Hi everyone,
I am starting a project to design an audio system for my home by reusing different woofers that I have taken from different speakers, systems, etc.
I read a student's thesis from 2009 on sound engineering and acoustics, and on the calculation and design of an acoustic box for a JBL woofer.This paper helped me a lot to understand the calculations behind the design of acoustic boxes, and to be able to calculate a box for my woofers.
The problem is that the design exemplified in the thesis is about a box for a single woofer, and I have several woofers, with different Thiele-Small parameters, which come from very different audio fields, so I don't know how to calculate a box. to be able to have all the woofers in the same box, or if it is not possible. I also don't know how I can see the theoretical response of the box with the woofers, since in programs for this, like WinSID (which is the one I use) you can only see the behavior of one or more woofers, but of the same model, not of different models.
So my main doubts are, how can I calculate the same box for the same woofers, given that they are different models, and with different Thiele-Small parameters. And my other question is also how can I see the theoretical response of the box?
I would also like to be able to better understand the Bode plot implemented for the responses of the woofers, since I do not understand it 100%
Any kind of advice, tips, etc. I would also greatly appreciate it.
PD:I don't speak English that well, so all this writing was translated by Google Translate, in case there are strange things in the writing, hahaha
I am starting a project to design an audio system for my home by reusing different woofers that I have taken from different speakers, systems, etc.
I read a student's thesis from 2009 on sound engineering and acoustics, and on the calculation and design of an acoustic box for a JBL woofer.This paper helped me a lot to understand the calculations behind the design of acoustic boxes, and to be able to calculate a box for my woofers.
The problem is that the design exemplified in the thesis is about a box for a single woofer, and I have several woofers, with different Thiele-Small parameters, which come from very different audio fields, so I don't know how to calculate a box. to be able to have all the woofers in the same box, or if it is not possible. I also don't know how I can see the theoretical response of the box with the woofers, since in programs for this, like WinSID (which is the one I use) you can only see the behavior of one or more woofers, but of the same model, not of different models.
So my main doubts are, how can I calculate the same box for the same woofers, given that they are different models, and with different Thiele-Small parameters. And my other question is also how can I see the theoretical response of the box?
I would also like to be able to better understand the Bode plot implemented for the responses of the woofers, since I do not understand it 100%
Any kind of advice, tips, etc. I would also greatly appreciate it.
PD:I don't speak English that well, so all this writing was translated by Google Translate, in case there are strange things in the writing, hahaha
Putting woofers of different T/S parameters in the same air volume would not be something I would contemplate.
I suppose you could build a tower speaker with multiple compartments, each compartment being designed to accommodate the T/S parameters of the particular woofer that will occupy that compartment.
P.S. The English read just fine.
I suppose you could build a tower speaker with multiple compartments, each compartment being designed to accommodate the T/S parameters of the particular woofer that will occupy that compartment.
P.S. The English read just fine.
Greets!
In theory you can sum the driver specs and find the mean specs to design by, which from experience can work well with two very different drivers, but a whole bunch? I'm guessing it will force you to only use a lower max power than the mean and of course you'll probably need DSP to frequency shape the results.
Regardless, I say go for it and look forward to how it turns out.
BTW this was a subject of hot debate decades ago for a two way with no XO and it turned out great.
In theory you can sum the driver specs and find the mean specs to design by, which from experience can work well with two very different drivers, but a whole bunch? I'm guessing it will force you to only use a lower max power than the mean and of course you'll probably need DSP to frequency shape the results.
Regardless, I say go for it and look forward to how it turns out.
BTW this was a subject of hot debate decades ago for a two way with no XO and it turned out great.
Thanks for the answers!
Galu, regarding the tower with several compartments, had considered it, but the ideal volume for maximally flat responses is already quite large, so adding all the volumes, in different compartments, would be a problem for me due to the size. Maybe there is a way to calculate how to shrink the enclosure without having to vary the answer so much, but I don't know how to do it.
And GM, I was interested in what you said, about taking the average of the parameters of each woofer, and then using that average to calculate the enclosure, but how would I do that, average all the parameters? Qes, Qts, Qms, Fs, F3, Fts, Mms, Cms, Lp, Sd, Dd, etc...?
And if I want to divide the frequencies between the different woofers that I have, do I also have to average them between all of them, or do I exclude the ones that I am going to use as a tweeter from the calculations, and only include the mid-range and bass ones? Or do I calculate mids and lows also separately?
And then, with that average I took, I calculated the enclosure as if it were a woofer, right? And I do all the calculations and see responses in WinISD as if it were that only woofer?
Thanks in advance,
Greetings!
Galu, regarding the tower with several compartments, had considered it, but the ideal volume for maximally flat responses is already quite large, so adding all the volumes, in different compartments, would be a problem for me due to the size. Maybe there is a way to calculate how to shrink the enclosure without having to vary the answer so much, but I don't know how to do it.
And GM, I was interested in what you said, about taking the average of the parameters of each woofer, and then using that average to calculate the enclosure, but how would I do that, average all the parameters? Qes, Qts, Qms, Fs, F3, Fts, Mms, Cms, Lp, Sd, Dd, etc...?
And if I want to divide the frequencies between the different woofers that I have, do I also have to average them between all of them, or do I exclude the ones that I am going to use as a tweeter from the calculations, and only include the mid-range and bass ones? Or do I calculate mids and lows also separately?
And then, with that average I took, I calculated the enclosure as if it were a woofer, right? And I do all the calculations and see responses in WinISD as if it were that only woofer?
Thanks in advance,
Greetings!
This depends on the woofer you use.
If you want to use many random drivers that's fine, but if you're trying to get a response so specific with them you're going to have to work harder.
I'd consider using EQ.
What do you want to know?
Me gustaría saber si lo he entendido bien. Lo que entiendo es que en el eje Y están los dB o volumen, y en el eje X las frecuencias, por lo que donde la línea del gráfico esté por encima de -3dB, se escuchará esa frecuencia. Pero no sé por qué la mejor respuesta es la máxima plana.
Are you having trouble with google translate?
Approximately maximally flat helps you to get more bass and it allows it to be presented smoothly.
Yes that is generally correct. Q=0.7 is about the average, a little more or less is OK. You can still hear below -3dB.
Approximately maximally flat helps you to get more bass and it allows it to be presented smoothly.
I understand,
What do you mean by Q=0.7, a little more or less is fine?
And if below -3dB you can still hear, why is it generally said that this is the cutoff of the frequencies?
Why does the maximum flat give more bass? What happens if I have a dB peak up to 2dB at 50Hz, then drops to -1dB at 70Hz, and then continues rising flat, wouldn't there be more bass in a response of this type than in a completely flat one? (the situation I raised is an example)
What do you mean by Q=0.7, a little more or less is fine?
And if below -3dB you can still hear, why is it generally said that this is the cutoff of the frequencies?
Why does the maximum flat give more bass? What happens if I have a dB peak up to 2dB at 50Hz, then drops to -1dB at 70Hz, and then continues rising flat, wouldn't there be more bass in a response of this type than in a completely flat one? (the situation I raised is an example)
Maximally flat means Q=0.7
It is the half power point.
More upper bass but less lower bass.
what do you mean by this?
Ah ok, thanks!Maximally flat means Q=0.7
I don't understand
Below this point, less than half the power is delivered compared to the passband.
If I was going to embark on a project that involved multiple woofers of different kinds, I would expect at the outset
that the electronics/circuit would be unusual and possibly complex.
My first step would be know (or approximate) the power capacities of the woofers, then flatten their impedances
using Zobel networks and then use Power Resistors in series to distribute power drive appropriately.
I would then figure out a wiring arrangement that resulted in an impedance of 4 to 10 ohms.
I would then create a speaker box of my size liking and install my woofers > and then start listening to them.
At this point, I would ascertain whether I was 'going 2 way or 3 way'.
After this I would go about standard methods of crossover design realizing that different component values
may be required for different woofers.
With plenty of TRUE DIY experimentation, I'm sure that reasonably decent sound can be achieved.
* Theil/Small parameters for this kind of project are in my opinion > 'Out The Window'.
that the electronics/circuit would be unusual and possibly complex.
My first step would be know (or approximate) the power capacities of the woofers, then flatten their impedances
using Zobel networks and then use Power Resistors in series to distribute power drive appropriately.
I would then figure out a wiring arrangement that resulted in an impedance of 4 to 10 ohms.
I would then create a speaker box of my size liking and install my woofers > and then start listening to them.
At this point, I would ascertain whether I was 'going 2 way or 3 way'.
After this I would go about standard methods of crossover design realizing that different component values
may be required for different woofers.
With plenty of TRUE DIY experimentation, I'm sure that reasonably decent sound can be achieved.
* Theil/Small parameters for this kind of project are in my opinion > 'Out The Window'.
Yes, I probably didn't have to sim it just now but I seem to recall last time I went looking for this same information I didn't like the plots. Some don't show the change in frequency, some might need to be explained since they have extra unneeded information.
In this one you can see 0.7 has the lowest f3
In this one you can see 0.7 has the lowest f3
And what do you think about what GM says above in the thread? What it says about averaging the parameters of the woofers, and based on that averaging, making the corresponding calculations
TL;DR,
Absurdly Ambitious Project #1, problem resolved in theory, thus I am now eagerly awaiting delivery of the KABD 4100 and its bits, due this morning, 15Jun24. Having returned to continue reading up on the KABD DSP docs,
the first thing I see is your OP. A well-communicated concept that made me smile ear-to-ear.
There are no coincidences; having bought only one DSP amp kit for my acoustically problematic project [Absurdity1],
I was teetering on the precipice of ordering another x4100 kit for [Absurdity2],
but was intimidated by some of the maths required to proto-mangle my nonsensical idea into anything near a minimal standard of /"Acoustically Meh"/.
That, and my inner Voice of The Ignorant,
"nah, I cannot begin to articulate it well enough to seek forum guidance about this unholy Frankenwoofer mutant enclosure idea"....
—the exact same idea—
about which you articulated very well, and in a secondary language, at that.
I salute you, my brother in Christ, for leading the way.
You made my day, man, and the sun isn't even up yet.
Uncanny, the cosmic stream and kindred souls. I will be going to school on your idea. Many thanks, and Godspeed!
Absurdly Ambitious Project #1, problem resolved in theory, thus I am now eagerly awaiting delivery of the KABD 4100 and its bits, due this morning, 15Jun24. Having returned to continue reading up on the KABD DSP docs,
the first thing I see is your OP. A well-communicated concept that made me smile ear-to-ear.
There are no coincidences; having bought only one DSP amp kit for my acoustically problematic project [Absurdity1],
I was teetering on the precipice of ordering another x4100 kit for [Absurdity2],
but was intimidated by some of the maths required to proto-mangle my nonsensical idea into anything near a minimal standard of /"Acoustically Meh"/.
That, and my inner Voice of The Ignorant,
"nah, I cannot begin to articulate it well enough to seek forum guidance about this unholy Frankenwoofer mutant enclosure idea"....
—the exact same idea—
about which you articulated very well, and in a secondary language, at that.
I salute you, my brother in Christ, for leading the way.
You made my day, man, and the sun isn't even up yet.
Uncanny, the cosmic stream and kindred souls. I will be going to school on your idea. Many thanks, and Godspeed!
However, it could be an interesting exercise just from the point of view of seeing what might happen.
Last edited: