Seas A26 kit

Discussion in 'Speakers' started by murphythecat, Dec 12, 2018.

  1. briskly

    briskly Friend

    Pyrate
    Joined:
    Oct 3, 2015
    Likes Received:
    59
    Trophy Points:
    18
    Location:
    New York
    Geddes tried to define a horn as a device built for the objective of the presenting a target impedance load for the driver, which follows from Webster's horn equation. The waveguide theory would concern a horn's wavefront and expansion, discarding Webster's plane wave assumption, with the new goal of predicting the directivity. In the loosest sense, it could be taken as anything intended to shape the radiated wave, with the acoustic impedance as a secondary concern, as it has become in practice for all high-frequency horns.

    The cone is normally of a conical [horn] shape, which presumably is used to create a spherical wavefront. There are a variety of practical problems with using the cone driver as a waveguide, but it can provide a means to control the directivity of the high-frequency driver being loaded.
     
    Last edited: Jun 27, 2019
  2. skem

    skem Friend

    Pyrate
    Joined:
    Nov 1, 2017
    Likes Received:
    1,911
    Trophy Points:
    93
    Location:
    Charles River
    Looking at the Calpamos again, and recalling Linkwitz’s rule that unbraced area of 3/4” plywood cabinets should never have more than 4-sq-in of unbraced area: I wonder of the decay of the Calpamos could be significantly improved by simply more bracing or stiffer/thicker cabs. Most of Troels’ stuff is thicker, for example.
     
  3. Serious

    Serious Inquisitive Frequency Response Plot

    Pyrate BWC MZR
    Joined:
    Sep 28, 2015
    Likes Received:
    2,594
    Trophy Points:
    113
    Location:
    near Munich, Germany
    I'd say for getting perfect phase response from a first order crossover it's better to have more drivers, kind of like what Vandersteen is doing vs the newer Marten series. The drivers just don't need as large a bandwidth.
    I do agree that more crossover components and also more drivers, etc hurt the sound in a way.
    Many speaker designers like to position the drivers to avoid floor bounce cancellation in their operating range, but I agree that it often hurts imaging. The reflections might be more tonally even than with fewer drivers though, especially when it's about adding a small mid between a tweeter and a bigger midrange driver.

    I strongly disagree that 24db/oct phase distortion is inaudible. There's software that allows you to change the phase response without altering the amplitude that you can use to test this yourself though. IME even the excess phase from drivers with whizzer cones is usually audible and we're talking about a maximum of 60° of phase shift here. This also means that phase shift from 2nd order filters is well within audible limits, too. However don't take my words for it, I also believe cable risers and power cords to make a difference.

    As with everything in life it's about tradeoffs.
     
  4. briskly

    briskly Friend

    Pyrate
    Joined:
    Oct 3, 2015
    Likes Received:
    59
    Trophy Points:
    18
    Location:
    New York
    A 4th order LR is more than a 4th order allpass, it changes the combined spatial response based on the crossover order. In the high order limit, the change in speaker directivity at the transition is abrupt: one driver is omnidirectional and the other driver is beaming. Personally, inserting all pass filters did little to music unless they were inserted one channel at a time. You may test for yourself, and the option of nullifying phase shift with digital filters is an option.

    The additional driver would normally help reduce the spatial variation in the pattern. The area of the radiators can be made more similar, reducing the inherent directivity discontinuity between drivers. The overall pattern can be made wider so that we are not forced into using a giant narrow coverage horn to match the woofer. We ease balancing a soft cone that does not create objectionable resonances and possible geometric distortions as bending modes form on its surface. There are nulls that form between drivers spaced apart wide compared to the wavelength (>0.5 length), but the driver may be able to cross low enough to avert intense nulls at the bass-mid transition.
     
  5. skem

    skem Friend

    Pyrate
    Joined:
    Nov 1, 2017
    Likes Received:
    1,911
    Trophy Points:
    93
    Location:
    Charles River
    Can you point me to a reference that describes the physics of this? It’s not obvious to me how a shift in electrical signal amplitude and phase changes the radiation pattern of a driver.
     
  6. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    Different dispersion patterns / polar response of different sized transducers. A steep slope means less blending. Not an issue if drivers are crossed over at the appropriate points, but this would definitely be an issue if a 10" woofer was mated with a 2" woofer crossed over at 1.4kHz using a super steep slope.
     
  7. purr1n

    purr1n Desire for betterer is endless.

    Staff Member Pyrate BWC
    Joined:
    Sep 24, 2015
    Likes Received:
    89,778
    Trophy Points:
    113
    Location:
    Padre Island CC TX
    It's less this than "crossovers" typically include EQ / frequency response shaping circuits too. Most stuff would not sound ideal if textbook LR2, BUT18 crossovers were applied without response shaping. There are a lot of tricks to crossovers to shape the response without adding more parts.


    Yup, that's the huge advantage. We get to have drivers operating in their sweet spots. Also we get reduced IMD, which I think might be a pretty big deal. One day I will build a 4-way and see for myself. A woofer handling a 100Hz and 500Hz signal simultaneously will have IMD at 400Hz and 600Hz. If we have two drivers, a woofer and mid playing 100Hz and 500Hz respectively, then no IMD (in this specific instance).

    Depends on the design, one could argue that using a lot of drivers that operate in their sweet spot would allow the use of cheaper non-exotic better behaved parts which encourage the use of 6db slope crossovers, a la Vandersteen.

    As far as the 24db/octave approach, I don't think it's any more audible compared to 18db or 12db. That being said, I do think the "time / phase coherent" approach Vandersteen does work.

    Well, at least let me put it this way: the slope steepness affects the sound differently, and I don't necessarily think it's a phase issue. Working with these metal domes with fast decay (Seas magnesium woofers and aluminum tweeters), I found that the steeper slopes offered a more precise and focused sound whereas the shallower slopes provided better driver integration and cohesiveness. Again, trade-offs and preferences. As I've gotten older, I started to care less and less about super articulation and more on cohesiveness and midrange quality. This might be because I have better quality recordings and don't listen to Prodigy or Fatboy Slim so much anymore.


    On the other hand, a tall speaker with many drivers, but dedicated tweeters on top taking on 1.8kHz up (think Wilson) image like crazy, even up close.

    Steeper slopes, more parts. More parts = suck the life out of music. The Linkwitz designs I won't touch unless I'm using a digital front end with DSP crossovers. Otherwise that active crossover box is like opamp + opamp + opamp + opamp. Try chaining four CMOYs together and see how it sounds.
     
    Last edited: Jan 21, 2019
  8. Hrodulf

    Hrodulf Prohibited from acting as an MOT until year 2050

    Pyrate
    Joined:
    Sep 28, 2015
    Likes Received:
    4,276
    Trophy Points:
    113
    Here's a good primer text on crossover slopes affecting loudspeaker polar response. Do keep in mind that what you're hearing are the acoustic slopes. A 1st order electrical filter can result a steeper than 1st order acoustic function both in terms of frequency and phase response.

    In my mind crossover always comes second to the driver. If the driver you're using can do shallow slopes and you still comfortably hit SPL targets, then go for shallow slopes. With that said IME many hifi speakers utterly fail at high SPL's due to too shallow filters, because:
    1. Tweeters run out of excursion
    2. Driver breakups aren't sufficiently attenuated and become audible (extra horrid with hard cones)
    I wouldn't worry too much about the phase response of crossover filters unless one goes beyond 4th order. If you see perfect frequency domain blending at the crossover frequency, the time domain shouldn't give you trouble.

    Here's the group delay of a 3-way speaker with 4th order filters.

    [​IMG]

    I don't see anything wrong and the amplitude response agrees:

    [​IMG]

    The problem with time domain is that its audibility is not too well understood and researched. The consensus seems that people can hear phase distortion in controlled conditions with test signals, but with musical material it's nigh impossible. Some speaker builders try to keep group delay under one cycle just to keep it safe. The worst thing one could do would be to sacrifice tonal response to attain some time domain goal.
     
    Last edited: Jan 22, 2019
  9. kongbakpau

    kongbakpau New

    Joined:
    Mar 22, 2018
    Likes Received:
    11
    Trophy Points:
    3
    Location:
    Singapore
    Hello friends,

    Thank you so much @murphythecat for sharing your impressions with the Seas A26 kit. I’ve since built a pair and thoroughly enjoy them.

    Coming from Elac Unifi BS U5 bookshelf speakers, the A26 are larger in their presentation. The A26 are non fatiguing @ 12 Ohm. Bass textures are plenty deeper than the Elacs although a tiny bit softer in presentation. I suspect that it is due to the tuning of the Elacs having a midbass hump that seems higher than that of the A26. I also notice that the tweeter — because of the 1st order crossover — seem to cause the reverbs or echos on certain tracks to be more audible and therefore gives a more realistic sense of the room/recording space.

    Furthermore, what is interesting is that I now don’t seem to use the tube buffer of the Schiit Saga pre amp anymore but rather use the Saga’s passive mode more and I tend to listen beyond 70db that I used to listen to with the Elacs. When in tube buffer mode, the effects of the tube seem more noticeable.

    Hey @murphythecat may I know if you still use the 15 ohm or have changed resistors around? What did you settle on eventually?

    Thank you in advance for sharing your setup ◡̈
     
  10. murphythecat

    murphythecat GRU-powered uniformed trumpkin

    Pyrate Banned
    Joined:
    Apr 4, 2016
    Likes Received:
    1,201
    Trophy Points:
    93
    hey! The A26 are great arent they? I agree about the softness in bass. this is due in part of the higer roll off of the aperiodic vent. the A26 rolls off around 60hz and are in room about -6db at 40hz. Ive played around with different foam density in the port, and eventually settled with a closed style foam, making the A26 Sealed. The A26 measurement with a aperiodic port vs with closed style foam indicate that with a sealed cabinets, I get more extension, adding a good 10-15hz of extension. In room now, im flat down to 40hz! I shall post measurements, but I suggest to experiment there. However, I think the Bass driver with its small magnet (which is necessary to make the driver work without a low pass) is the culprit with the bass softness and cannot be overcome completely. I still find the bass great, with just enough weight and punch to satisfy my inner bass slut.

    Also, I suggest to add high density Roxul foam for the inside cabinet and cover every inner wall with Roxul.
    Also, I suggest playing with the toe in of the speakers. I prefer to have basically 3-5 degree of toe in, not more. basically firing straight away. this helps with the roll off of the bass driver, integrates better with the tweeter.
    Ive also changed the mundorf caps to AN copper caps which is a great improvement albeit a bit pricy.

    I still play around with 12 ohm and 15 ohm. I notice that its actually brand dependant and that different resistors have subtlle but noticeable difference of SQ. I use mills 15 ohm right now.

    Hope this helps!
     
    Last edited: Aug 23, 2019
  11. murphythecat

    murphythecat GRU-powered uniformed trumpkin

    Pyrate Banned
    Joined:
    Apr 4, 2016
    Likes Received:
    1,201
    Trophy Points:
    93
    Back to mundorf 12 ohm resistors. Ill have to order a 12 ohm mills resistor but the 15 ohm mills seem to sound too muffled/rolled off. I find the mundorf sometime just tiny bit too bitey but i prefer the 12 ohm mundorf sound over the 15 ohm mills.

    Also, Kongbaku, the tweeter offset must be on the inside. measurements indicate if you palce the speakers with the tweeter offset on the outside, youll have a 4 db dip between like 2 khz to 5khz. really not good
     
  12. kongbakpau

    kongbakpau New

    Joined:
    Mar 22, 2018
    Likes Received:
    11
    Trophy Points:
    3
    Location:
    Singapore
    Thank you so much for sharing all those @murphythecat! I might just experiment with the aperiodic stuffing to see how it changes bass presentation. I’m not a bass head but certainly won’t mind a little more heft. And so looking forward to seeing your measurement results if you do measure them :). Although, I recall from the SEAS brief that increasing stuffing might affect timing?

    I have placed the tweeters on the outside instead of the inside though. I might try putting them the other way around and have a listen again. The A26 are 1.2 meters apart in my living room and I’m sitting 2.4 meters from them with no toe in. As such, I don’t know how much of an audible impact the tweeter being inside or out would be. Especially since I do think I want to tame the 2Khz frequencies.

    Regarding building; I lined all the internal walls (except the front baffle) with 8-10mm thick felt and also loosely distributed the 50 grams of polyester at the back panel following several build guides (including yours) as inspiration. The materials you use are not readily available where I am, and so resorted to the felt. No weird cabinet effects so far — or so I think. I have no way to measure so I don’t know. Oh well :/.

    I’ve also briefly tried the 15 ohm resistor option but did not like it compared to the 12 ohm. I think it is because the treble was too rolled off. And so went back to the 12 ohm.

    A friend has kindly loaned me his Loki mini. I like that it gives the option to tame the 2khz and boost the 20hz to my liking.

    I might just settle on Loki in the system so that I can tweak till my heart’s content! Overall, entirely satisfied!

    Have you heard of Esperanza Spalding? You should give her self entitled album a listen on the A26 if you fancy Latin Jazz. Superb.

    Good day!
     
  13. murphythecat

    murphythecat GRU-powered uniformed trumpkin

    Pyrate Banned
    Joined:
    Apr 4, 2016
    Likes Received:
    1,201
    Trophy Points:
    93
    - tweeter inside vs outside. This is very critical. on the tweeter outside, it gives a sort of large dip of about -4 db if i recall correctly from about 1.5khz up to 5khz. when the tweeter are on the inside, the response is flat. I cannot take measurements now but this outside vs inside thing really troubled me at first cause I thought my speakers were defective as when I got them I setup them with the tweeters on the outside and I thought the speakers were badly voiced with such a pronunced dip. Try both but for sure, if you want a neutral response, tweeter on the inside gives the more neutral presentation and ruler flat measurements.

    - 2khz: The woofer break up mode is around 2khz. the woofer as you know doesnt use a low pass, and the amount of toe in is very critical because of that. if you listen to the woofer on axis, you will get the mid play up to about 2khz -6db and 4khz -8db. not good. However, if you listen about 25/30 degree off axis (so if you listen to your speakers pointing strang forward so you have no toe in at all), you will get a much better result, audibly it sounds more clean. the woofer measurements will now be -6db at 2khz and -15db at 4khz. Much better.

    the Loki may be helpful in your setup, but the 2khz knob is very wide. I have a Loki and when you turn the 2khz knob it affect the entire treble and upper mids. If you use no toe in, this will also lower slightly the entire treble region, as well as making the wofoer behave more like if its having a low pass.
     
  14. murphythecat

    murphythecat GRU-powered uniformed trumpkin

    Pyrate Banned
    Joined:
    Apr 4, 2016
    Likes Received:
    1,201
    Trophy Points:
    93
    [​IMG]

    here are the measurements with Foam in the ports vs closed cell foam (Sealed) in yellow. measurements taken from the LP. No toe in.


    cheers
     
    Last edited: Aug 30, 2019
  15. murphythecat

    murphythecat GRU-powered uniformed trumpkin

    Pyrate Banned
    Joined:
    Apr 4, 2016
    Likes Received:
    1,201
    Trophy Points:
    93
    Ive been playing more with the A26. even though measurements seem to imply the A26 are better sealed, I actually prefer the sound with the foam in the port like Seas designed. it seem to relax the bass and make it seem more agile for whatever reason.

    Im back with the mox 15 ohm and some 10 degree toe in. 12 ohm is too bright for my taste. even with 15 ohm, i find the sound veering into the bright side. really wonder why since measurments really doesnt point to anything that could cause brightness. this may be due to the 10 inch controlled break up to roll around 2khz. not a big deal.
     
    Last edited: Sep 16, 2019
  16. murphythecat

    murphythecat GRU-powered uniformed trumpkin

    Pyrate Banned
    Joined:
    Apr 4, 2016
    Likes Received:
    1,201
    Trophy Points:
    93
    I now use 20 degree toe in and 18 ohm resistor

    I get about -4db from 1.5khz up to 15khz compared to 100 hz to 1khz. I definitely prefer this presentation compared to the 15 ohm resistor. not neutral, but more to my liking.
     

Share This Page