Better if the tweeter is at least a little bit off center. You always want distances to side edges to not be equal all the way around.
Also, left and right speakers should be mirror images of each other.
In Jack's simulations, it was better for the left tweeter to be left of the centerline and the right tweeter to be right of the centerline.
Also, left and right speakers should be mirror images of each other.
In Jack's simulations, it was better for the left tweeter to be left of the centerline and the right tweeter to be right of the centerline.
If I correctly understood the arguments presented here, just in case of extremely rounded corners is when offsetting the tweeter can be avoided, the other extreme is the usual sharp corners the normal diyers would do, and hence the offsetting as showed by the OP would have the bigger beneficial impact.
Besides this is my humble contribution to the "The degradation of Proper English" thread😊
Besides this is my humble contribution to the "The degradation of Proper English" thread😊
I'm sorry, I didn't read all the replies. If you don't want to use WGs etc, the best way to get a constant directivity is to use the drivers according to their membrane diameter. Yes, you can add mechanical-acoustical gimmicks like cardioid side 'vents' etc but the easiest way is to add another way to stay within a certain dispersion angle. Ie, for a 8" mid-bass and a 1" dome tweeter, a 2" cone midrange driver will fulfill the dispersion pattern nicely.
How wide or narrow you want your dispersion can be controlled by the xo frequencies and the driver membrane surface/diameter. The bigger the diameter, the more it beams and the higher it has to play, the narrower it becomes. Yes, there are formulas and tables to read and plan but since there are a lot of drivers which divert from that pattern (BMR, full-range drivers with a whizzer, special membrane geometries) and every driver got its own breakup where only parts of the membrane are active or even move the opposing direction, the baffle having a big impact on that too and the crossover having a big impact on that too, that's just a rule of thumb.
My suggestion is - and that's way more important than anything else - determine, what dispersion you really want/need and go from there. Everything else is just fishing in the dark and hit/miss. If you know what dispersion you like or need, that's a point to get results and not just guessing or someone elses preferences.
How wide or narrow you want your dispersion can be controlled by the xo frequencies and the driver membrane surface/diameter. The bigger the diameter, the more it beams and the higher it has to play, the narrower it becomes. Yes, there are formulas and tables to read and plan but since there are a lot of drivers which divert from that pattern (BMR, full-range drivers with a whizzer, special membrane geometries) and every driver got its own breakup where only parts of the membrane are active or even move the opposing direction, the baffle having a big impact on that too and the crossover having a big impact on that too, that's just a rule of thumb.
My suggestion is - and that's way more important than anything else - determine, what dispersion you really want/need and go from there. Everything else is just fishing in the dark and hit/miss. If you know what dispersion you like or need, that's a point to get results and not just guessing or someone elses preferences.
Yes that’s right.If I correctly understood the arguments presented here, just in case of extremely rounded corners is when offsetting the tweeter can be avoided, the other extreme is the usual sharp corners the normal diyers would do, and hence the offsetting as showed by the OP would have the bigger beneficial impact.
Besides this is my humble contribution to the "The degradation of Proper English" thread😊
@Juhazi As much as I'm a big fan of those fancy Audax neo dome tweeters, I don't like the TW025A26 soft dome version as much. The top end is really fizzy on this tweeter with alot of breakup. Its a fairly rough sounding driver for a smaller dome. I'd actually rate the much less expensive TW025A0 overall better sounding in some regards. The catenary dome shape of most smaller Audax domes provides a higher on axis rising HF trend. This can be used as an advantage if the driver is used in a 2 way with a larger woofer.
Brainstorming the non-WG CD principle, takes me to a different design approach. Perhaps If you array a smaller FR driver like the Tymphany TC9 in a circular quasi coax arrangement with a somewhat larger than usual tweeter in the center, you can achieve similar directivity results as with a 6" to 8" coax, but without some of its drawbacks.
By stretching this multi driver arrangement into an elliptical shape (narrow width, taller height), the vertical directivity narrows to reduce floor and ceiling reflections and increases higher on axis sensitivity. You'd have to splay the driver angles horizontally to reduce HF combing and get the desired directivity trend that meets up with the tweeter directivity at desired xover point. This also requires EQing the entire outer array with a rising HF response to counter falling array sensitivity gain towards the top end.
I wonder if there was a way to expand this type of driver arrangement into a system which can allow for variable directivity ie. the newer B&O Beolab 90 system. I realize that I'm treading on unorthodox ground here, but the reality of having variable directivity to suit specific listening environments without the need of larger 2 way WGed designs which typically don't sound as good in the midrange. Same goes for larger coax systems with inherently rough HF linearity.
Brainstorming the non-WG CD principle, takes me to a different design approach. Perhaps If you array a smaller FR driver like the Tymphany TC9 in a circular quasi coax arrangement with a somewhat larger than usual tweeter in the center, you can achieve similar directivity results as with a 6" to 8" coax, but without some of its drawbacks.
By stretching this multi driver arrangement into an elliptical shape (narrow width, taller height), the vertical directivity narrows to reduce floor and ceiling reflections and increases higher on axis sensitivity. You'd have to splay the driver angles horizontally to reduce HF combing and get the desired directivity trend that meets up with the tweeter directivity at desired xover point. This also requires EQing the entire outer array with a rising HF response to counter falling array sensitivity gain towards the top end.
I wonder if there was a way to expand this type of driver arrangement into a system which can allow for variable directivity ie. the newer B&O Beolab 90 system. I realize that I'm treading on unorthodox ground here, but the reality of having variable directivity to suit specific listening environments without the need of larger 2 way WGed designs which typically don't sound as good in the midrange. Same goes for larger coax systems with inherently rough HF linearity.
You shouldn't be seeing a problem with Geddes style midrange. Dialling a speaker in to the room is quite a process, so is the cross. The woofer should be selected and used to manage breakup. A speaker of this calibre has the potential to let smaller issues stand out if they are allowed to stay.
https://salfordacoustics.co.uk/sound-waves/diffraction/diffraction-through-slits
Interesting simulation......could this part of the allure of the Behringer 2031p?
Interesting simulation......could this part of the allure of the Behringer 2031p?
No doubt IME; the Altec A-8 used a 12:1 slit to better blend a HF horn designed for <12" drivers to a 15" to make a 'full' sounding cinema speaker in a very compact box for small cinemas, surround sound speaker apps.
Basic specs
Basic specs
" Bunching" otherwise known as normal filter summing. +3 dB rise from Butterworth is well known behavior.Notice some "bunching" around 2500Hz. This bunching comes entirely from the tweeter, as the crossover is 1700. At 45 degrees off axis, the output is +4dB higher than on axis.
BUT guess what… if I set the crossover at 4900Hz I get this:
And can happen either on or off axis depending on the normal everyday phase relation of the 2 drivers. This case off axis had a 3 dB rise.
5" driver usually wont get crossed over at 1700 Hz that is closer to a 7" driver. I would have just started at the usual 3500 expected beaming frequency.
Going to a active filter model you will likely find " Bunching" or the 3 dB rise from a Butterworth filter sum rather easy to observe.
regardless of frequency or baffle diffraction.
No such thing as " Baffle beam" or " Baffle bulge" it is called baffle diffraction. For accuracy or the term accuracy seems to have a loose definition thanks to YouTube and audiophile reviews. System accuracy is the same old flat response. And for accuracy or flat response, yes asymmetrical driver baffle placement.
It has been and will always be, based on constants found in a 1.6 to 1.7 ratio. Problem is for 25 years or more the hifi market likes to sell what " looks" right
not based on any real world diffraction science. Basically wasting early pioneers time who figured this out 70 or so years ago without computers.
Also no such thing as a polar "heat" map, it is just a Polar Map based on frequency response.
Tweeters are non directional till 8 or 10k. Summing them to a mid or woofer with a tilted response is just done with filters and padding.
Any bunching again uneven summing is fixed in many ways. It is a phase relation. And over time since computer aided design shows the phase response rather well in a graph on the screen. If the model is accurate. You will begin to notice where more friendly points will be in the falling slopes.
Anything from dead centered drivers to MTM is just a dead failure which is from marketing things that " look" right. Or otherwise people making up pseudo science to explain why they work and then gets accepted. When luckily computer aided design or straight simulation can punch holes in most imaginary theories. And immediately show accuracy issues, and real world off axis holes vertical or horizontal. Which do nothing for " image" or real room response.
Last edited:
This 3dB rise is not from a butterworth filter. It's an anomaly in the polar pattern caused by diffraction of the tweeter's output. It also doesn't have anything to do with active vs passive crossovers. Those are horizontal polar plots, not vertical, and what I refer to as "baffle bulge" is an increase in tweeter output at +/-45 degrees off axis over a range of frequencies specified in the model.
Last edited:
- Home
- Loudspeakers
- Multi-Way
- Constant Directivity without Horns or Waveguides