3-way vs. 4-way

[Dr.db]

Thanks for the link to wiki and the TAD-shematic!

We will use an active crossover for this project to get started. Probably we´ll stay with this, as I have very limited skills in passive crossovers.
Using the TAD crossover with our JBL components probably isn´t working at all, so we`ll better stick with an active xover.

Ian, what I had discovered before asking is that phaseshift caused by a filter changes over the frequency-range. This made me wonder, if the attempt of an electronic offset (crossover) has the same effect as a physical offset regarding the time-alignment. Reading your link, the pictures do say so.
I´m aware of my lacking knowledge, so I´ll try to avoid going into detail any further.

Just let me ask this simplified: if a compression driver is located significantly behind the woofer because of a deep horn, is a higher crossover-slope on the woofers lowpass a smart idea in most cases?
Did JBL use a 24db lowpass (woofer) and 12db highpass (midhorn) on the DD66000 for the same reason of time-alignment? Or is this because the acoustical roll-off of the midhorn(~12db) sums with the electrical roll-off (12db) a 24db slope and therefor the lowpass needs 24db slope as well to match up?

[dn92]

JBL 2" drivers in TH-4001 response is a bit different but not so much. I've EQ settings for this that have been tried on the loudspeakers of a friend (2446J on TH-4001 horn clone).

[Ian Mackenzie]

Thanks for the link to wiki and the TAD-shematic!

We will use an active crossover for this project to get started. Probably we´ll stay with this, as I have very limited skills in passive crossovers.
Using the TAD crossover with our JBL components probably isn´t working at all, so we`ll better stick with an active xover.

Ian, what I had discovered before asking is that phaseshift caused by a filter changes over the frequency-range. This made me wonder, if the attempt of an electronic offset (crossover) has the same effect as a physical offset regarding the time-alignment. Reading your link, the pictures do say so.
I´m aware of my lacking knowledge, so I´ll try to avoid going into detail any further.

Just let me ask this simplified: if a compression driver is located significantly behind the woofer because of a deep horn, is a higher crossover-slope on the woofers lowpass a smart idea in most cases?
Did JBL use a 24db lowpass (woofer) and 12db highpass (midhorn) on the DD66000 for the same reason of time-alignment? Or is this because the acoustical roll-off of the midhorn(~12db) sums with the electrical roll-off (12db) a 24db slope and therefor the lowpass needs 24db slope as well to match up?

It’s a case by case question

The Jbl drivers and horns are different to yours. Their combined electrical-driver slopes where symmetrical AFAIN. Read the white paper

In theory a steep slope on will have less lobe related problems in the crossover region

But every design has a hierarchy of design goals and they are often competing goals.

Ideally you want to set of complementary compromises

But listening tests are also required in each case.

The Tad woofers breakup quickly (l have been told) so a steep slope made sense.

They chose a lower slope for the horn. There may have been several reasons for that?

You can’t guess that.

What l am saying is there is a hierarchy of considerations.

It’s too early to be theorising. Don’t get hooked on over thinking a solution or over reading.

There is a tendency to become self obsessed about a particular topic if you over read.

It’s not real and you are not present.

It’s like a person who reads a book on how to swim.

They then dive in and sink. It goes on.

A successful loudspeaker is NOT 95% thinking and design, 5% build.

Discussions here have very limited impact on your success unless you are closely following a Jbl design.

First are your drivers. Understand them, listen to them.

Get your hands dirty and setup modular test boxes

Do your measurements and listen on your drivers

You then use a dsp crossover for exploration of flexible crossover options.

Pink noise or off station FM noise is useful to assessing the blend.

Listen and measure.

You can then make some informed decisions

Bread crumb: there tends to be a very narrow sweet spot in terms a fine adjustment of the crossover in both listening and measurement

That is why you should invest in a measurement device

[Dr.db]

@dn92: Would you share your eq-settings?

@Ian: Thanks a lot so far, I´ll continue with extensive listening sessions now

[Ian Mackenzie]

Below is an example of two way system vertical crossover lobes

Passive network 36 dB slope , crossover point 800 hz

The key is to obtain exact phase tracking at the crossover point.

Attached is the frequency response from the design point and vertical polar responses for both in phase and reverse phase of the drivers.

The data is actual raw driver measurements of an AE15M and a Tad 4003 on AH700 conical horn

The computer can very accurately simulate the proposed crossover design schematic within the crossover region.

By reversing the phase of the drivers cancellation can be seen of nearly 40 dB at the crossover point.

The primary impact in the vertical lobe is at 800 hz due to steep crossover slope.

In phase the lobe is exactly where l designed it in the crossover region.

This is only possible with quality measurement data, powerful simulation software and experience.

Even with the computer this took a while with several different crossover attempts.

It is optimal outcome is very sensitive to small changes in the crossover design and location of the drivers.

I am not setting the bar here but to illustrate the theory and the practical outcome of crossover network operation.

[Dr.db]

Thanks Ian, those graphs are very descriptive.

Maybe reversing the polarity on one of the drivers while listening and playing a test tone of the crossover frequencie gives a first guess if phase is correct?

[Ian Mackenzie]

In the crudest sense a sine wave sweep at 1 m with your trusted spl measurement of the dip.

The actual crossover slopes need to be correct prior to that approach

I have previously indicated an active crossover solves the problem of complex and often difficult passive crossover design and implementation.

You still need a trustworthy measurement device.

The alternative is closely follow an existing Jbl design.

The thing is Andrew Jones (renowned loudspeaker designer) can design and get to market a really good sounding cheap loudspeaker.

So the tipping point for your own design diy loudspeaker is shelling out a modest outlay on an active crossover and a trustworthy measurement kit.
That really is the gateway for rolling your own diy loudspeaker.

Conversations are best about measurement techniques. Not the end solution.

That will prove a valuable long term investment.

[Dr.db]

Sounds reasonable, thanks a lot for your help so far!

[Dr.db]

Regarding the crossover between 15"-woofer and 2"-driver on yuichi horn there are at least two contrasting demands:

On the one hand you want the crossover to be as low as possible to increase group-delay of the lowpassfilter on the woofer to help compensate for time-alignment to the deep horn-profile. So a ~650hz crossover would be desired.
On the other hand you want the crossover to be a lot higher so that the 15"-woofer reaches almost the same dispersion as the yuichi horn and they match up nicely. So a ~1000hz crossover would be desired.

I know it always comes down to some kind of compromise, so which of these two aspects is more important and where about would you set the crossover?

[Ian Mackenzie]

The best course of action is to commence evaluating your drivers yourself and the facts will help you make a decision.

[ivica]

Regarding the crossover between 15"-woofer and 2"-driver on yuichi horn there are at least two contrasting demands:

On the one hand you want the crossover to be as low as possible to increase group-delay of the lowpassfilter on the woofer to help compensate for time-alignment to the deep horn-profile. So a ~650hz crossover would be desired.
On the other hand you want the crossover to be a lot higher so that the 15"-woofer reaches almost the same dispersion as the yuichi horn and they match up nicely. So a ~1000hz crossover would be desired.

I know it always comes down to some kind of compromise, so which of these two aspects is more important and where about would you set the crossover?

Hi Dr.db,

I will not pay any attention about 650Hz dispersion characteristics of A290 either 15"-woofer.....
talking about group delay of the woofer with the passive network near 650Hz all the elements have to be taken 'into account'.
1ms delay would correspond to 34cm. May be 4th order passive filter would help

regards
ivica

[Ian Mackenzie]

In reality depending on the dimensions of the horn mouth and profile the dispersion may be wider than you think

The woofer may also sound terrible above 650 hertz

So you need to do some home work and determine what the horn and woofer is doing

[Dr.db]

I got you...
So far I have to say, the TAD 1603 isn´t super snappy in the upper bass, but I was impressed how clean it sounds even in the mids. I´ve had woofers that thrilled me more in the upper bass, but the TAD isn´t doing anything wrong at all in the mids. As I said, very clean.
I wouldn´t mind crossing it over higher...
But so far I can´t judge the yuichi, as I havn´t build or listened to one.

[Ian Mackenzie]

In that case to close the loop on your questions l would simply email Yuichi and get his feedback or recommendation on a crossover point.

Auditioning is often the only way to do this.

He has helped numerous people with these horns and would no doubt help your get started.

The A290 might look similar to a Tad horn but small details are important.

Ie the depth and other dimensions are probably different to the Tad and you might not be using a Tad driver. Your woofer is a 1603 not a 1601 .

Therefore any desire to clone a Tad monitor is out the window

http://www.technicalaudiodevices.com/lf-units

[Dr.db]

I will email yuichi