There is a quandary that I can’t resolve.

There is a hypothetical horn-loaded chanall operating for instance flat form 100Hz to 5000Hz. Presumably I need to cut this channel off at 300Hz, very sharp-with 3-4 acoustical orders, but according to the rules of the game I must not use electrically any other thing then a first order coil. Now, here is the question: how might I do it? How to truncate sound from the channel without introducing the high order filter? The solutions around making a recursively-bent horn with intention to notch-out the “sonic tail” in the resonance chambers do not offer.

If someone would be able to offer anything that moves me in this direction then a wonderful musical handouts, not available by any other mean, are available.

Rsg,
The Cat

PS: Forgot to mention - also do not propose any DSP solutions.

Hello , Roma ! I think that you can use only a back chamber with decreased size , and a serial capacitor . I think that there are no ways besides  ...

Dima

deemon wrote:

Hello , Roma ! I think that you can use only a back chamber with decreased size , and a serial capacitor . I think that there are no ways besides  ...

Actually it would be a serial coil, as I need a low pass not a high pass.

Unfortunately the back chambers, for what I need to accomplish, are worthless. The back chambers are very affective for lower knee of roll-off but they do nothing with high knee. Well, they do provide very-very-very minute effect for a natural HF roll-off of the drivers loaded into the horns but considering that I need to cut the frequencies in the very middle of the range (of actually in the beginning of the range) the back chambers are useless.

I still have no idea how to roll it off abruptly without going to a separate amp and a high order crossover (I can’t change driver). I wonder it would be possible to construct some kind of active wall within a back chamber that would be sitting inside of own sealed chamber. The wall would have a heavy equalized driver running in contra-phase to the main deriver… I think this would help to actively damp the LF but not the HF. I think for killing HF I would need to use an EQed contra-phase transducer inside the bell of the horn… Of course no one would be able to say how this will sound… Most probably it would be horrible… or at least it certainly will me more harm then utilizing 4th order filter :-)

I juts do not know at this point any other solution and I’m just purely fishing…

The Cat

Of course , if you need low pass , the chamber won't help .... but in seems to be that there are no good solution in this case . Too narrow bandwidth , and it will sound like a barrel :-) .

Dima

Actually another idea that I was "considering" was to put a sound-absorbing screen inside of the bell. I used it a few years about and it worked very well. I needed very slightly to roll-off a MF driver and I was not able to use indictors. I ended up placing at the mouth of the horn the acoustically transparent screens and then I was editing the thin absorbers until I dialed up the necessary response. It worked with a slight alternation of HF but I do not think it would be possible as a brutal intrusion in the middle of the range… unless some king of overly affective Helmholtz Resonators would be used… Some kind of well-tuned whistles extending into the bell.. :-)

The Cat

I think a parallel notch filter might do the trick. I find that parallel circuits do not get in the way of the music as much as components that are placed in series with the source. A C-L-R circuit wired across the woofer’s input terminals should do the trick. Of course I don’t know what values you would need to use. However, I can run the calculations to give you “ball park” figures to start with. The main fault of parallel circuits is if designed poorly, it can screw up the impedance that the amplifiers see. Still interested?

Rgs, JLH

Thanks, JLH

I would consider it, however my experiments with notching lead me to feel that high level parallel notching is as bad as series. A series network removers meaningful tone out of acoustic pressure. The parallel network does not do it but it eats transient response. In both cases the “bad things happens”.  Also, what whatever reasons the parallel network provides different effect with the different but identical drivers. I think I have to explore more opportunities of a parallel notching at line level…

Rgs
The Cat

Romy,

     What you say about both series and parallel filters is true. The parallel circuit does eat transient response, but if you keep the minumin impedance higher, it does not stress the amplifier so much. For your 16 ohm woofer I would try a 20 ohm resistor and a 27uF or 28uF capacitor in series and place it across the woofer terminals. This acts some what like a Zobel network, but will maintain a minumin inpedance of 11.3 ohms to the amplifer at 300Hz. As frequnency rises, of course impedance will drop, but at higher frequencies the amplifer should not be too challenged. It is possible it could screw with the mid range horns, but this would have to be listened to in order to determine how much damage it does. Other than this, I'm out of ideas.

Rgs, JLH