Manger

[phase_accurate]

I was also trying to move the "hump" to the steep filter branch(i.e. lowpass) but the maths doesn't seem to like this.

Regards

Charles

[capslock]

Originally posted by phase_accurate
The original Manger crossover performs a 180 phase change during the transition, what a subtractive crossover won't do.
Therefore the first zero-crossing of the step response occurs after 700us approx.

I am still puzzled by that statement. The Manger itself (as per Manger spec sheet) has almost 180° phase shift around its resonance. I guess the change is that slow because of the high R_ms of the Manger. A first order high pass above resonance will introduce another 90°. I am currently confused about the sign convention, but I believe it will lower the total shift to 90°, rather than adding to it.

Anyway, in the test of the Zerobox 109 in K&T 5/99, I don't see any evidence of a zero crossing between 0 and 1 ms.


And as Nelson pointed out in his article on phase coherent XOs, if one of the drivers is operating near its resonance, the acoustic signal will not be phase coherent. The only way I see to change this is to use a Linkwitz to correct the Manger to an f_s well below 50 Hz. But then, you will probably need to put the high slope side on the Manger and have the woofer live with the hump and the 6dB slope.

Regards,

Eric

[phase_accurate]

Hi Eric

If you have closer look at the Zerobox 109 datasheet at

http://www.e-speakers.com/PDF's/Manger%20109%20Project.pdf

you will clearly recognize that there is a first zero-crossing after 500us approx and the second one after 4.5 ms.
An optimal transient perfect speaker would only have ONE zero crossing.

The points mentioned by NP are indeed the ones that make the application of a subtractive crossover with the MSW cumbersome.
Although I regard transient response and minimal difference between both speaker's amplitude response the most important things for this project, I do still aim at an amplitude response that is flat within +- 2 dB. And that is where I struggle most at the moment.

As you correctly mentioned, there is indeed the possibility of getting around the problem with a 6 dB slope on the woofer. I do however fear that even the best woofers will introduce some midrange artefacts that will mask/destroy the MSW's excellent midrange clarity.

Regards

Charles

[capslock]

OK, I guess I misread the axis captions in K&T (the print is really poor).

I gather this 180° shift is mainly due to the resoance of the MSW, not the filter itself. So how does your subtractive crossover get rid of it without resorting to a Linkwitz transform circuit to move the MSW resonance out of the way?

Regards,

Eric

[phase_accurate]

It does not get rid of it.
I do two things:

1.) I have a network that flattens the MSW's response around fs a little.
2.) I have approx 2 dB more gain on the woofer than on the MSW (taking the driver efficiency into account).

This leaves me with a 2 dB dip around 200 Hz. With this dip I could live, but in practice it is more than that unfortunately.

Regards

Charles

[Oz_Audio]

Charles, (phase_accurate),

What Audio Technology drivers are you using with your Manger?

After living with them for a time, would you change the Audio Technology driver to a different size or spec?

Mark

[phase_accurate]

G'day Mark

My AT drivers are Flex Unit (i.e. custom made) 8 B 77 25 06 SD.

Many people say that the C-Qequenze drivers were a little better in terms of frequency response than the Flex-Units (and they have a nicer looking basket).
This fact combined with the cheaper price and the fact that I don't need as much x-max as I first assumed (one's listening habits change over time !) I would even go for a C-Quence 23 I 52 ....... which would even be cheaper.
For me the exact TSPs are not that important since I use an LTF (and the aforementioned C-Quence is also quite close to mine BTW).

Regards

Charles

[phase_accurate]

Well, since I am a grass-widower currently our living room has been changed into a speaker-testing lab. I am currently experimenting with a crossover topology that is a variant of this one:

http://www.diyaudio.com/forums/showt...789#post292789

Attached you can see the in-room measurement of the step-response. You'll have to note that the room is very reverberant, the "echo" at 4ms approx is most probably floor-bounce since it correlates quite well with the path-difference divided by c.

Regards

Charles

[phase_accurate]

sorry, above link was wrong, it should be:

http://www.diyaudio.com/forums/showt...070#post292070

Regards

Charles

[inertial]

Hi,

congratulations Charles, a true step-response!

regards,

Inertial