@Anglofun think about this for a minute. As a whole, a multi way speaker's sensitivity will be constrained by the least sensitive driver. Similarly, that speaker's dynamic linearity will be constrained by the driver with the least dynamic range. One of the drivers will hit it's output limits first, unless by happenstance they all have the same limits.
If you linearize the whole speaker, it will result in dynamic linearity, up to the point where the driver with the least clean output capability shows it's limits, which, if you made an informed choice on the drivers involved will be beyond levels you listen at.
Now, linearity demands padding down more sensitive drivers to match the least sensitive one. The woofer you're using is less sensitive than the EV mid by several db. When you cobbled this all together, did you account for that? Did you model the crossover between the EV driver and the woofer, or measure after applying requisite filters to confirm it reflects the model? Based on the following, it appears you did not.
It seems to me that (at the very least) you have not one but two crossovers to dial in: EV to woofer, and woofer to sub. Aside from that, your choice of drivers is bizarre, and how they're employed seems like a concept that emerged from the OB/dipole branch deep in the audiophool rabbit warren. Doesn't that 15" mid beam like a laser? At 3 khz, that ribbon tweeter won't. On top of matching amplitude, any considerations to directivity matching?
Even if you dial in the crossovers, I fearyou're still trying to make a silk purse out of a sow's ear.