Originally Posted by Ricci
I don't really see anything wrong with the measurements of this system. Other than there are none from the listening position. This driver is probably a version 1 working just fine. The tummie exhibits about an 8dB roll off to 100Hz from a peak near 40-45Hz in a 100L sealed box. He shows a few DB more roll off but as I have mentioned earlier a close mic does just this so that is consistent with what you would expect. .
Do you have any examples of a close mic in free air showing a steeper roll off than the driver in a sealed box at GP? I haven't observed that to be the case... ever. Also, as I noted earlier, my close mic measurement of the Tumult in 100L shows the same roll off on the top end as Ilkka's GP and actually a more accurate roll off (2nd order) on the bottom than Ilkka's (10dB/octave).
In free air, the measurement should accurately reflect the Fs and the roll off at the top should be notably less than when in a sealed box, not more.
Also, he shows a heckuva lot more roll off in-box than a couple of dB. JJ's own measurements of the AV15H show a couple of dB drop at 100 Hz, but the OP's comparo shows a 12dB difference between box w/Tummie vs box w/AV15H when it should be 1/2 that at most.
To the OP, in your DATS exercise, you entered 13.75" for driver diameter. The Tumult is 12.2", according to Adire's published Sd and my measurement. That would change the result for the Vas and sensitivity calculations quite a bit, as far as modeling goes. Try the proper diameter input and use the result to try a new model.
IOW, it's impossible to make much of the posted data because it's not accurate. That means I disagree with Josh's assessment that there isn't anything wrong with the posted measurements of the system.
If you're not willing to whack together a 100L box (or use a larger one and displace the proper amount of interior volume to make it 100L) and measure close mic for a comparison against a known reference, which is the simplest way to get pointed in the right direction, then all that will happen here is a lot of useless speculation.
The posted data doesn't crunch. Regarding the model, it's incorrect T/S input. Regarding the measurements, is it the measurement rig, the driver, the box, the PRs, etc? No way to know from what's been posted, IMO.