OK, glad to hear the Qtc will be tweaked to an acceptable level. Some tips...the plywood construction is preferred for the reasons noted, however, finding the correct type plywood maybe difficult. You want plywood that is "void free"...those without air-spaces in the layers, it can be expensive. Further, your dimensions as I see it are a cubed shape....this is a no-no..as the symmetry of a cube can result in frequency "pile ups", frequencies within the box "adding up", matching wave-fronts reinforcing each other, an additive effect. Various ratios exist, my favorite... 0.7937:1:1.2599 and 0.618:1:1.618 or any such combination...within reason of course.
-------------------------------------------------------------------------Rick...........
-------------------------------------------------------------------------Rick...........
ie use irrational numbers. Beware √2, if you use it twice then you get the long dimension twice the smallest.
It should be noted that with subwoofers, even the longest dimensions are smaller than any frequency generated and thus the usual proportion rules can be ignored.
dave
Love the cube
I now only build cube shaped enclosures. I spent time with boxnotes software
Boxnotes - Free Speaker Design Software
to see what frequencies my box would resonate at. When you stray from the cube shape the lowest resonance for the box goes lower and lower, right into the pass band of your driver where you don't want it. Building small enough cube enclosures you can have a resonance free response. To get the cube as small as possible you can let the installed Q get up to 0.8. This can then be compensated out using the Linkwitz transform to what ever you want with an ideal response range and Q. Because the Linkwitz filter cancels out the response of the speaker and replaces it with an ideal response, it will give better performance than can ever be achieved by simply adjusting parametric equalizer parameters.
The small panel dimensions of the cube also help raise the panel bending resonances so that they can be moved out of the woofer operating range as well.
I now only build cube shaped enclosures. I spent time with boxnotes software
Boxnotes - Free Speaker Design Software
to see what frequencies my box would resonate at. When you stray from the cube shape the lowest resonance for the box goes lower and lower, right into the pass band of your driver where you don't want it. Building small enough cube enclosures you can have a resonance free response. To get the cube as small as possible you can let the installed Q get up to 0.8. This can then be compensated out using the Linkwitz transform to what ever you want with an ideal response range and Q. Because the Linkwitz filter cancels out the response of the speaker and replaces it with an ideal response, it will give better performance than can ever be achieved by simply adjusting parametric equalizer parameters.
The small panel dimensions of the cube also help raise the panel bending resonances so that they can be moved out of the woofer operating range as well.
Sorry, but this isn't the case.
A Linkwitz Transform is a line-level EQ curve which can be emulated just fine by a parametric curve. The difference with an LT is that it's designed for a particular Fc and Qc, and has a precise curve to compensate for the sub's innate frequency response.
That's all an LT is, though - an EQ curve to replace one Fc and Qc with another, via electronic EQ.
Chris
Depends on what is "just fine"
Mathematically if is was possible to generate the exact LT transfer function by adjusting the parametric EQ, you are correct. If the parametric EQ is simply adjusted to get an approximate flat gain response, the phase will response will not be as good as simply applying the LT. The beauty of the LT is you still have the response of a second order system when you're done, it simply has the desired lower cut off frequency and desired damping ration. You start twisting knobs on the parametric EQ and you have a 4th or 6th order system with 2 or 3 resonant frequencies and damping ratios combine to make an response.
Mathematically if is was possible to generate the exact LT transfer function by adjusting the parametric EQ, you are correct. If the parametric EQ is simply adjusted to get an approximate flat gain response, the phase will response will not be as good as simply applying the LT. The beauty of the LT is you still have the response of a second order system when you're done, it simply has the desired lower cut off frequency and desired damping ration. You start twisting knobs on the parametric EQ and you have a 4th or 6th order system with 2 or 3 resonant frequencies and damping ratios combine to make an response.
What damping do you enrol with EQ or LT? You can damp with motional feedback but not, as the term is ordinarily understood, by manipulation of the one-way signal.
I accept as a mathematical truth that various signal adjustments can compensate for the measured resonance imperfections of a speaker in running an FR and in playing sine waves (which I assume assumes the speaker has no harmonic distortion and all kinds of other covert assumptions too).
But my intuition says you still have a speaker in your system that will have ringing that (might) be audible.
Will some of the Fourier transform advocates educate me please.
B.
Last edited:
I try to hold my opinions with a strength that is proportional to the weight of evidence supporting them, so that they can be easily changed with arrival of stronger evidence.
My intuition matched yours exactly. I went and played with simulations and saw it work there. I then built a system and tested it and the measured response was as predicted and the sound was fine.
It is response cancellation with cascaded filters verses pole placement using feedback. So everything must match for it to work perfectly. As the system becomes non linear operating beyond Xmax things will shift, so it's hard to say what might become audible. Of course things are pretty loud at that point.
Much appreciate your intelligent and courteously worded reply and doubly appreciate the empirical testing.
I've been thinking about tests for the last year.
The Fourier-equivalent displays in REW are not "tests" in the usual sense. They are what might be called "a sample of one" in that the single FR data collection is convoluted and transformed back and forth. In a sense, they "buy into" the Fourier model and thus can't critique it.
One the other hand, 'scoping square waves would be considered an independent test.
I wish I knew enough to go beyond just intuition.
B.
In the region that is being EQed the room dominates. Given that any textbook “solution” is very likely not to be the best solution. The room needs to be “measured” and the levels, phase, etc of all the woofer sindividually adjusted to get a nice smooth bass with few anomalies.
An LT trades box size against power & excursion usage, and in room will still need EQing. By using intelligent PEQ one skips a step.
dave
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.