400Hz-20KHz horn for BMS 4595

A Klipsch K-402 HF horn on top of a LaScala. It's conical in the throat area, with a tractrix mouth.



It's large enough to have good horizontal directivity control to 400hz, and still have good dispersion at 16Khz (without a slot).

http://forums.klipsch.com/forums/storage/4/1959035/K402_hor.pdf

http://forums.klipsch.com/forums/storage/4/1959037/K402_ver.pdf

H=25.5" W=39.5" D=16.25" 2" throat.


Edited by _djk_ - 28 April 2013 at 2:03pm

Pavel, the main advantage here is getting a true point source. Meaning one source is always better then two sources.

Edited by MarjanM - 28 April 2013 at 2:31pm

there was a comment about the sound quality at such a low frequency. I only have experience with the B and C coaxil compression driver, which also has a similar x/o point. These drivers are made to go that low. I'm not sure how the BMS driver sounds at the lower frequencies, because I believe that the larger diaphragm is also metal. But the larger diapragm in the B and C driver is a paper cone, and I guess in part, this also combats any harshness at lower frequencies. I think this gets a better result for upper mids than depending on the high end of a 12" speaker to do the same job.

MarjanM wrote: One more thing, the response is within 4db (+/- 2db) from 500Hz to about 18Khz following one central line all its way. And it is much smaller horn then the other one. It is 29x29cm square front 24cm deep 60x40 dispersion.

As I said, the low end response is set by the included angle. There is not much difference between yours and Timebomb's dispersion angles so the low end should be similar even though your horn is smaller.

The effective length of a horn can be slightly longer than the physical length. Add to that the additional length in the driver, between the diaphragm and exit and let’s say it is 34cm long. If we go with 340msec^-1 for the speed of sound, to make things easy, the length of your horn is 1 wavelength at 1000Hz. If the mouth area is too small the horn will exhibit resonant peaks at multiples of ½ wavelength of the horn, so 500Hz, 1KHz, etc. This can be seen in your measured response. Timebomb’s larger horn does not exhibit these ripples.

This may or may not be a problem. Size versus slight uneven response it is the designers choice. The thing is, if you don’t know the theory, how do you decide which way to go if you wish to modify your horn.

The other thing that doesn’t show in either response graph is the off axis response. The larger horn will have a lot better control at lower frequencies. Again the loss of control (wide dispersion pattern) might not be a problem depending on what device is being used in the next passband down. With a compression driver of this size and output, I wouldn’t expect it to be used above a single 12” cone.

_djk_ wrote: It's large enough to have good horizontal directivity control to 400hz, and still have good dispersion at 16Khz (without a slot).

I guess it depends on what you call good dispersion at 16KHz. At 1.6KHz the coverage is 90˚ whereas at 16KHz it is down to 50˚, highlighted with red arrows. The horn also exhibits fingering, highlighted by yellow arrows.

I’m not saying that it is a bad design, but a tractrix flare means that it will have narrowing dispersion at higher frequencies. Assuming that the designer knew what he/she was doing, it will have been a conscious decision, and the advantages of the flare curve will have been considered to outweigh the disadvantages.

I think that this all goes to show that Timebomb already has a horn that can compete with commercial designs.

Cheers Steve, i chose a linear / conical expansion as it was easy to build and has more consistent directivity compared to exponential, hyperbolic, tractrix etc.  I built some tractrix flares for some other drivers and they sounded great on axis but the beaming was too much for me.   I knew the large mouth would cause lobing at the top end but i think you need a large horn to get down that low cleanly,  its not something you can really get away from.  And i thought it was best to make it as big a possible but fit inside the cases that fit inside the low mid horns.  I have some smaller OHM horns that the drivers came with and they are not as clean as my big horns, even with a much higher high pass. 

I still think its worth trying some mouth correction, i thought of a different way to do it, more like the Klipsch horn DJK posted, i could possibly truncate the existing horns and add a roundover / front baffle.  I still have a OSB prototype that i could cut down to test with, and i think i have some 6" plastic pipe somewhere.   

I will have a dig on my hard drive and see if i took any off axis measurements of this horn,  i think ARTA will do polar plots so i recon my next step is to make a scabby prototype with the osb horn and take a load of measurements.

Something like this.