass channel anyway. In fact, since the rears do not have to work all that hard, I may end up with less midrange drivers. I can start with two drivers and see how it goes, as I can always add more drivers later. The front channels will need 4 midrange drivers for sure, but the rears are another story in my opinion.

                                    The timetable for my project is, well, flexible. But I hope to complete the front channels this summer. I have the samples of the half rounds en-route at this point and trapezoidal baffle sections will be cut in the next 2 weeks. I am hoping to complete and set-up the RD-75 baffles by the end of June, as I’ll need at least a month to finish the electronics crossover for them. I still have to make a decision regarding how will I make this crossover. Will it be mono or all channels together with one big power supply? Well, mono is better but more work for a questionable return, we will see…Perhaps I’ll make one circuit board with all electronics for one channel. Anyway by the end of summer the towers will ‘sing’.

                                    Regards,
                                    Victor

                                    Comment

                                    • JonMarsh
                                      Mad Max Moderator
                                      • Aug 2000
                                      • 15259

                                      #19
                                      So, you're going to have a really busy summmer, too!


                                      The big BAT (I think the model number is VK-6200) is solid state. It's an impressive beast. Some of their other stuff is just weird enough (Russian radar tubes and all) to be really interesting, though I haven't built anything with tubes since the mid seventies. I'm an old EL34 and 6550 kind of guy myself. Well, except since the late seventies, when I've been more an NPE5565 and 2n4401/4403 MAT-01/MAT02 kind of guy, and these days more a 2SC3298/2SA1302 kind of guy. :LOL:

                                      A lot of the better single ended differential opamp circuits do require well matched pairs; my old trick is combining relatively low voltage monolithic dual transistors with hand matched high voltage transistors in cascode. I've had a tendency in the past to get a little carried away; these days I'm trying to simplify.

                                      The last power amp I designed and built using feedback had paralleled NPD5565 for the front end differential pair, cascoded with matched MPSA42, loaded by an differential active current source load using a MAT-02 differential pair, which were buffered with a MAT-02 pair (collector's biased from a regulated reference 6 volts below the regulated driver rail, driving a matched set of 2N2219, which were cascoded with matched MJE350's. The "back side" of the 350 differential pair is connected to a an additional ground biased MJE350 cascode, and the output of that drives a buffered cascoded wilson current mirror using MJE340 to make a pushpull connection with the "front side" of the upper MJE350. Both the upper and lower cascode outputs have anti-saturation clamps, to avoid sticking during clipping.

                                      The output used a compound output stage, with a bipolar class A buffer driving a MOSFET classA output referred follower, where the drain output drives to rail referred MOSFETs for the outputs, using British made Magnetec lateral MOSFET transistors. None of that "transconductance droop" that Douglas Self likes to talk about with that output stage...

                                      Very clever, (or so I thought at the time), but the newer stuff sounds better. Go figure....


                                      Best regards,

                                      Jon




                                      Earth First!
                                      _______________________________
                                      We'll screw up the other planets later....
                                      the AudioWorx
                                      Natalie P
                                      M8ta
                                      Modula Neo DCC
                                      Modula MT XE
                                      Modula Xtreme
                                      Isiris
                                      Wavecor Ardent

                                      SMJ
                                      Minerva Monitor
                                      Calliope
                                      Ardent D

                                      In Development...
                                      Isiris Mk II updates- in final test stage!
                                      Obi-Wan
                                      Saint-Saëns Symphonique/AKA SMJ-40
                                      Modula PWB
                                      Calliope CC Supreme
                                      Natalie P Ultra
                                      Natalie P Supreme
                                      Janus BP1 Sub


                                      Resistance is not futile, it is Volts divided by Amperes...
                                      Just ask Mr. Ohm....

                                      Comment

                                      • Victor
                                        Senior Member
                                        • Apr 2002
                                        • 338

                                        #20
                                        Jon,

                                        Very interesting! I must congratulate you on having such a profound experience with amplifiers. My own experimentation started in a way like yours. When I just started building power amps for the first time, in 1985, I was in a graduate school in Boston. I tell you, the complexity of my first endeavor was truly unbelievable. I had a regular 3-stage deal with fully differential front end, gain stage and the output buffer. All bipolar! I had cascodes and active loads everywhere with a Wilson-type current mirrors. My power supply was around 100v both ways with a full regulation implemented by bypassed 3-terminal regulators with the pre-regulators to boot. Still I only got 300 Watts out of the entire thing and it weighed a tone. The THD numbers were good but not fantastic.

                                        Over the years I experimented with transistor arrays including MAT-04 and some Motorola chips. I also tried FETS pretty much everywhere. In the end I came to a realization that active loads are not all that beneficial and simple resistors do just as well. I still like the idea of cascodes though. My fundamental design philosophy got formed around the thought that you must start with an overall very linear system before any kind of feedback is used. So then the feedback is simply an added benefit and not a panacea from all evil.

                                        With this in mind I dispensed with active loads and differential set-ups. I found that the noise they added is more detrimental then the linearity they can bring. So by putting a lot of effort into matching parts I can achieve a very comparable linearity with simple resistors. Also by reworking the feedback structure so that there is no electrolytic capacitor (the one that keeps the DC gain at unity) present in the signal path, I achieve another little bit in non-linearity reduction. It is a small gain, but it is measurable. I endeavor to use a smaller value resistor where it is possible (and important) to minimize the noise further.

                                        I decided to put the most effort into the output stage. After all, it is the output buffer that is the most responsible for the distortion. Early on I found that the BJT’s are actually more linear then FETs. It has to do with the gate capacitance that is impossible to linearise. Although the collector-base capacitance is not that easy to deal either without cascodes. Anyway, I discovered the Bryston approach, variation of which I now use in my amps. The remarkable THD behavior of Bryston amps is due singularly to the output stage.

                                        Like Bryston, in my amplifiers the output stage actually has gain. I know it goes against the textbook, but as long as this gain is low (5 in my amps) then the stability can be handled. There is a local feedback loop around the output stage and then another global one. The global feedback is not ‘huge’, it is actually can be deemed to be a ‘low feedback design’. Interestingly this output stage feedback keeps the output transistors in the linear region even at high power levels. You see, when the high power is demanded, the output transistor collector-emitter voltage gets very close to the saturation ‘knee’ and thus the THD suffers as the global feedback cannot really correct for this, but the local feedback can do it to a greater degree. This results in the THD of less then 0.01% even at max power.

                                        So that is where I am today as far as the power amps go. I have been looking into two other very promising implementations, the current feedback and the sliding rail approach. I did build both types over the past few years. However, lately with my interest in speakers I have not done much for the amps. The current feedback design is particularly intriguing as it allows for a huge slew rate. I had an amp swinging at 200 Volts per microsecond at 150 Watt with 0.05 % THD! The sliding rail is perhaps the most elegant approach. It is a lot like what Carver did with his Lightstar, but it uses a linear amplifier for the power rails. Also Y. Didden published a great design in Audio Amateur in 1983 that worked on the same principle. Anyway I hope that once my speakers are done I will come back to this extremely intoxicating pursuit.

                                        Funny, it would seem that my proposed speaker system is, complexity-wise, a lot like my first amp. Go figure…

                                        Regards,
                                        Victor

                                        Comment

                                        • JonMarsh
                                          Mad Max Moderator
                                          • Aug 2000
                                          • 15259

                                          #21
                                          Hi Victor,

                                          Man, am I glad this two weeks of travel is over!


                                          It sounds like we've been down some very similar lanes with regards to some of the amplifier design concepts and issues. The compound FET output stage I used is fairly similar to the Bryston output stage compound, except that there was only the driver set directly referred to the outputs; more like Walker's compound; but that was a dual chip 250 watt transistor, which could handle a 100 watt output at eight ohms by itself. Like the Bryston circuit, it was essentially a combination of a follower and common emitter/source stage, but not using a fully complementary (in number) set of transistors in both follower and common emitter mode. (Do you know what output transistors Bryston is using these days?) The Magnetec MOSFETs had a fairly low input capacitance, and much more linear Ciss than the the IR types Pass and others use; and have almost negligible feedback capacitance (CGD) over most of the drain to source voltage range- that, plus the local loop in the output stage and the number of transistors kept the open loop THD and IM quite low. I'm really not clear on why people keep trying to use vertical MOSFETs in audio (I'm a power semi apps engineer to pay the bills); they're not at all suited to the application. They're designed to be switches, not linear amps.

                                          These days I'm pretty much just using the extended beta bipolars, mostly the On-Semi copies of the Toshiba and Sanken parts; their linearity is quite good in a slightly modified classic T-circuit. The trick for open loop operation was figuring out how to stabilize them for capacitive loads without using an output decoupling inductor or base input resistors. Remember grid neutralization of vacuum tube RF amplifiers? Works fine with FETs and bipolars, too. :B

                                          I think I go through swings of complexity and simplification myself at times; both for speakers and electronics; I think your project will be very interesting, and I hope you'll keep us informed of your progress, Victor! Too bad I don't get out to Ontario very often!

                                          Best regards,

                                          Jon




                                          Earth First!
                                          _______________________________
                                          We'll screw up the other planets later....
                                          the AudioWorx
                                          Natalie P
                                          M8ta
                                          Modula Neo DCC
                                          Modula MT XE
                                          Modula Xtreme
                                          Isiris
                                          Wavecor Ardent

                                          SMJ
                                          Minerva Monitor
                                          Calliope
                                          Ardent D

                                          In Development...
                                          Isiris Mk II updates- in final test stage!
                                          Obi-Wan
                                          Saint-Saëns Symphonique/AKA SMJ-40
                                          Modula PWB
                                          Calliope CC Supreme
                                          Natalie P Ultra
                                          Natalie P Supreme
                                          Janus BP1 Sub


                                          Resistance is not futile, it is Volts divided by Amperes...
                                          Just ask Mr. Ohm....

                                          Comment

                                          Working...
                                          😀
                                          😂
                                          🥰
                                          😘
                                          🤢
                                          😎
                                          😞
                                          😡
                                          👍
                                          👎
                                          Searching...Please wait.
                                          An unexpected error was returned: 'Your submission could not be processed because you have logged in since the previous page was loaded.

                                          Please push the back button and reload the previous window.'
                                          An unexpected error was returned: 'Your submission could not be processed because the token has expired.

                                          Please push the back button and reload the previous window.'
                                          An internal error has occurred and the module cannot be displayed.
                                          There are no results that meet this criteria.
                                          Search Result for "|||"