Maximally-flat frequency response, being the default answer from most box design softwares, it often leads to a curious looking almost "banana shape" curve for lower Qts (< 0.383) drivers in QB3 alignments. Each LF alignment has its pros & cons, however with regards to frequency response cuve the B4 alignment is something desirable to try to approach. It cannot be a real B4 type alignment since that would require a driver Qts of 0.383. With little will and imagination we can get closer to it's curve, starting from a QB3.

If you're not familiar with QB3 and B4 alignment response curves, they can be found easily on the Web, or in Bullock on boxes, 1991, p. 4, to visualize what I'm explaining here. He says "The main differences between a B4 and a QB3 are that a QB3 has more "droop" in the passband and its cut-off frequency his higher." (p. 5)

Since these softwares initially seek to design for max flat response, the response curve you get may not necessarily look uniform, particularly in the low-end. It is your job to vary the box parameters (Vb, Fb) in the software until such time you reach a proper box result with the driver you have in mind. Here's what John Eargle (JBL) says about low-frequency alignments: "What we are aligning is, of course, the driver and the enclosure both with their own resonance frequencies in a coupled system. (Assuming the enclosure is in a half-space boundary condition). There is a continuous range of alignments possible with any driver and enclosure, but Thiele and Small have labeled certain of the possible alignments ..." John M. Eargle, Loudspeaker Handbook, Chapman & Hall, 1997, p. 58."

Experimenting with the software (within reason), as well as with other software to double check results if need be, is part of the speaker box design process to find what's the best scenario. This has to be logical also because too much Vb or Fb or not enough creates problems. I don't worry much about a 0.5 db positive Ripple (bump in response), though I prefer to avoid negative Ripple (dip in response) leading to weaker bass. When you have a response curve that suits you, see what it means in terms of Ripple, Xmax vs input power, Vb/Fb/F3/Vent dimensions vs box dimensions. Acceptable or not? Then make some changes as required.

As I write this, I see in front of me a graph with a response curve done with a JBL 2226H in a QL 7 box tuned to 40 hz. From that graph it can be seen the output level in the 40-50+ hz range is about 3 db lower compared to the level from about 250+ hz. The ear can easily distinguish a 3 db level difference, and this box may very well sound mid-bassy, which I don't like, unless floor, wall or corner placement is used to boost the lows. Also, to me that type of dip in response curve is often a sign of too big a box and/or tuned too low.

My aim is to achieve a more straight line response by raising the low-end dip in output level (in the 40-50 hz or so range for the above example ) with tuning the box a little higher, while avoiding too much Ripple (< 1 db). That way making the speaker sound more balanced (less mid-bassy). Seen otherwise, getting away a little from the "max flat" QB3 curve in order to get a little closer to the B4 type curve, I'm modifying the classic max flat QB3 alignment (for sort of a "hybrid") by increasing slightly box tuning frequency.

Starting with the Vb/Fb given by the software with max flat default alignment, then I do things MY way in the software (i.e. refine Vb and/or Fb number) to flatten response some more regarding different output levels in the lows. Usually the higher of two or three Fb's considered not only brings me closer to a B4 type of response curve, but also gives me a little more loose on cone travel (Xmax) VS maximum input power.

Fb "optimization" for a more linear curve, is NOT a magical cure to correct a largely wrong box volume. It cannot be stretched far to avoid other problems, since box tuning is fine tuning the driver/box combo, not re-defining box volume.

"Re-working" QB3 alignments that drop too fast in terms of low-end response compared to B4 alignments, is not an absolute must have or religion to be followed to the letter. Aligning the lower frequencies levels (for less dip) with the mid-bass frequencies level is rather a nice to have (trick or tip) for balancing (trying to match or so) driver output levels at different frequencies to minimize hearing predominantly mid-bassy sound. While having a minimal Ripple (< 1 db) and a bit more headroom on SPL before reaching Xmax, if that is what you like.

Tuning refinement also works the other way. The use of a lower tuning frequency usually means a dip in LF response which is useful to minimize a bump. In a little too small box with peaking low-end response, you may want to try reducing the tuning frequency a bit to flatten somewhat that peak. Be aware however, that with lower tuning comes inconveniences: first, an issue of more strain put on the woofer with a minimal, small or larger increase in cone travel, depending on how much lower you're tuning and on the sound level used. Its better to reach Xmax later than sooner re distortion. So, proceed with caution to avoid driver damage. Second, lower tuning will require a little longer vent tube of the same diameter. You may or may not have enough box depth for a little longer vent, but plastic piping with an elbow may help. If you are still at the computer design phase in software, try this other tyrick: a slight net volume increase to the proposed box and a bit higher tuning frequency will shorten that new required vent.

As standard practice, I model all of the above for my boxes in Win ISD Pro and Winspeakerz to decide on an Fb, comparing the effects on response and cone travel, in chosing between a few logical tuning frequencies (e.g. 40 VS 42 VS 45 hz).

Richard