A question on directivity

[swak]

I am working on a non-horn 3 way system. I want the directivity to be controlled (i.e. somewhat narrow) and that the driver integration is apropriate "directivitywise". How many dB´s is the maximum desirable differece between drivers at 15, and at 30 degrees respectively at the crossover frequencies (mine are about 350 and 3000Hz) ?

Thanks in advance.

 

[John Sheerin]

They should be as close to the same as possible (at all angles).

 

[swak]

Perhaps I wasn´t very vlear with what I meant with controlled radiation. I´m not aiming at a certain fixed dispersion all over the spectrum, but only 1) trying to mimimize directivity jumps between drivers at crossover points 2) that the directivity is somewhat narrow, at least from a point up.

I`ll be using dipole woofers which have a cardioid-like dispersion, an 8 or 10" midrange, and a tweeter on top, perhaps a ribbon(-like)one.

Any ideas or suggestions are appreciated.

 

[soongsc]

Perhaps I wasn´t very vlear with what I meant with controlled radiation. I´m not aiming at a certain fixed dispersion all over the spectrum, but only 1) trying to mimimize directivity jumps between drivers at crossover points 2) that the directivity is somewhat narrow, at least from a point up.

I`ll be using dipole woofers which have a cardioid-like dispersion, an 8 or 10" midrange, and a tweeter on top, perhaps a ribbon(-like)one.

Any ideas or suggestions are appreciated.

Take a look at this site and see if it makes sense.

http://www.kettering.edu/~drussell/Demos/BaffledPiston/BaffledPiston.html

 

[poptart]

I don't suppose anybody has the animations from that page saved do they? They don't seem to work anymore.

 

[Rudolf]

I`ll be using dipole woofers which have a cardioid-like dispersion ...

... so it would be best to follow that dispersion pattern with a dipole midrange. A 8 or 10" will start to beam at less than 3000 Hz. Therefore the tweeter dispersion should start with less than 180° too.


BTW.: Dipoles don´t radiate cardioid-like, but in a figure 8 pattern.

 

[soongsc]

I don't suppose anybody has the animations from that page saved do they? They don't seem to work anymore.

I don't recall them animated on line, but they just seem to be animation results from either MATLAB or MATHEMATICA or some analysis program like those.

 

[John Sheerin]

Perhaps I wasn´t very vlear with what I meant with controlled radiation. I´m not aiming at a certain fixed dispersion all over the spectrum, but only 1) trying to mimimize directivity jumps between drivers at crossover points 2) that the directivity is somewhat narrow, at least from a point up.

No, I got that. I guess I was not clear. That is what I meant - that you should have smooth transitions between drivers. Ie, the woofer should have some polar response at the crossover frequency and the mid should have exactly the same polar response at that frequency. That's what I meant by ' They should be as close to the same as possible (at all angles)'. I didn't mean they should be omnidirectional - just that the output level at 60 degrees, for example, of the woofer should match the output level at 60 degrees of the mid, etc.

Basically you will probably have a rough time without resorting to some sort of horn or waveguide to restrict the pattern. Especially with a ribbon - you will have essentially omni radiation in one plane which will likely not match up to any decent sized driver covering the frequency range below it. Unless of course you did something like a super wide range ribbon and crossed that to a sub, but there are still issues there.

For some examples (with horns), see the polar maps on my webpage.

 

[swak]

Rudolf:

Perhaps you could share a word or two about the dipole configuration. As I said I plan to use pair of dipole subs, each with two 15" drivers, probably cheap Pyle or the one of the equally priced from steelsound.com.

If I cross to an 8" driver at about 300 Hz (I know I might have to equalize the peak moving to the crossover point), how would the dispersion of the W-frame dipole at 300Hz be?

Would the directivity of an 8" driver like a B&C 8PE in a closed box aproximately match the dipoles directivity at the crossover point (300Hz)? I`m afraid the mid will be wider radiating than the sub, but I don`t know the functions for directivity to calculate how much.
Thanks in advance.

 

[swak]

John:

I was completely aware you had got the point. I should have specified this when writing, sorry for that.


What you point out about the tweeter is true, I think there is no other way. The horn resonances should be a smaller penalty at higher frequencies, therefore triying to get the midrange to go higher (moving the crossover freq. conveniently out of the most troublesome range) would be a nice thing. Perhaps stepping down and using the widerange 6" B&C, but only when the resonance and lower crossover point is more anoying that the slight lack of dinamics from a bigger mid.

Thanks for your reply.

 

[poptart]

I don't recall them animated on line, but they just seem to be animation results from either MATLAB or MATHEMATICA or some analysis program like those.

sorry, it was my computer not the webpage that wasn't working. They display fine here at work. I thought they were animations because at the top of the page it says "Acoustics Animations"

 

[Rudolf]

If I cross to an 8" driver at about 300 Hz ... how would the dispersion of the W-frame dipole at 300Hz be?

A W-frame dipole is not as effective in cancelling radiation at 90° as a plain open baffle - but you can still expect attentuation of 10-20 dB @90° and 300Hz.
An 8" driver like a B&C 8PE in a closed box would be ca. -5 dB @500 Hz and 90°, but almost radiating full circle @300Hz.
This match certainly is less than perfect.

In an open baffle the radiation of the 8" driver @300Hz is fully determined by the dipole effect. It would follow the radiation pattern of the W-frame closely up to ~60°, from there on approaching almost null radiation @90°.

Looking at the next transition @3000Hz:
Around 2000Hz the natural radiation pattern of a 8" driver is almost the same as in dipole configuration - it doesn´t "see" the baffle anymore. Up from there the radiation beam is steadily narrowing.
So you certainly don´t want a simple dome or ribbon tweeter to take over @3000Hz. You will need a horn or waveguide of some kind. If you are lucky, a small cone tweeter could do the job.

I don`t know the functions for directivity to calculate how much.

In their discussion of H- and U-frames both Linkwitz and Kreskowsky are explaining the radiation patterns in detail AFAIK.

 

[AJinFLA]

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.

 

[Rudolf]

@AJinFLA

I almost waited for you to chime in. Your design was exactly was I was trying to explain to swak. Have you got any polar diagrams of your setup as a proof of concept? Seems almost nobody has ever done a polar for his H-frames.

 

[AJinFLA]

Thank's Rudolf, I'm glad to know I've become so predictable . When I'm posting on Saturdays,it usually means that I'm too hung-over to be doing anything else . The FLA part of my name should tell you that there are LOTS of things I could, or probably should, be doing other than polar plots (the beach comes to mind ).
Sorry, that design is now approaching 2 yrs old and was done largely on the fly, I did lots of measurements - but kept virtually nothing. I was doing DIY long before becoming a member of these sites. It never even crossed my mind that I should have kept a detailed record of what I did like some of the nicer folks like Zaph,etc. My Orions were built over a period of a year at my office - on weekends. I live in a condominium - no garage/yard. Neighbors. In the limited time that I have to play my setup now, I'd rather listen than remeasure at this point. Maybe at a later date I will do such things (like when I get a house with a garage and back yard!). Just typing this is making my head hurt. I'm going to go lay back down now

Cheers,

AJ

 

[swak]

Thanks Rudolf, very informative. I had arrived today at the conclusion that an OB mid would probably be the way to match it to the dipole sub. Now, I would find it nice to use the mids as high as possible (up to what point it is really helpful to avoid crossover overlapping is a little unclear, some tend to see 4k a safe point while others only 7k or even more), but the penalty would probably be 1) lack of dynamics for using smaller 6" drivers 2) cone breakup. One compromise would be to use 2 6" drivers to a point where they would cross to the tweeter without beieng so directional as the single 8".

For the tweeter a dome in a waveguide could do the trick. Have you seen small pseudo-ribbons in waveguides, do you think they are theoretically possible? The small BMS (which is very well reputed by the way) CD would be an option, in a waveguide.

AJinFLA:

Vey nice system. Looking at your pictures made me think how close I have come to think about a system like the orions, although with much more sensitive pro drivers and trying to circumvent the heavy EQ used by SL. Enyoy thpse beautiful beaches

I would apreciate if anyone can point me some ideas for the tweeter wave loading.

regards.

 

[AJinFLA]

Hi Swak,

I have long planned on doing a HE design myself, but was far away from doing so until I saw these: http://www.shanghaiproaudio.com/specsBM8CXA.htm
Available here: http://www.usspeaker.com/paudio bm8cxa-1.htm at a price I could not resist (awaiting delivery of a pr. ).
Just about all of my future designs will incorporate certain elements. First, unless a wall speaker, they will be OB. No closed box. Unless it is for low (sub)bass or a very small design. The mids will always be boxless. Second, I'm gravitating more towards coincident driver arrangement. Like the P.Audio. I've heard many good things about that brand. Keep in mind that you can always upgrade to a BMS HF compression driver (which I have also heard are very good) at a later date - if you chose to do so. If I thought a HF cd (like BMS) gave me superior overall performance to my WG XT19s (which replaced Millenniums) or the top SS (like these http://www.d-s-t.com/link/scs/data/D2904_710000a.htm - which would have replaced the XT's - had I decided to continue developing that design) I would not hesitate to get it. They do seem to have the low distortion that is also a must for me. I dont need that type of huge SPL (120db+) capability for my listening environment that the pro units are quite capable of. What I do need (3rd criteria) is broad (but controlled) dispersion - wide directivity if you will - throughout the entire operating range of the system. This is where I think the cd is weakest, 14-20k, off axis. Perhaps the BMS (or P.Audio!) will change my mind. You almost never see off-axis measurements for cd's. Now I realize that they can be mated to quite a few different horn lens, which will obviously affect dispersion. Ok I seem to have rambled quite a bit here, so let me just say that the BM8cxa will be used in a cardioid type design (the 3mm xmax is decent) so that it can be crossed reasonably low (125-160hz range for me). I don't like going higher than that with a dipole woofer system. Probably a asymmetric H-frame with HE woofers for this project. You may want to take a look at the P.Audio to see if it fits yor needs. Then again you may not.

Cheers,

AJ

 

[swak]

AJinFLA:

Have you seen the coaxial B&C drivers? There is some information on this forum about them, specificaly someone who sells both and has recommended them above the PAudio. Anyway, the PAudio could be very good.

Why not go with a bigger coaxial, like a 10 or 12 incher? That is what I satarted thinking about and then I moved on to non concentric drivers. While you don`t need 120 dB spl, the dynamics and (non)compression will be better with the larger cone.

http://www.cadaudio.dk/8cx21.htm

 

[AJinFLA]

Swak -

Have you seen the coaxial B&C drivers?

Yes. Looked at those, Beymas, Radians, Ciare/BMS, etc,etc.
Show me one with the P.Audios performance for $109.
swak -

Why not go with a bigger coaxial, like a 10 or 12 incher?

Same reason I use a 8 and not a 10/12 in my Orions. If you understand the reasons why SL chose the W22 (an 8"), then you will know why. For the intended purpose/range, the 8 will be superior to a 10 or 12. Those will be beaming/breaking up way lower. No good for me. The 8 will have plenty of dynamics in my design. Again, this is for my home/listening tastes, not an arena.
Good luck with which ever path you choose. Have fun building.

Cheers,

AJ