Alignment tool don't correspond to measured curve

[lillnysta]

Yesterday all was good, did a nice curve and it align well with the simulated aligned curve and got a nice flat curve from the EQ.
Today I was starting all over (don't ask me why.) But can't seem to get the Align Sum curve and the measured curve to match. Something must be terrible wrong with my set up some where.
Please help me, been experimenting hours with this now and can't get it to work like it did yesterday.
Blue aligned sum and purple measured.

My settings:

Levelmatched both subs with the L speaker on 75db


MiniDSP

 

[sm52]

Check polarity in all places. In programs and physically.

 

[lillnysta]

How to check? Newbie at this, sorry. everything is connected with a LFE cabel, the subs for themselfs measures exactly like yesterday, but when i combine them I got this massive drop, and the drop can be a little bit different between the measurements aswell. The big null disappear when i invert one of the subs, but then the graph looks worse in other places. Totally lost at this now, was so happy yesterday when the curve look soo good.

 

[sm52]

It's hard to find the answer. In REW, inversion marks can be made in several places (at least 6). MiniDSP can also invert channels. Some other filter could have been created by accident, which instead of raising the frequency response makes a drop of the same level.

 

[lillnysta]

When I raise my SVS sub that is in the front left corner, the null disappear, the more i raise the volume the more the null raises. +5db and it is gone.
So strange!
Maybe try another position for that sub?

 

[lillnysta]

***Solved***
After using Timing Preference and align both subs, the problem with the big null is gone and the curve nog replicates the alignment tool curve much better.