The Plywood Buddha: a dipole testbed

Discussion in 'Speakers' started by philipmorgan, Feb 5, 2020.

  1. philipmorgan

    philipmorgan Member of the month

    Friend BWC
    Joined:
    Sep 28, 2015
    Likes Received:
    2,770
    Dislikes Received:
    1
    Trophy Points:
    93
    Location:
    In the wind, so to speak
    Home Page:
    Since the dawn of my speaker experimentation time, I've never had a single spare *inch* of space to spare in terms of depth. I always needed speakers I could place all the way against the rear wall because I lived in a tiny ass house in expensive ass California, and WAF was always a constraint as well.

    We've since moved to Taos, are working on building here, and are renting a place where my office is HUGE relative to what I'm used to. It's a 15 x 23' room that used to be a photography studio. Lots of light from a bank of north-facing windows, and finally... enough room to experiment with dipole speakers, which I've wanted to try for some time.

    [​IMG]

    Some people are detail whores. Detail or plankton or whatever gets their mojo workin'. For me, soundstage depth is the equivalent property. I also really dig big, muscular Bruce Lee bass. People have raved about the unboxyness of dipole bass, and sometimes said diples with sufficient back space do soundstage well, so I figured I'd give it a go.

    I don't have the time, tooling, or desire to invest right now in fancy woodworking, so I grabbed a few 2x4' sheets of cheap fir plywood, some 10" wide pine board, a cheap circular saw (too cheap, it turned out), and over a weekend put together a dipole testbed I'm calling the Plywood Buddha because it's short and squat, kind of like those laughing Buddha statues you see at Asian restaurants.

    [​IMG]

    The woofers are 18" Beyma SM-118N I had from a previous project centered in 2x2' baffles with 2x2' "wings" on either side. The midranges are Fostex FE108EZ, and the tweeters are Peerless BC25SC06-04 soft dome tweeters I picked up for cheap from Madisound. Mid and tweet are mounted in a 10" pine baffle, and that baffle is mounted atop the bass baffle wings using some shelf brackets. Everything's screwed together but not glued. This lets me easily change things around, and attempt to time-align the bass/mid sections, though it means the mid/high drivers are pretty low to the ground. Which is fine because right now, I'm just trying different things out and listening seated cross-legged on the ground at a midfield distance.

    [​IMG]

    [​IMG]

    [​IMG]

    To keep things super cheap and super easy/fast to experiment with, I'm using this MiniDSP DDRC24 I had lying around for both DAC and crossover and (later, if necessary) EQ/DRC duties. It feeds lows to a Crown X1000 and the rest of the range to a JLH69. Later I might try a passive XO once I settle on design parameters I like.

    As I mess with these, I hope I learn things that are generally useful.

    An underrated attribute of floorstanding speakers for me, at least early on, is ease of moving them around to try different placement options. I've built a few heavyass 3/4" baltic birch cabinets with no handles and spike feet, and they are a massive pain to move around. With the Plywood Buddhas, I just reach down and grab the woofer basket, which is both confidence-inspiring in size and cast aluminum in composition and affords this really useful handle for picking up and moving the speakers.

    This whole thing needed to be done on the mega-cheap, since we're trying to start a house build in 2020 and business still has some months where cashflow is a bit wavy. Thus the mega cheap circular saw (lesson: never buy the cheapest circular saw if sawing straight lines or having the blade be aligned with the saw base matters to you)

    I prolly shouldn't be spending any money at all on this, but I get bored if I'm not doing _something_ exploratory with audio.

    Measurements were made with the dayton imm6 mic. I got it during a particularly broke period after my umik1 quit working after I opened it up to change the DIP switches, and acoustically the imm6 is probably fine (probably uses the same/similar cheap capsule as more expensive measurement mics anyway) but you need like a fucking roach clip or something to hold it while you measure, and lacking any actual roach clips, I gafffer taped it to a tripod.
     
    uncola, shanchandi, Azteca and 31 others like this.
  2. philipmorgan

    philipmorgan Member of the month

    Friend BWC
    Joined:
    Sep 28, 2015
    Likes Received:
    2,770
    Dislikes Received:
    1
    Trophy Points:
    93
    Location:
    In the wind, so to speak
    Home Page:
    first experiment.

    Back wall distance matters, or so I hear, when it comes to dipoles and their bass response. Let's do some measurements and see!

    Again, the room is 15' x 23' or 17' x 23' depending on where you measure. The Plywood Buddhas are set up with the speaker cones facing the short wall.

    I measured a frequency sweep with REW with the speakers at progressively greater distances from the wall that's behind the speakers. I went in 1-foot increments and stopped at 8' away from the wall behind the speakerrs. Because of the rear wings on the speaker baffle, the closest I could get to that wall was 2 feet.

    I measured with one speaker active and placed roughly where it would be for listening (not centered, but placed off to the side the way it would be if listening to music) and the measurement mic roughly in my usual listening position.

    Here's all the measurements. It's an unreadable mess, but it'll give you a sense of the variation that speaker placement causes:

    [​IMG]

    And here are the extremes, to create a feeling of dipole-shock-and-awe in my readers:

    [​IMG]

    I picked the "extremes" here mostly based on overall frequency response smoothness. There are greater extremes in the loudness of certain bass frequencies at certain localized points along the curve. But overall, I'm probably going to have to use EQ to make the bass quantity not suck on these, so I'd rather need to use less EQ than more, thus I'm favoring overall FR smoothness for now.

    And finally, here's the REW data file if you wanna inspect it yourself: https://pmc-dropshare.s3-us-west-1....s-back-distance-measurements-for-sharing.mdat

    What I learned

    I had assumed that dipoles require massive distance to the back wall to work well. If what I'm looking for is an overall smooth-ish FR, then **that's not true here**. There is more overall FR smoothness with the Plywood Buddhas at 3 feet from the back wall than at 7 or 8 feet from the back wall.

    Other qualities, like soundstage depth or ultimate bass quality, might not follow this same pattern. It might be that soundstage depth benefits and FR smoothness suffers with greater back distance. I guess that's my next experiment.

    Are there any tracks that y'all like for evaluating soundstage depth or other soundstage-ey qualities?
     
    yunie_, Philimon, captkirk and 3 others like this.
  3. Biodegraded

    Biodegraded Friend

    Friend
    Joined:
    May 28, 2017
    Likes Received:
    2,938
    Dislikes Received:
    0
    Trophy Points:
    93
    Location:
    Vancouver BC
    I presume your sweeps start at least as low as 20 Hz; if so, are you surprised not to get more LF response from those big woofers?

    Quarter-wavelength cancellation distances between 2' and 8' should be between 140 and 35 Hz, but I don't really see a correlation in the measurements, so I guess reflections and cancellations from other boundaries are affecting things too. Have you plugged the distances into REW's room simulator to see where the modes and holes are?

    (Can't open your file - I'm using the current release (v5.19) for Windows).
     
  4. mitochondrium

    mitochondrium Friend

    Friend
    Joined:
    Sep 1, 2017
    Likes Received:
    475
    Dislikes Received:
    0
    Trophy Points:
    63
    Location:
    A Cell
    Thanks for sharing, being a speaker only guy myself, this is interesting. Ruled out open baffle because of one corner of my room being windows an the other one concrete on the basis that open baffle is more problematic than boxes in that case. Seeing that the measurements show so few difference, I wonder whether the wings play a role in this, alas I am not knowledgeable enough to be sure about that.
     
  5. sphinxvc

    sphinxvc Gear Master (retired)

    Staff Member Friend BWC
    Joined:
    Sep 28, 2015
    Likes Received:
    2,778
    Dislikes Received:
    5
    Trophy Points:
    113
  6. philipmorgan

    philipmorgan Member of the month

    Friend BWC
    Joined:
    Sep 28, 2015
    Likes Received:
    2,770
    Dislikes Received:
    1
    Trophy Points:
    93
    Location:
    In the wind, so to speak
    Home Page:
    Super surprised. The bass response at this point is disappointing. But! There's lots of room for experimentation. Thanks for the REW sim tip. I'll give that a try soon.

    EQ does improve things somewhat. This is with a massive low shelf:

    [​IMG]
     
    Psalmanazar and Ntbm3 like this.
  7. philipmorgan

    philipmorgan Member of the month

    Friend BWC
    Joined:
    Sep 28, 2015
    Likes Received:
    2,770
    Dislikes Received:
    1
    Trophy Points:
    93
    Location:
    In the wind, so to speak
    Home Page:
    I guess the latter? I'm still learning here, so not sure why I'd want it to be a pure dipole response?

    Thanks for the further reading link!
     
  8. yotacowboy

    yotacowboy McRibs Kind of Guy

    Friend
    Joined:
    Feb 23, 2016
    Likes Received:
    2,478
    Dislikes Received:
    10
    Trophy Points:
    93
    Location:
    NOVA
    Home Page:
    Cowboy Junkies: Trinity Sessions

    I think my high water mark for hearing what image depth can sound like on this album was hearing it on an MBL setup. Diffuse, yes, but shockingly convincing.

    I'm keen to see where this leads!
     
    Aeron likes this.
  9. ergopower

    ergopower Almost "Made"

    Contributor
    Joined:
    Mar 21, 2018
    Likes Received:
    235
    Dislikes Received:
    8
    Trophy Points:
    43
    Location:
    South Central PA
    Echoing others, this is a cool project. Question for you - why did you put the wings on the baffle so that they project front & rear? I know low frequency = long wavelengths, etc.; but surely there must be some diffraction effects? I wonder if there would be a difference in measured response if the wings were flush with the front? Center of gravity of the woofer is well behind the mounting flange because of the weight of the magnet, so it would still be very stable.
     
  10. peef

    peef Almost "Made"

    Joined:
    Oct 5, 2015
    Likes Received:
    111
    Dislikes Received:
    0
    Trophy Points:
    33
    This is a super cool project.

    I may have missed the specifics of your mic placement, but generally your best bet when measuring and EQing the bass is to place the mic right in front of the dust cap to swamp out any room effects. This might be why the bass is measuring lower than you'd expect. You'd then measure the mids/highs with the mic a few feet out at tweeter level (not necessarily in the listening position) and apply gating which discards any information after the first reflection point. There's some really great info on how to do this in Jeff Bagby's whitepaper.
     
    sphinxvc likes this.
  11. sphinxvc

    sphinxvc Gear Master (retired)

    Staff Member Friend BWC
    Joined:
    Sep 28, 2015
    Likes Received:
    2,778
    Dislikes Received:
    5
    Trophy Points:
    113
    Still learning myself, but from what I understand it comes down to exciting fewer room modes and a better polar response, which translates to better off-axis performance nearabouts the listening position, and also keeps the spectral signature* of reflected sounds as close as possible to direct sounds.

    *I've never been able to pin down a nice definition for spectral signature in audio, but I presume it's similar to FR
     
    zerodeefex likes this.
  12. rlow

    rlow A happy woofer

    Friend
    Joined:
    Jul 18, 2017
    Likes Received:
    3,412
    Dislikes Received:
    3
    Trophy Points:
    113
    Location:
    Canada
    Here’s a few suggestions:

    Eric Clapton, MTV Unplugged: Nobody Knows You

    Miles Davis, Kind of Blue: So What

    Norah Jones, Day Breaks Deluxe Edition (I get it on Tidal): Peace (Live in NYC)

    Bela Fleck, Flight of the Cosmic Hippo: Title track (check out the “Reprise” as well) - also a great bass test track

    Live recordings and Jazz tend to be the best for depth that I listen to. Don’t listen to a lot of classical, so can’t point much in that direction.
     
    Last edited: Feb 8, 2020
    yonson and StageOne like this.
  13. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    53,259
    Dislikes Received:
    89
    Trophy Points:
    113
    Location:
    Antarctica
    I missed this thread the other day. Very awesome. Thoughts below:
    • Nope, we don't need or want necessarily 7' of space behind the OBs. 2' from the back wall may be pushing it. 3' is plenty in my experience.
    • As far as purer figure 8 polar response, it's theoretically ideal because the bass (front wave and back wave) will cancel at the plane of the baffle. No bass this angle means nothing to bounce off the walls and ceiling at this plane. What sparse data I have would suggest that room modes are halved at best with OB. Any little thing helps because EQ for one position, say in the middle of the couch, can cause drastic issues at the end of the couch.
    • The wiggly up and down frequency response might be because of the OB effect (related to baffle width). You can break this up a bit by using asymmetry. This can be done by offsetting drivers left or right on the baffle, or extending a flap on one side at the back of the H frame of the woofer or attaching a flap one side on the back of the mids/high baffle. Don't have to use wood. Cardboard works well enough for test purposes. Uneven or jaggy patterns help.
    • If you want deeper bass, cover up the top of the H frame. There's that open gap at the top of the H frame where the baffle of the mids is offset behind the H frame. Use cardboard to cover up that gap. This may be why you are getting less bass than expected from an H frame.
    • If the above doesn't work, create a compartment behind the H-frame and stuff to varying degrees. This will create less of a pure figure 8 response and more of a cardioid polar response instead. Maybe more room excitation, but perhaps more and deeper bass. Tradeoffs.
    • Similarly, you can construct a basket behind the mids to contain stuffing. Note that damping the rear of the driver will have some effect on its transients response. However in this case with the Fostex's with low Qts, I don't think there will be any deleterious effects.
    • However, another method is just to leave the drivers be and install absorbing panels on the wall to catch the rear wave of the mids (and to a lesser extent the bass). The extra reflection from the back walls does give a grander sense of the concert hall effect, but it can also muddle the imaging precision.
     
    Last edited: Feb 7, 2020
  14. purr1n

    purr1n Finding his inner redneck

    Staff Member Friend BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    53,259
    Dislikes Received:
    89
    Trophy Points:
    113
    Location:
    Antarctica
    Looking at the non-EQ'd response in green. I just realized something. That's actually close to a theoretical model for an H frame. You may be able to get less of a bump at 200Hz an move up 100Hz and below by going with a U frame.

    Better yet, a super wide flat baffle (no U or H, or minimal U or H) will yield the flattest response bring up 100Hz and below and without the bump at 200Hz and dip at 350Hz. U and H frames are nice for space concerns, but better off widening the baffle (assuming all things equal, that is distance around the surfaces from front to back of the driver remains the same).

    Answer above regarding effects.
     
    Joshvar likes this.
  15. lehmanhill

    lehmanhill Almost "Made"

    Contributor
    Joined:
    May 3, 2018
    Likes Received:
    127
    Dislikes Received:
    0
    Trophy Points:
    43
    Location:
    Michigan US
    +1 for covering the top of the H frame. Also for cardboard and assemetry to smooth out response. I couldn't find a picture of my test baffle, but this picture from GR Research shows how much assemetry works in satellites that go down to about 200 Hz. Of course, the concept works with low bass, but gets complicated in the lowest octaves.

    https://www.dropbox.com/s/qsf9heeyf4fq8eg/wedgie.jpg?dl=0
     

Share This Page