Akabak 3 tuning vs. reality

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I've read through the Akabak thread and not found answers to the 2 problems I'm having with Akabak 3.
First a little background. I'm working on a small surround with 2 round ports. Depending on where I place the ports, I get different tuning, so I thought this would be the perfect test for Akabak, since it should do a better job of predicting. I built the box and took measurements with the ports in various locations, then wanted to use Akabak 3 to simulate. But I have 2 problems. The first, and largest, is that the tuning is way off. Here's the frequency response (yellow is measured, black is Akabak model):
FR.jpg
[/URL][/IMG]
But here's the impedance:

Nothing I do seems to affect the lower humps or the tuning. I've measured the driver on Klippel and am using the data (tried both the "Klippel" setting and by curve-fitting within VACS).
After a couple days banging my head against a wall, I started fresh with a different box and driver that I had data on. I modeled in SolidWorks, meshed in gmesh and used Akabak. Same problem. The tuning is off by 8Hz and moving the port has no effect in the model, even though it does in real life. Shouldn't a BEM model be able to do a better job of predicting tuning than WinISD?
The second problem is minor, but also frustrating. The voice coil excursion is always zero. I even opened all the examples and I get zero excursion in all of them. Is this feature just broke? It worked in ABEC.
Thanks.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.