I made this LONG post over at AVS and I thought I'd pull it over here and add to it over time. Many DIYers are new to modeling subs so I thought I'd keep a "guide" of sorts. This is a start.
Point #1. Use the right scale! Don't look at a frequency range your not going to use. Most T/S modeling programs default to 10-1000Hz. Since we are building a sub, I don't give a hoot about anything over 100Hz.
This is too much information!
This is where we want to look.
Point #2. Traditional box alignments: Don't be traditional! If you go by the numbers, and let software optimize a box size for the Shiva-X it will come out large in most cases. Lets look....
I used the LSPCad box optimizer, took all the default values and loaded up the Shiva-X. Here is what LSPCad suggest for alignments:
SBB4:
QB3:
SC4:
You will notice the box sizes, with the exception of the SBB4 alignment are all large. The QB3 & SC4 are both > 200L so the "traditional" alignments suggest the Shiva-X is a large box driver.
Pish Posh.... ignore the traditional alignment suggestions. I don't have the time or space here to go into the historical/mathmatical reasons for these traditional alignments but there is no reason to limit ourselves to a them. These are just points on a curve and all that really matters to us is getting good bass in your room.
Point #3.
Use Polyfill! I'm not sure why people are so hesitant to use fill in ported enclosures. You gain a good 15-20% MORE effective volume. The downside? You have slightly more box loss but its not near as meaningful as the extra effective volume you gain. It provides some marginal help with internal resonance issues also. The only other downside is cost but a 5lb. box of polyfill at Walmart is cheap. Go down the isle with the old ladies and pick some up because its going to help keep your box size manageable and your wife happy.
Oh... how much polyfill should you use? There are different numbers floating around out there but from my experimentation, I'd say 12oz. of fill per cubic foot of enclosure volume is just about equal to 100% fill.
Point #4.
Don't fret over minor differences! Differences of 1dB or so are completely swamped by the room contribution. You also have to remember, that this is a simulation and there is room for error and tolerances. The T/S parameters are small signal level parameters and the changes in T/S are drastic at high signal levels. The simulation is not real...... its just an approximation of what your going to get so don't get too caught up in small differences.
LSPCad has the ability to simulate some of the high signal level non-linarities. Once again... these are estimates that are based upon some "assumptions" made by the programmer of the software. These are not absolute but they give us some idea of how the non-linear driver behavior will effect the output.
Here is my recommended 4 cubic foot box (113L) with 100% fill and an 18Hz tune as you see in most simple T/S models.
Now lets model this same alignment with power compression, suspension non-linearities, inductance non-linearities, port non-linearities and BL compression with the driver under 700W of power.
Kind of ugly compared to our simple model huh? This is an exaggeration, the 700W power compression is not representative of what you actually see in real life. Music/movies is about a 30% duty cycle and the graph above represents a 100% duty. The effects of power compression in real life are much less. I'm just trying to prove a point though. The nice graph that WinISD pumps out is NOT what your going to get in real life. There are other factors at play so don't get caught up in small differences, they will be swamped by other factors in real life.
Point #5.
Understand the ROOM! The room is by far and away the biggest variable in this game. If you have an enclosed small room you get significantly more gain under about 30-40Hz than you do in a wide open space. Even in wide open spaces though you get some cabin-like gain under 30Hz. How much varies but you need to keep it in mind when your designing your box.
Here is our 4 cubic foot box tuned to 18hz. The anechoic F3 is around 20Hz. Now lets stick this baby in an average sized HT room. The room dimensions are in the simulation but for you Yankees, its 21ft deep, 14ft wide with a 9fth ceiling. The sub is in one of the front corners and the listening position is seated at 39", centered on the room about 2/3 back from the front wall.
In this situation, you get tons of gain from 40Hz down. In real life the amount of pressure vessel gain is going to vary substantially room to room. Even in situations where you have a completely open house with minimal room gain your going to have gain. Plan on, it and build your sub to account for some of the low frequency gain that your going to see.
Point #6.
Keep your priorities straight! I know everyone like to look at 10hz extension and dream about how your sub is going to perform with those handful of movies that have extremely low content. In real life though, most of the action is above 20Hz. If you like organ music and want to cover about 99% of what is on the disc, look at your subs output from 16-18Hz on up. That is where its really going to spend most of its life. Don't design around the exception, design around your main usage.
Here is the suggested 4 cubic foot enclosure with an 18Hz tune from 17Hz on up.
You will see in the room graph that you get a nice broad hump, centered on about 25Hz in this case. Keep in mind your going to be using a low pass filter with a corner anywhere from 40-80Hz and that the final response will end up looking like a large hump, and the sub will transition to the main speakers somewhere in that range.
Thats how I like to design them. You get good in-room integration into the main loudspeakers area of performance, excellent extension into the teens without too much of a 20Hz hump and the box size is relatively small and manageable. The 18Hz tune is achievable with reasonable length ports, about 43" for a 6" diameter port in this case.
For the dual box configuration you just double the volume, and increase the port size to handle the extra output.
