As I said, this is a question about C-C. Imagine a baffle height of 60 cm, wherein you need 28 cm for a 10-inch driver, structural support at top and bottom of the baffle, let us say 2 cm each, then 28 cm are left for the waveguide. You are restricted to C-C of 28 cm. If XO frequency is supposed to be optimised at 1-1.2 lambda, this forbids a low XO frequency. And so on ...
Why don't you do your own c-c investigations. I understand it isn't simple. In my experience there was nothing available to help, I had to do the calculations by hand, by writing my own software, wavefronts in CAD etc... it would take months on that part of a crossover. Therefore I wouldn't take it lightly to say hey, all I have is 60cm. I'd give as much as needed, and if that's all I had, I'd have to do it all over again.
Instead, we hear this 1-1.2 rule of thumb that few seem to know the reason for. I looked at it once and I noticed that the person who started it didn't use some of the available techniques for optimising it. Therefore, if it works, then it isn't the only option that should work.
Instead, we hear this 1-1.2 rule of thumb that few seem to know the reason for. I looked at it once and I noticed that the person who started it didn't use some of the available techniques for optimising it. Therefore, if it works, then it isn't the only option that should work.
You can always bite into your axi-symmetric horn with the driver to decrease the CC.
Something I've been pondering about is if it's possible to design the crossover and CC distance in such a way that the cancellation matches the floor reflection at the listening position.
For a certain height of the speaker and distance from it.
Something I've been pondering about is if it's possible to design the crossover and CC distance in such a way that the cancellation matches the floor reflection at the listening position.
For a certain height of the speaker and distance from it.
^ That is exactly why one would want to play with the c-c spacing, mind about how vertical early reflections affect sound at listening position. Besides c-c distance height and number of transducers in a system in relation to bandwidth they play, room height, listening height and distance can be adjusted. Watch out not to introduce unintended side effects like biting into perfectly nice waveguide could possibly do
And don't forget about the ceiling!
Adjusting c-c is a way to affect frequency response towards the first vertical reflection points around crossover, thats all. One tool to affect and balance things out for given application.
And don't forget about the ceiling!
Adjusting c-c is a way to affect frequency response towards the first vertical reflection points around crossover, thats all. One tool to affect and balance things out for given application.
Last edited:
Looks like there is a high q= value in that contour which makes the overall diameter quite large, reduce that down and you won't lose too much but you may need to tweak the other parameters to compensate.
This post gives a simple way to check the assumptions. No need for months of work or writing software.
http://www.htguide.com/forum/showthread.php?44128-VituixCAD-v2&p=639916&viewfull=1#post639916
You seem to ascribe more to it than kimmo does himself, in no way is it the only option or even the best if the C-C can be below 1/3 wavelength, but in a lot of practical situations it works out to be a good option, but only for in phase crossovers. Worth thinking about and trying out at the very least.
For mathematical foundations, fluid probably has the best overview on Kimmo’s postings on the matter so far. For someone trained in sciences/engineering, it must be easy to understand the relation of radiating pattern vs. takeover frequency from there. It is spacial frequency summation and cancelation, and can be observed by the layman (i. e. me) in VCad.
(e: Fluid was faster)
Increase.
Limiting the vertical radiation to an extend that avoids floor and ceiling reflections requires too much length when combined in this sort of radiator. At least I don’t know of a good way how to keep the lengths of the vertical walls the same as horizontals, when it’s angles are steeper. They’d need to protrude.
In this case, the Idea was really about reducing the vertical dimension of the waveguide, without damaging its performance too much and maybe limiting vertical a little bit on the way.
To me the best way is to use a single source for the largest possible bandwidth. It could be a MEH but I have yet to see or hear one without other drawbacks. That's why I'm putting quite a lot of hope into the ESP as it could allow to use a single source maybe 500 - 15k without a glitch, with a suitable driver (perhaps even a direct-radiating one).
Last edited:
And I am look forward for this. It will need a wide diameter waveguide, though, implying: a bigger speaker. I really set my mind to compromise on “optimum” performance and build to a size that I find beneficial in my living environment. Speakers like JBL 708p and Genelec S360A show a very high performance can be achieved.
I thought I was already restricting it a lot with a q = 0.991 and introducing a good amount of ripples. But I will play with more abrupt terminations after returning from Xmas travels.
That isn't very high, maybe a high s value or something else, the transition to flat baffle looks to be quite drawn out which isn't ideal when you are trying to squash it in
One more note: the commercial waveguide I started with has an abrupt termination and effective size of about 26.5 by 16 cm. The small vertical dimension leads to an early loss of pattern control and I thought this was the whole reason for excessive of axis energy above takeover frequency. But as Kimmo has demonstrated, this is not only the effect of pattern control, but positive summation. This means even a fairly controlled waveguide that can work down to i. e. 1.2k will introduce an off axis energy excess. And the effect is audible anyway: I have equalized the on-axis frequency response in this area with the current version, because it sounds to hot/rough if it’d be flat.
Correct, it is s and n that have been modulated on the vertical axis (using: x*sin(p)^2) to balance off the changing effects the parameters have due to change in coverage angle around the axis.That isn't very high, maybe a high s value or something else, the transition to flat baffle looks to be quite drawn out which isn't ideal when you are trying to squash it in
Are you sure? Kimmos post is a helpful starting point, an approximation. I don't think it's meant as a complete solution and it skips some important steps.
Don't forget, the point I was making was the amount sheeple was counting on flexibility.
That depends heavily on the crossover - overall, the total radiated power ("off-axis energy") can still be flat, even though there are vertical lobes.
I understood you that by extending the passband of a single source, through a phase plug of a low extended compression driver in this case, you want to optimize the passband wherein on-axis/off-axis are equally controlled. Then it is important what size of the waveguide is needed to fulfill this task at i. e. 600 Hz.
I'm sure that it is enough to demonstrate the point he is trying to make. A basic analysis of ideal drivers and varying the CTC and crossover frequency shows how the sources interact well enough to illustrate it. If you want to elaborate on the other complications you mention it might make it clearer what point you are trying to make because that is the not clear to me.