Totally 🙂 And one of the best features of multiway speaker: try and push issues out of band. Drummer goes first 😀 I get my coat
"Ideal" ? No such thing. Everything is a tradeoff. The skill is picking the best set of tradeoffs you can do.
I have never needed to exceed what I can do with 3/4 MDF. The box is the easy part. The crossover is where you should send your time. The drivers is where you should spend your money. I am amazed in how much effort and money is spent on boxes for $30 tweeters. Don't overthink the box as it is the least significant part. Personally, I brace internally to keep any given panel under about 40 sq inches. All different of course. I found 3/4 inch dowels to work quite well. Easy, cheap, little additional reflections internally. I have seen bracing that does little than reducing the volume excessively. Too much bracing in corners where they need none.
I have a full wood shop and can make about anything I can imagine, but it turns out, I have not exceeded what can be done with a box. Next increment of time, next increment of money is still for drivers and crossover tweaking.
I have never needed to exceed what I can do with 3/4 MDF. The box is the easy part. The crossover is where you should send your time. The drivers is where you should spend your money. I am amazed in how much effort and money is spent on boxes for $30 tweeters. Don't overthink the box as it is the least significant part. Personally, I brace internally to keep any given panel under about 40 sq inches. All different of course. I found 3/4 inch dowels to work quite well. Easy, cheap, little additional reflections internally. I have seen bracing that does little than reducing the volume excessively. Too much bracing in corners where they need none.
I have a full wood shop and can make about anything I can imagine, but it turns out, I have not exceeded what can be done with a box. Next increment of time, next increment of money is still for drivers and crossover tweaking.
Sorry, I feel opposite the box (construct with drivers with particular size and position) is the most important part and the crossover is the trivial part. Drivers have to be at least problem free, picked to work within the construction.
But, I don't know it all. It is just a trivial task to make a crossover in VituixCAD, really the simplest thing in comparison to figure out how to make such construct that the crossover makes the end result "perfect". I mean it is easy to perfect crossover for given measurements. And the measurements depend on the construct!
If the measurements are wonky, there is no such crossover that makes the acoustic output of the system good. Any single axis / observation point can be made perfect (with EQ at least), but try and make it good to all directions (that are relevant)!
Think it this way, assume ideal drivers that have flat on axis response 20hz-20khz or what ever, but had size that makes them beam! Alright you could use the 1" tweeter all the way to subs except the excursion would be so wild you'd have to protect your eyes. Other than that, it would be still you construct that determines the performance! You'd still probably want multiple drivers and would have to figure out to avoid resonances, diffraction and where to put the crossovers to avoid off axis problems. Consider the room etc. Drivers and xo are just part of the system, the construct.
But, I don't know it all. It is just a trivial task to make a crossover in VituixCAD, really the simplest thing in comparison to figure out how to make such construct that the crossover makes the end result "perfect". I mean it is easy to perfect crossover for given measurements. And the measurements depend on the construct!
If the measurements are wonky, there is no such crossover that makes the acoustic output of the system good. Any single axis / observation point can be made perfect (with EQ at least), but try and make it good to all directions (that are relevant)!
Think it this way, assume ideal drivers that have flat on axis response 20hz-20khz or what ever, but had size that makes them beam! Alright you could use the 1" tweeter all the way to subs except the excursion would be so wild you'd have to protect your eyes. Other than that, it would be still you construct that determines the performance! You'd still probably want multiple drivers and would have to figure out to avoid resonances, diffraction and where to put the crossovers to avoid off axis problems. Consider the room etc. Drivers and xo are just part of the system, the construct.
Last edited:
Modding cross overs is a little beyond my competency at the moment. I know roughly how capacitors and induction coils work.
Will look into minionkens.
I am not smitten with love for sda effect. Always happy to hear critique of anything in my life including Polk sda.
Will be building up boxes next week with tilted middle wall or V shape. Depends on what space constraints dictate.
Very much appreciate all posts. This is a great place to learn
Will look into minionkens.
I am not smitten with love for sda effect. Always happy to hear critique of anything in my life including Polk sda.
Will be building up boxes next week with tilted middle wall or V shape. Depends on what space constraints dictate.
Very much appreciate all posts. This is a great place to learn
If you are going to build a speaker with 4 midrange drivers in a rectangular array and a single tweeter somewhere in the middle, the size and dimensions of the baffle and where the drivers are located relative to each other and the edges of the baffle are far more important than the shape of the back of the box. It's not clear if you have really solved that problem or not.
There was a similar discussion in a thread back in March of this year. Much of that thread is relevant to the discussion here:
https://www.diyaudio.com/community/threads/midrange-enclosure-internal-shape.369316/
We did not reach a firm conclusion, by the way...
j.
https://www.diyaudio.com/community/threads/midrange-enclosure-internal-shape.369316/
We did not reach a firm conclusion, by the way...
j.
It's 4 drivers + 2 passive radiators for each vertical array of speakers. I don't want the speakers to become too big.
At the moment I am looking to make a v shape divider of the 2 vertical arrays. So each array of 2 drivers + 1 passive radiator will share the space.
My only concern is position of passive radiators. To make speakers compact I have to put them in the rear or down firing. Need to read through the other thread hifijim posted
At the moment I am looking to make a v shape divider of the 2 vertical arrays. So each array of 2 drivers + 1 passive radiator will share the space.
My only concern is position of passive radiators. To make speakers compact I have to put them in the rear or down firing. Need to read through the other thread hifijim posted
With these cross talk cancelation cross overs inner array of 2 speakers plays the music and outer array plays opposite speakers sounds out of phase to achieve this cross talk effect. Always happy to hear criticism of the design.
I built a pair of triangular boxes in the early 1990's using a jigsaw power tool, screws, screwdrivers and lots of glue. That is, there was a triangular bottom panel, three sides and a triangular upper panel (lid) with a downward slope. No idea if it actually helped against standing waves; I presumed it did, but had no way to quantify that.
In 2002 I sold them to a then colleague, whose wife didn't like them because she couldn't put a plant on top of them, because the flowerpot would slide down the sloped upper panel.
In 2002 I sold them to a then colleague, whose wife didn't like them because she couldn't put a plant on top of them, because the flowerpot would slide down the sloped upper panel.
What design criteria are using to establish the positions of the drivers on the front baffle?
Horizontally the centers of the drivers need to be more or less equal to distance between ears. That's what cross talk paper matt Polk wrote and most online sources recommend. So 7" -7.5". Positioning drivers further apart widens the stage. Bringing them closer narrows.
Vertically trying to make them as compact as possible. Tweeter is 3" x 4" approximately.
Passive radiators seem to play 180 degrees out of phase. I asked this question elsewhere. Is there a position of passive radiators that will cancel the 180 degree out of phase play . If they are on the same front baffle plane seems they playing out of phase. But if pointing to the back would this sort of align the bass waves with bass coming from drivers?
With 4 mids surrounding a tweeter you want them as close as possible. Ideally one wants to get to one-forth the wavelength at teh XO centre-to-centre. If this can be achieved then the drivers effectively become coincident and there is no lobing… as one gets further and further away from this ideal one gets more and more combing. Typically, because getting mids this close to the tweeter is difficut, you see an MTM arrangment which will comb/lobe in the vertical, not the horizonatl (this is eqivalent to Richard’s "Horizontally the centers of the drivers need to be more or less equal to distance between ears”).
Given a circular array around the tweeter, it will start lobing both horizontaly & vertically, the horizontal can get quite annoying.
You either need a really small bezel on the tweeter and small mids, or a T that can XO really low.
As the listening distance increase, the affect sof the lobing become less noticable. Except maybe the way room reflectiosn happen.
dave
Given a circular array around the tweeter, it will start lobing both horizontaly & vertically, the horizontal can get quite annoying.
You either need a really small bezel on the tweeter and small mids, or a T that can XO really low.
As the listening distance increase, the affect sof the lobing become less noticable. Except maybe the way room reflectiosn happen.
dave
Reading about acoustic lobing. And 4th of a wavelength
All 4 drivers 6.5" dynaudio 17w75xl
The frequency cut off for drivers and tweeter is 2000hz with these particular sda cross overs. 2000hz+ goes to tweeter, everything below 2000hz goes to midrange/midbass drivers. Some other sda models had 2500hz cross over points . Not sure if it's the cross over frequency relevant here.
Thanks you for this diagram. Helps with understanding
You might want to consider an arrangment of mids that works to kill the lobing.
Or egt far away. Or maybe it wo’t be bad enuff to give you issues.
dave
Or egt far away. Or maybe it wo’t be bad enuff to give you issues.
dave
Isn't the lobing effect that Polk sda effect tried to eliminate with cross talk cancelation? It seems like horizontal lobing is resolved. But vertical is introduced since I am replacing one 8ohm 6.5" drivers with a vertical array of 2 6.5" drivers
What if your proposed position of 4 drivers is rotated 90 degrees. Basically I making a horizontal sda effect+ vertical sda effect. Just random thought. Not sure it's gonna work though
It's a 'horses for courses' choice to meet specific performance goals, so numerous 'right' ways. If just defined as an acoustic ideal it would be a matching impedance on both sides of the diaphragm, so if the room had a net volume of 1000 ft^3, then so should the backside, or in approximate/practical terms, all point source speaker alignments should be near enough IB = > 10x Vas.
So got everything wrong
6.5" drivers start to lobe at 1285hz. If using this formula
f = (2 * c) / (π * D)
Where :
Trying to find out the vertical distance between the drivers to minimize lobing..
6.5" drivers start to lobe at 1285hz. If using this formula
f = (2 * c) / (π * D)
Where :
- f = frequency where the speaker starts to beam (Hz).
- c = speed of sound (343 m/s).
- D = effective diameter of speaker (m).
Trying to find out the vertical distance between the drivers to minimize lobing..