I didn't check this, even 86dB seemed pretty loud to me. When I set up the equipment again, I will try to measure also THD at 96dBspl.
There is no special bracing inside, except that chamber of the midrange driver is stiffening the box on one corner and the BR potrt, a piece of hdf glued on slats, does the same on the opposite corner. Internal modes are damped with polyester fiber filling.
The idea was that the directivity of a larger tweeter dome would be easier to match with directivity of the midrange driver, assuming no waveguide on the tweeter. As a side effect, the DI reaches 12 dB at 20 kHz, but that's not a big problem, considering that it's the edge of the band. The S-curve at 3-6 kHz (the effect of mismatched directivity) would be a bigger problem I think, and it was (partly) avoided.
What does the vertical response and contour look like?
Is there anything preventing the LF driver to be flush mounted as well?
There is actually no reason not to flush-mount the woofer, other than that DSA135-8 manufacturer's specification says "Cosmetic frame with low profile lip, designed for front mounting—no countersinking required!" which simplifies the cabinet construction.
Vcad, not measured right? Didn't want to try to get phase a bit tighter? Otherwise - very nicely done. Hard to believe you can get 40 out of that little woofer without a ton of distortion.
This is VituixCad simulation based on raw drivers measurements in the target enclosure (full orbits, 10deg step).
Extending the band towards low frequencies usually comes at the expense of sensitivity. My setup wasn't calibrated to measure sensitivity. Trying to estimate it basing on the manufacturer specs and filter transfer function magnitude I'm getting values ranging 80-83dB (so it's not a very accurate method).
How about phase? I am asking because I am learning. I would be trying to get the phase more alighned. Is that necessary ?
I'm still learning too, the approach I'm currently using involves reaching a compromise in balancing the system's SPL vs PIR and PR characteristics, and if this results in some phase mismatch, I just let it go.
I would think estimating from mfg specs is probably pretty accurate. Your woofer is 87dB and you probably have 5-6dB baffle step compensation, so you are almost surely at 81-82dB sensitivity. Whether it is actually 81dB or 82dB seems pretty irrelevant.
Here is phase of tweeter and midrange one octave above and below the crossover point circled in red and phase of woofer and midrange one octave above and below the crossover point circled in blue. I don't see anything concerning. You could have disabled the pink target curve in VituixCAD and it would have been a little easier to see.
The woofer isn't going to measure differently flush mounted or not, but your tweeter is likely diffracting off the woofer frame edge some (and maybe the surround a little more than if it was flush mounted). But if I was building it, I wouldn't be concerned about it one way or the other.
