I decided to try Xsim for the first time to simulate a 2nd order crossover network, but I'm having trouble understanding why the SPL dips so much at the crossover point. The crossover is set to 3.2khz, Q values are 0.707. I feel like I'm missing something basic here.
The individual SPL graphs look good, but together it's a mess.
The individual SPL graphs look good, but together it's a mess.
You probably have an excellent crossover, once you reverse the polarity of one of the drivers. If your slopes are acoustically truly 2nd order (not just electrically 2nd order) and your measurements of the two drivers are phase coherent, then this dip is expected until you flip the polarity of one of the drivers. A 2nd order xo results in the drivers being 180 degrees out of phase, which is typically fixed by polarity flipping.
I actually have a second question for anyone reading this. It looks like I can obtain the flattest frequency response by having my 2-way crossover at 2 different frequencies due to some peaks in the woofer. It looks like 4.5k for the tweeter and 3k for the woofer works best. Is there any reason not to do something like this?
That is OK. Spreading effects like this does happen, but it's also important to not expect each way to help too far out of their preferred band.
Yes, exactly. I play with knee frequencies of the two drivers (somewhat) independently and then see where the two slopes are intersecting, and call that my crossover frequency. So the Fc is a consequence of my design and simulation, not strictly an input to it.