Sub/Main Integration

[brucek]

You know you're done when you measure +6dB at the frequency where your signals intersect. +6dB from -6dB results in a perfectly flat crossover transition.

Nordo wrote:
That doesn't make sense to me

Did you see this pictorial in my post above? Is the resultant signal not +6dB at the frequency where the signals intersect?



brucek

[Nordo]

brucek wrote:

Did you see this pictorial in my post above? Is the resultant signal not +6dB at the frequency where the signals intersect?
brucek

I understand Xovers and I understand your example graph. I just didn't understand what Mike meant by "...measure +6dB at the frequency where your signals intersect...". I thought he was saying that the combined result of the sub and main had to show a +6dB peak at the Xover. Whereas your graph correctly shows that to get a flat graph at the Xover, both signals have to be -6dB at the Xover.

[atledreier]

Yeah. What is meant here is that the combined signal is +6dB from the measured signal of either sub or mains at the crossover frequency.

[clubfoot]

Mike & brucek you're on a roll here, keep it coming, pictures are worth a thousand words....

[DrWho]

atledreier wrote:

Yeah. What is meant here is that the combined signal is +6dB from the measured signal of either sub or mains at the crossover frequency.

Exactly.

So I dug around for a bit and found a graph that kinda illustrates what you're trying to do. This is a measurement of my Klipsch Chorus II's being powered by an active amp I'm building for school (so basically I'm making a DSP front-end to a multichannel amplifier for active bi-/tri- amping of speakers).



Green is the filtered response of the woofer, blue is the filtered response of the squawker, and then green is the filtered response of the tweeter.

Red is the final combined system response.

The minor differences between the individual and summed responses are the result of differences imparted by the smoothing filters. DO NOT use smoothing filters when dialing in the system...I only added them to make it easier to see the concept.

My DSP isn't finished yet so I wasn't able to add any PEQ. The woofer and the squawker both have rising responses around 700Hz, so it would make sense to put a PEQ on each one to tame things a bit before the crossover (not a single PEQ on the summed system response). My tweeters are also tired which is the reason for the rise in the tweeter response - I need to replace the tweeter diaphragms (not add PEQ).

Here's another measurement of the system before the tweeter diaphragms got tired. This is the same filter settings as above, but without smoothing and without time-alignment. Notice the steep notches at 700Hz and 5kHz.



I should probably post an unsmoothed example of the first chart to illustrate how the combined red response at 700Hz actually isn't the broad peak that it looks like.

Anyways, the point here is that you get +6dB from perfect acoustic summation.

[Wayne A. Pflughaupt]

Great illustration there, Mike!

Regards,
Wayne

[Nordo]

atledreier wrote:

Yeah. What is meant here is that the combined signal is +6dB from the measured signal of either sub or mains at the crossover frequency.

That makes sense. That's the basis of all Xover designs, whether your integrating a sub to your mains, or whether your designing a 2-way or 3-way speaker. I've done a fair bit of Xover design for my diy speakers (6 out of 7 in my HT).

But I'm still learning how to use REW (I previously used EFT), and a bit of ambiguous advice throws me out.
It takes very little to throw me out.