Need advice on omni-directional TQWT design

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi all,
I found some cheap and nasty full-range 6-inch drivers at a junk shop...... couldn't resist and now I'm contemplating cabinet design. I've been reading the Shahinian website about omni-directional sources, have seen pictures of Duevel omnidirectional horns and am keen to give omnis a go. However, I'm also keen to try a folded TQWT design with cabinets similar to those shown in the picture (made from MDF) - credit here to to Bob at http://geocities.com/rbrines1/ (I really liked the stuff on his web site.). Is it necessary for the driver to go on the front baffle near the top?? I was planning on placing it on the top of the speaker firing directly upwards; is this inadvisable????

My plan at this stage is to build the TQWTs with upward firing drivers and then add the dispersion/'reverse horn' element on top later. (for a picture of the Duevels see: http://www.cd-konzert.com/Produkte/Evenus.htm).

Does anybody have any advive on the geometry these "reverse horns" should have? I haven't a clue on this matter - the main objective would be as a dispersing "omnifying" element rather than as 'acoustic transformers' - but that may change.


I know I should really be experimenting, but unfortunately this year I haven't got quite as much spare time as I woul like, so I'm just keen to get some good quality sound from my second system in my study. Hence, the attraction of this simple design.

Thanks for your help,
Joseph.
 

Attachments

  • fe164.gif
    fe164.gif
    14.2 KB · Views: 438
Hi Joseph,

I have done TQWTs and omnidirectional speakers, but never combined both principles, so this is pure theory:

To be scientifically correct you would have to recalculate the TQWT when moving the driver upward. But since it is only a small "travel" around the corner, you should get away with the proposed dimensions of the TL.

For a start the dispersion element doesn´t need to be as elaborated as the Duevels. A plain 90° cone will be sufficient. I have seen comparisons between different cone shapes (straight, hyperbolical etc.), and the straight cone was as good as any. My cones have been turned from a multiplex cube, but you could simply fold a sheet of paper in a cone shape and fill it with plaster or something comparable. Just look for a hard and smooth surface of the cone. It does not need to be polished! The diameter of the cone should be at least the same as the driver cone.

Be prepared to do some work with the frequency response. Unfiltered it will be VERY different compared to the on axis response. So you will need some notch filter taming a wide elevation of the SPL somewhere between 500 Hz and 5 kHz.

Hope this helps

Rudolf
 
Thanks Rudolf,

I never even thought about off-axis response. I only have a multimeter, so have no way of designing the correct notch filter if I don't know where it should be (on the frequency spectrum). Is there anything I could do with the dispersion element to create a better on-axis response???

Should I re-calculate the TQWT dimensions so that when the driver is on the top of the speaker it is exactly half way along the length of the pipe??

Thanks,
Joseph.
 
Joseph,
you will have a hard time doing a correct notch-filter only with a multimeter :D

For wavelength that are very long compared to the driver dimensions an omnidirectional speaker will almost behave like a normal speaker.
For wavelength that are very small compared to the dimensions of the cone the driver will behave like in a normal box with one BIG difference: the energy is not radiated into one ~90° cone in space, but approx. into four 90° cones in space (resulting in 360° dispersion horizontally). So the energy fed into your room will remain the same, but the energy radiated to your listening position will be only a quarter.
For wavelength that are comparable to the dimensions of the driver and the cone, dispersion will be heavily relying on the individual geometry of those elements.

So there is no easy way to come to an adequate result. :xeye:
The best you could do IMHO would be to take a proven omnidirectional design (the only one with a fullrange driver I know uses the Jordan 92) and change its sealed or BR enclosure to TQWT.

If you are familiar with Martin Kings worksheets for TQWTs, you could certainly recalculate the dimensions. But you wouldn´t look for the driver to be exactly half way along the length of the pipe (as this is NOT the optimal position!), but for the pipe to be correctly distributed to both sides of the driver.

Hope this was not too desillusionating :angel:

Rudolf
 
Thanks Rudolf,
Well, I'm not too disillusioned - I'm glad I know of these problems now and not finding them out when I st back to listen to my newly-completed speakers. Thanks for you comments.

I think I'll just build a conventional TL. The omnis would be nice but they're a bit beyond me with my limited resources of time and equipment.

Regards,
Joseph.
 
Hey! I recognize that drawing! ;)

If you want to put the driver on the top, you really should recaculate the fold. TQWT's are rather touchy to driver position. If you screw it up, you get a large suck-out in the 100-200Hz range.

I have no clue about using a cone as a dispersal element. My only experience with up-firing speakers is with HT surrounds, an application where up-firing works very well.

Bob
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.