11/21/12


Triangular Cross Section Labyrinth Loudspeaker story.

There was a customer request to design shelf loudspeaker with rather controversial input data and “wish list”:
- only shelf size form factor was accepted,
- 7 inch Sonido 175 SFR to be used (driver supposed to be normally used in floor standing back loaded horns),
- customer was fed up with bass reflex sound, so no bass reflex.
So we have got such options for enclosure to design:  transmission line, “Rogozhin” labyrinth, horn, sealed box.
Sealed box – no bass at all with this full range driver, horn – to big to be done as shelf speaker.
So could try short TL or labyrinth. Personally I prefer “Rogozhin” type of labyrinth (of course it`s still an variation of TL).
So, decided, we try to go with labyrinth enclosure design.

Here we met another factors.
First, to get needed bass level normally all TLs should have port opening close to either floor or wall surfaces. Floor is rather problematic with small shelf loudspeaker, but luckily customer confirmed that it will be placed really close to the front wall.
Second, length of the labyrinth channel should be at least 150 cm be tuned for 57Hz, and cross section not less than speakers Sd (132 cm2).  How to convert these points in shelf speaker with limited dimensions, especially height and depth?  Quite a tricky task….

This second constraint initiated some brain storming with lots of paper used for sketches, extensive soft usage for simulations. As a result we came to the solution, which probably would not be born if not customer insistence.
Interesting labyrinth layout was done: Triangular Cross Section Labyrinth (TCSL) turned by circle, or spiral, depends how to look at this layout.

One moment still was not clear enough. How the driver acoustic loading will be finally tuned if we have such uncommon triangular turns of the channel?  Previous experience with rectangle cross section TLs was that without smoothing out the turns we will get tuning lower than calculated. Ok, lower...but how much lower? How to predict it on the stage of calculations?


We were pushed to make draft full scale prototype to understand how to correct calculation in future for such designs. This was found out, and now is know how of  S.A.M.

After listening tests and measurements minor corrections were done in labyrinth parameters. Process of making finished loudspeaker was started.  Additionally front panel was changed to correspond to Golden ration principle, after diffraction estimation was done.   




Acoustical treatment of the labyrinth was done with felt.  TCSL allowed us to avoid modes between inside walls and back reflections of the mid-highs to the Sonido diffuser. As a result, there is no need of internal damping with materials like wool, etc. If you ever tried to tune TL with wool you know the situation when too small quantity of wool means muddy sound, to big quantity means weak, lazy bass….TCSL solved this problem in cardinal way.

Final loudspeaker is on the photo.
In room measured response (45Hz at -3db) is quite good for 7 inch full range driver in “on shelf form” enclosure.  All advantages of Sonido were realized and due to labyrinth principle speaker is much more “friendly” to room modes which are less excited than with bass reflex enclosures.  Bass itself is accurate, and not interpreting all sounds in one way, like “one note bass”.  No delay between low end and mid high can be observed by listener, as group delay is changing very gently, without big peaks.
Of course, still optimal enclosure for Sonido SFR 175 is floor standing back loaded horn. In case of certain room (or wife))) limitations this labyrinth is perfect solution.

P.S. Recently received some comments from  DIY community that above mentioned speakers are "copy" of some Bailey`s design from 1970s.  Well, quite interesting to know that, especially that for me it was first time to hear about those speakers from 70th. 
The design is not exactly the same, after checking old articles: Baily`s one was tapering TL without 1/3 displacement, three sections instead of four, but still trinagular shape. Quite a lot of differences not to be a "copy", both for design and for the sound.

20.11.2012  Volodymyr Saburov,  S.A.M.






1 comment:

  1. What an excellent idea. Great work!

    ReplyDelete