Reversed tweeter deep notch?

[Raka]

How big have you managed to get the notch with the tweeter reversed?

[kelticwizard]

Raka:

I am not sure of your question.

In the conventional 12 dB/octave passive crossover, if the tweeter is not reversed in polarity, there generally is a notch. If the tweeter is reversed in polarity with respect to the woofer or midrange, there is not supposed to be a notch.

[Raka]

Yes, you are right, I didn't explain myself. I mean the notch you get if the tweeter is not correctly connected according to the xover, but reversed. Theorically, if the phase tracking is perfect between the two drivers in the xover region you should get an infinite notch, but this is not achievable. I see it as the final test of quality in a xover.
In my last filter (active) I'm getting around 20db, but like to know what values are common out there.

[Salas]

what kind of smoothing and measuring method you use?

[Zaph]

Originally posted by Raka
Yes, you are right, I didn't explain myself. I mean the notch you get if the tweeter is not correctly connected according to the xover, but reversed. Theorically, if the phase tracking is perfect between the two drivers in the xover region you should get an infinite notch, but this is not achievable. I see it as the final test of quality in a xover.
In my last filter (active) I'm getting around 20db, but like to know what values are common out there.

About 40 dB is as deep as it can get.



Still, I think the importance of a deep reverse null is greatly overemphasized. Some designers tweak and tweak until it gets as deep as possible, but the off axis lobing at Fc can still look bad.

[Salas]

In my experience for 6th octave smoothing and MLS method everything over 15 dB is enough. If you use no smoothing it can go to 30dB and over. But I prefer to optimise off axis first and foremost.

[sreten]

Originally posted by Raka
I see it as the final test of quality in a xover.

Hi,

you can regard it as a test of certain acoustic crossover topologies.

However even with the "best" implementation of a topology the
dip will quickly reduce if you vary the vertical measurement axis.

Widely spaced drivers can have a deep notch, but be inherently poor.

I'd suggest the range of the vertical axis the notch is maintained is a
much better test of xover "quality" than a single axis depth of notch.

/sreten.

[Raka]

Well, you are right about the polar response, the funny thing is that at the moment the polar diagram is taking most of my time with the measurements of my dipoles. And as I target a 24LR the lobing problems are less of an issue that with other alignments. But thanks for reminding me the vertical axis existence.

I use various smoothings, and use Speaker Workshop, sometimes ETF, and Sweepscope which is an application made by some friends of mine that includes distortion evaluation and pulse exportation. Very nice, btw.

[Raka]

FWIW, here you are the polar measurements:

This is the horizontal dispersion at tweeter level

[Raka]

This is the dispersion 15º below the tweeter level