SB Acoustics Satori Monitor

[gornir]

I am using the 8 version. My design axis is at the height of the tweeter - my stands are 25" tall. Of course that means the tweeter is off-axis about 14 degrees due to baffle tilt. I found this to be fine since the TW29 has a rsing top octave anyway.

It's interesting to see that the 8 Ohm version seems to have a deeper suck-out at the edge resonance than the 4 Ohm version, even though the difference isn't huge.

Originally Posted by Jeff B

My phase response is measured from the spike of the impluse of the MLS with excess phase removed for time of flight. At other distances it will pretty much look the same, as it will follow the frequency response. In other words if you move off axis, or get far enough away that the top end is dropping then this will be reflected in the phase response as well. Phase and frequency response are intimately tied together in this type of crossover, hence the term "minimum phase response".

Sorry, I didn't read your post carefully enough.

/Göran

[gornir]

Originally Posted by Jeff B

I don't know if this will be of interest or not, but here is an article I wrote explaining how to determine the relative acoustic offsets between two drivers using simple frequency response measurements and a crossover simulation program. In this version you simply assign the tweeter's acoustic center to be the reference point, even though its exact location may be unknown. Then using this method you can very precisely determine the offset distance between the reference point and the woofer's acoustic center.

As you know, the location of the acoustic center of a driver is not always obvious. For example, for many years it was regard to be the position of the voice coil, so if voice coils were aligned then we had time alignment. Later research found this to not be correct. In most midwoofers the acoustic center is actually very close to the attachment point of the coil former and cone, sometimes, depending on the driver design and how the dust cap is implemented, it can even be forward of this. With dome tweeters it is typically close to the base of the dome.

This method can very easily be used to determing how to offset drivers to exact time alignment and it easier to dial in and read than by viewing the impulse and step response.

Here is a link to a pdf file with the article in it:
https://www.box.com/s/ouxjjsx0m8bs00cil5iq

Jeff B.

Yes, I've bee using this method for years. It's fast, simple and accurate.

I'm using LspCAD 6 Pro for simulation and design and it's very easy to import it in LspCAD and determine the off-set and it's easy to recalculate for different distances etc.

Here's another article using the same method Finding Relative Acoustic Offsets Empirically

Regards

/Göran

[Joachim Gerhard]

The acoustic center can be found by eroding the frequency response.
Do a Hilbert transform of the transfer function. Calculate the excess phase and then the
excess group delay. The amount of delay that is necessary to make the excess group delay flat gives the acoustic center.
Anyway in practical terms that does not mean much. A real transducer seldom has a dead flat response and there will also be some all pass terms in the area where the driver breaks up in partial resonances. Richard Heyser made a point that the acoustic center can also move with frequency.
I have now measured the drivers in the test cabinet and found that the acoustic centers match well when i measure on the tweet axis at ca. 1m.
The woofer sits on a baffle that is 38mm thicker then the tweeter baffle.
The crossover i made is a L/R second order with the tweeter reversed.
The acoustic phase responses match very well.

[Joachim Gerhard]

I will publish the data soon.

[Jeff B]

Originally Posted by Joachim Gerhard

The acoustic center can be found by eroding the frequency response.
Do a Hilbert transform of the transfer function. Calculate the excess phase and then the
excess group delay. The amount of delay that is necessary to make the excess group delay flat gives the acoustic center.
Anyway in practical terms that does not mean much. A real transducer seldom has a dead flat response and there will also be some all pass terms in the area where the driver breaks up in partial resonances. Richard Heyser made a point that the acoustic center can also move with frequency.
I have now measured the drivers in the test cabinet and found that the acoustic centers match well when i measure on the tweet axis at ca. 1m.
The woofer sits on a baffle that is 38mm thicker then the tweeter baffle.
The crossover i made is a L/R second order with the tweeter reversed.
The acoustic phase responses match very well.

My sloped baffle results in setting the tweeter back 37.5mm, so we are very close on this.

[Joachim Gerhard]

Cool, i had some 19mm thick chipboard so 2 x 19 = 38.
That is a tiny bit more then the calculated 32 mm but on the tweeter axis in the near field,
say not more then 2m away the woofer is a bit further away because of geometry.
Anyway, i will try to simulate your solution too. I will not copy it though without your permission.

[Joachim Gerhard]

Here is a first attempt at a crossover. It is an L/R2, the lowest order crossover that has the drivers in phase so the lobing is symmetrical in the vertical direction. It also has less phase distortion then an L/R 4 although in the Step Response you can see that the tweeter starts negative.
The ETC shows that this design storers very little energy.
More later.

[Joachim Gerhard]

Here you can see the effect of the edge between tweeter and woofer.
When it is not chamfered there is a suck out that amounts to 5dB in a certain range.

[Richidoo]

Why is it preferable to design with tweeter at ear height rather than midway between woofer and tweeter? Thanks

Thanks for posting links to acoustic center articles.

The crossover looks very good!

[Jeff B]

Originally Posted by Richidoo

Why is it preferable to design with tweeter at ear height rather than midway between woofer and tweeter? Thanks

Thanks for posting links to acoustic center articles.

The crossover looks very good!

I am not sure who the question was directed to, so I will chime in, and anyone else is welcome to as well.

However, the tweeter axis is typically used as the design and listening axis simply because it is the more directional of the two drivers, especially at the upper end of its frequency range. Midwoofers are usually crossed over before they become directional, but we don't have that luxury with tweeters. If a tweeter is large, like these 29mm SB tweeters then the top octave can roll-off quickly as you move off axis.

Jeff