Final measurement help on a simple 2way

[tuxedocivic]

Hi all,

See this measurement:



That's on axis at 1m. The rising top end was deliberate as I listen off axis and I'll be padding the tweeter down a couple db and remeasure. Here is the in-room, in position measurement smoothed 1/6th. Blue line. Red is in position @ 1m on axis:



Here is a nearfield measurement of the woofer:



My question is with respect to the 5-6db dip through 2 - 3khz. It's not that audible, but annoys me that it's measured that way, and as this is my very first from scratch XO design, I wanna learn. I don't really see it on the in room measurements. Is it a phase issue? Are the LP and HP to far apart? Is the gating causing some kind of anomaly?

Any help is appreciated. Thanks.

[AllenB]

You said that you allowed the rise on axis as you planned to listen off axis. I assume that the dip is not there off axis. Your in room response may be showing this as it includes reflected off axis sound.

[tinitus]

Originally Posted by tuxedocivic

Is it a phase issue?

it could be

how does it look if you reverse polarity on tweeter

the expert guys says that when design is right you will have a 'deep null' at xo point, with wrong polarity on tweeter
its my understanding that they use it to check correct function

well, in my book a good speaker is made for on axis listening
to me, off axis sounds, ehh....off

[Inductor]

Originally Posted by tinitus

...
the expert guys says that when design is right you will have a 'deep null' at xo point, with wrong polarity on tweeter
its my understanding that they use it to check correct function

Doesn't look bad to me tinitus.
It shows very much like a null, but, polarity is very right, from output graph above. Could be a null from the angle with the tweeter (not phase issue) of from the mid/woofer only (nearfield measurement of the woofer). Doesn't show like a "practical" problem.

[tuxedocivic]

Hi guys again, and thanks for the thoughts.

So I was resisting reversing polarity because I had it all soldered up already :P But I did anyways. This is what I got:



So it looks like that corrected it. Having said that, I did have two 2ohm resistors paralleled together in series with the tweeter, and snipped one out at the same time as flipping the polarity. The tweeter level seems better now. And I don't think that would have played much into making that look flatter.

My model in PCD says the null should be deep, but that's just a model and I'm new to measureing. I somewhat wonder if I didn't screw up my measurements. I did compare them to zaphs with similar results.

So... was this a polarity thing after all? Is the XO not very good?

[tuxedocivic]

Oh and I almost forgot, I did this measurement in position unfortunately due to time. So there are early reflections, hence the huge amount of gating. The red is my last measurement from my OP for comparison, but I gated it the same as the blue for a fair comparison.

[Henry8]

I think you need to do more tests.
Measure the drivers with and without xovers, tweeter at 1m, woofer n/f
Then you can see if your xovers are acheiving the desired filtering acoustically.
With drivers of equal ampiltude output, and correct filtering, a good reverse null (say of 20db or so ) can be seen.