Announcement

Collapse
No announcement yet.

A new design: Bob's B3 project

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • A new design: Bob's B3 project

    It's been a while since we had a new project on the board, so if you don't mind, I would like to show you the speakers I've been working on for the past two years. I call it simply the B3 project. I took these to DIY Atlanta unfinished in summer 2009 and they were well received, tying for first in the unlimited class. I'll explain the name later, but first, some pictures and a quick description of components.







    This is not a small speaker. As you can see, this is a four foot tall, dipole hybrid. As such, they have to be pulled away from the back wall so that the front is about two and a half feet out. It uses some fairly standard DIY drivers: The 8" Dayton RS-225S-8, the 6" Dayton RS-150S-8 and something a little different, the 1" Seas 27TBFC-GTV shielded metal dome. The total cost for drivers and crossover components is around $550 for the pair - not cheap, but not terribly expensive either.

    You can see from the first picture that I still have a CRT type TV, so I chose all shielded drivers. I probably didn't need to, given the distance to the TV, but I also wanted to make use of the flatter low end of the shielded tweeter vs. the standard 27TBFC. I also have a center channel in the works and I wanted to use the same driver. More on that later.

    Despite its somewhat standard driver lineup, there are some suttle refinements. First, the drivers are mounted so that their acoustic centers are very closely aligned. The wooofer is mounted on an extended front baffle and the tweeter is rear mounted using waveguide loading. These two details match their ACs with the mid on the main baffle.

    Something I did that not everyone will agree with is that I used first order electrical crossovers for both the woofer-mid at 250 hz and mid-tweeter crossover at 2000 hz. I also used notch filters as needed for the metal drivers. I'll discuss more about the crossover choices later, but I found that all drivers in phase sounded much better than anything else I tried. Here is the working crossover. I have some more testing and there are two small areas I want to improve, so don't consider that a finished design just yet, but I don't foresee any changes other than modifying values on the existing components. The minimum impedence is over 5 ohms, so any amplifier can drive them.



    And finally, here is the raw ungated, in room, 4 meter measured frequency response. I used ARTA and a calibrated microphone for this, and I also showed the response with and without my sub.



    That's it for now. In following posts, I'll walk you through my design decisions and construction details. For those of you that are a little gun shy about designing your own crossovers, I'll show you what each part in the crossover does and how I arrived at the values. It's not always textbook, but you will at least see why I did what I did.

    Enjoy
    -Bob



    The normal DIY disclaimer applies: This design is free for non-commercial use. Commercial use requires my permission.
    Attached Files
    -Bob

    The PEDS 2.1 mini system
    My A7 Project - another small desktop speaker
    The B3 Hybrid Dipole - thread incomplete and outdated

  • #2
    I'm interested to see more details. Looks like fun.
    diVine Sound - my DIY speaker designs at diVine Audio

    Comment


    • #3
      Looks good. I'm looking forward to the additional details to come. :T
      - Ryan

      CJD Ochocinco! ND140/BC25SC06 MTM & TM
      CJD Khanspires - A Dayton RS28/RS150/RS225 WMTMW
      CJD Khancenter - A Dayton RS28/RS150/RS180 WTMW Center

      Comment


      • #4
        It's getting late, so I'll just get started on the background tonight. Since college, I've been using a pair of Dahlquist DQ10s as my main speakers. Some of you guys will remember them. They were the first successful non-electrostatic dipole designs and to this day, they are hard to beat. Here is one next to my unfinished B3:



        Fast forward to this decade. I built a pair of box speakers for my daughter and another for my son. Granted, the speakers for my kids were not up to the level of the designs here, but they did get me back into the speaker building hobby. I had a tech school degree in electronics and I took an audio engineering class at Georgia Tech in the 80's, so I had the backgound. I decided that it was time to build something that at least matched my venerable old DQ10s and this project was born.

        I had to build something with at least a dipole midrange like the DQ10s, but for a first major project, I didn't want to invest the resources on a full dipole bass section. I considered it, but I decided that a dipole tweeter was not something I wanted to try for this project either. Maybe next time.

        So here is what I came up with as my criteria for this project.
        1. Bass. Sealed or ported. This was a tough choice. I had alway used sealed bass speakers, but I liked the way ported and PR bass went much deeper than the sealed boxes I had used. I also liked the way ported speakers seemed to hit harder but they did tend to sound boomy. But after building my kids speakers ported, I decided I could do ported without sounding boomy so ported it was. To keep box size under two cubic feet, I decided on an 8' driver. The ported models for the 10" drivers were up in the three cubic foot range and that was just too big.

        2. Midrange. If I was going to beat my DQ10s, the project had to be at least a three way so I could run the mid as a dipole. The DQ10s are a five way speaker, but only dipole from 400 to 1600. With better drivers from today, I thought I could extend this to cover more of the midrange with a single driver.

        After looking at drivers, I decided that 250 to 2000 looked reasonable for a single driver in dipole mode - a full octave more than the DQ10s. It was pushing it, but it was doable. So there was my first crossover point - 250 hz. This corresponds to the B just below middle C on the piano, which is also called B3, and the name was born. It also seemed a good choice for a first crossover point. The woofer would handle fundamentals below middle C and the mid would cover the three octaves from middle C up.

        3. Tweeter. Since I was only going to 2000 on the midrange, I had to choose a tweeter that could handle a moderately low crossover of 2000. After reading John Krutke's experiment with wave guide loading a tweeter, I decided that would do the trick. It also allowed me to push the tweeter back to the same acoustic center as the mid. Later Dan Neubec also ran some tests on small wave guides and that confirmed it. A cross at 2000 should work just fine and the wave guide would match the dispersion of the midrange pretty well.

        So there you have it: A three way crossed at 250 and 2000 with a dipole mid and waveguide tweeter.

        As a side note, just after I finalized this design, Paul Carmody published his Sunflower design which further confirmed that this was a reasonable project. Although they are similar, this is entirely my design as you will see and is in no way meant to be a copy of Paul's excellent designs.

        I know I've bored you with the background, but it might help you understand where I'm coming from.

        -Bob
        Attached Files
        Last edited by BOBinGA; 16 September 2010, 22:22 Thursday.
        -Bob

        The PEDS 2.1 mini system
        My A7 Project - another small desktop speaker
        The B3 Hybrid Dipole - thread incomplete and outdated

        Comment


        • #5
          Very cool project, Bob! I like the nice clean lines. :T This forum can really use some new designs, IMO! The way you supported the OB back panel with the curved central piece and circular cutouts is almost exactly the solution I arrived at back when I did my NaO mini's.

          How does the excursion model on the RS150 crossed at 250 in a OB dipole? I guess the Seas did fine in the shallow waveguide. That's good to know.

          What was your thinking behind the significant driver offset on the OB panel?
          Dan N.

          Comment


          • #6
            Thanks, Dan. I tried my best to make them presentable. I wish I had your woodworking skills. I learned woodworking just to make speakers. But my guess is that you had plenty of practice before you learned speakers.

            The excursion of the RS150 stays within xmax limits as long as I keep the high pass filter cap at 150 uF or less if I remember correctly. You can see from the latest version of my crossover that its up to about 175 now. I'm still working on that part. But I have a few more tweaks to try, so I might not need that much.

            The tweeter turned out great. You can see that it's almost dead flat from about 2500 on up to the rolloff over 12,000. I probably could have moved the tweeter a little closer to the mid, but I was trying to get part of the back brace to fit in between them. They are still just 6" apart center to center, which is just barely under one wavelength at the 2000 hz crossover, so it's OK.

            I also moved the centerline of the drivers over to the side after playing with the Edge program during development. This gave the flattest response in the program, but after I built it and measured, it didn't really work out that close to what the Edge said. I still had to make big adjustments in the crossover to compensate for the dipole peak I was hoping to minimize by offsetting the drivers.

            I'll go into more detail later. That is just one example of where I found theory not to match reality as closely as I hoped.

            Thanks,

            -Bob
            -Bob

            The PEDS 2.1 mini system
            My A7 Project - another small desktop speaker
            The B3 Hybrid Dipole - thread incomplete and outdated

            Comment


            • #7
              Love your work Bob. Aligning acoustic centres certainly makes crossover design a lot less of a headache.

              Comment


              • #8
                Very nice Bob, I wish I had your skills. :T ;x(
                Gavin

                BAMTM Build

                Comment


                • #9
                  Nice work Bob!
                  Seas Idunn
                  ZA5.2
                  ZA5.3CC
                  SB Acoustics 12" sub
                  Statement Monitors
                  CLD M5B

                  Comment


                  • #10
                    Thanks guys.

                    Let's get into some of the details. I'll start with the bass tonight.

                    I've gotten pretty good with WinISD, so that's what I used to get a bass alignment. I wanted to get down to 40 hz with the bass and the RS225S-8 modeled pretty well and met that goal. I also wanted to keep the box size blow 2 cu. ft. So here is what I came up with.



                    As I said before, I had some hesitation using a ported bass, so I purposely designed a little less than maximum bass extension. I was afraid if I maxed it out using a 2 cu. ft. box, it would begin to sound boomy. I think the slightly smaller 1.75 cu. ft. box worked out very well. It also allowed me to use this standard 2.5" X 8.5" port from PE without modification to tune it to ~27 hz and get an F3 of 32 hz:

                    http://www.parts-express.com/pe/show...number=260-478



                    Next I wanted to see how much output it was capable of. It looks like 50 watts would produce 104 db while just barely exceeding Xmax. That looked like enough for me. (I assumed a 20 hz high pass filter for woofer protection):





                    (Darn, I just realized that I forgot to add .25 ohms for the crossover inductor resistance. Oh well, I just checked and in this case it doesn't make enough difference to put together new jpg's - maybe half a db more at 35 hz.)

                    So far, so good. I'll show you the midrange modeling later, since it's now after midnight and I have to work in the morning.

                    -Bob
                    Attached Files
                    Last edited by BOBinGA; 16 September 2010, 22:27 Thursday.
                    -Bob

                    The PEDS 2.1 mini system
                    My A7 Project - another small desktop speaker
                    The B3 Hybrid Dipole - thread incomplete and outdated

                    Comment


                    • #11
                      To model the mids, I used a 5 cu. ft. closed box as a proxy for the open back. This results in a Q only .02 higher than the free air parameters. Open back speakers are not well modeled in WinISD, so I had to use something close. This was close enough for modeling SPL and excursion.

                      For the mid, the same 50 watt input yields the same 104 db SPL:



                      And doesn't exceed Xmax if you use a 135 hz or higher first order electrical high pass filter (about 150uF):



                      What this showed me was that only one woofer and one mid was needed, since I wasn't looking for head banging loud. 104db is still louder than I want to listen at for any extended period of time. Even adjusting for a 4 meter listening distance, it's at 92 db which is almost freight train loud.

                      -Bob
                      Attached Files
                      -Bob

                      The PEDS 2.1 mini system
                      My A7 Project - another small desktop speaker
                      The B3 Hybrid Dipole - thread incomplete and outdated

                      Comment


                      • #12
                        Arrrgh!!

                        I spent yesterday running some measurements. I ran the gated FR from 4 meters and it's much more bumpy than the ungated response. The trouble is, the ungated response is more like what I hear. But nonetheless, it showed me some areas to investigate in the crossover. I thought all I needed to do was lessen the dip at 2000 and smooth the peak and dip at 120-150, but the gated measurements showed a peak at 1000 that I wasn't expecting, since it wasn't in the ungated response. Measuring dipoles is not easy. I want to run a few more measurements and make a crossover change (or two) before I start into the crossover details. More to come...

                        -Bob
                        -Bob

                        The PEDS 2.1 mini system
                        My A7 Project - another small desktop speaker
                        The B3 Hybrid Dipole - thread incomplete and outdated

                        Comment

                        Working...
                        X