Inviting design ideas for these drivers and cabinet (files and pics inside)

[percy]

I have three woofers and one tweeter for each side. Enclosure is ~34l, ported. You might have seen my other thread about 'deciphering crossover'. Its for this speaker. I would like to see what is involved in redesigning the crossover from scratch. If you had these drivers and this volume to work with what would you do ? What do you recommend ?

I am attaching the ZMA and FRD files for the woofer and tweeter. and also screen shots of a ungated farfield response including distortion components to give an idea of the real FR and distortion profile of these drivers.

Let me know if you need any more information in particular.

[percy]

FRD and ZMA - (you will have to remove the .txt extension on the files, it wouldn't attach without it).

[sreten]

Hi,

With 3 very low impedance drivers I see very little point in trying a "redesign".
I cannot see any sensibly alternative topology arrangement you could use.

rgds, sreten.

[percy]

nothing series either ?

[Jay]

Originally Posted by percy

nothing series either ?

Of course you can try series crossover. Second order for example.

First filter: L=0.15mH, C=10uF (in series with 8R2)
Second filter: L=2mH, C=22uF

Three woofers in series, tweeter polarity flipped.

[percy]

What about BSC ? the speakers will be up against the wall as show in the pic but the measurements were taken away from the wall. Also, I think the dip between 150-250hz is something to do with the internal(cabinet) or external(room) resonance. Dimensions of room are 15x12x9 and one of the 12' sides opens into another room.
And the nearfield response of the driver is very flat all throughout, it does not show the dip. The dip starts appearing and increasing as I move the mic 6" and more away from the driver.

Now are those textbook values or did you run the files through a XO designer ? Would PCD7 allow me to model those 3 drivers in series as "one woofer" ?

[Jay]

Originally Posted by percy

What about BSC ? the speakers will be up against the wall...

Now are those textbook values or did you run the files through a XO designer ? Would PCD7 allow me to model those 3 drivers in series as "one woofer" ?

BSC? Why you need it? The above values are "copy" of the original crossover (L=2mH, C=10uF). So basically it doesn't change the low end response. Effect of baffle step loss is depending on the speaker placement against the wall, so you don't really need electronic compensation.

I don't know PCD7, but been a programmer once I guess you can model the 3 drivers as one, probably with work around.

Here is I'm attaching a "better" crossover. I still use common values. The 1.8mH can be achieved from your 2mH (but you need LCR meter).

Of course I can create a better looking chart, but it is not necessary because in my experience, tuning by ears will not yield final filter corresponding to the best looking chart. What is IMPORTANT is the phase tracking shown in the chart, meaning that around the filter there is a "sweet spot".

The components values in the tweeter filter is friendly enough.

The woofer response doesn't have problematic peak (as you have measured). And being in series the cones movement is minimal.

The minor issue may be the driver placement and the internal volume if the 3 drivers do not have equal volume.

[Jay]

Originally Posted by Jay

Of course I can create a better looking chart, but it is not necessary because in my experience, tuning by ears will not yield final filter corresponding to the best looking chart. What is IMPORTANT is the phase tracking shown in the chart, meaning that around the filter there is a "sweet spot".

Correction: your data is a real time measurement of the EXACT real drivers, so what you see may be what you will exactly get.

[tuxedocivic]

Your frd and zma files don't work for me. I converted them, but the text format looks wrong. You have 3 different amplitudes for each frequency so PCD just chooses 0db or something. Idunno.

[percy]

Thanks Jay I will check it out and will post my feedback. btw, is that Speaker Workshop ? XO design is the only thing I haven't tried in SW. Perhaps I should try again.

Do you by any chance have the total system impedance ? you dont have to redo the whole thing. I was just curious to see it.

I was hoping to push the XO crossover further up (~3khz?) than what it is right now (2.3Khz) since I am actually hearing audible distortion between 1.5Kz-4Khz, and it reveals in the chart I posted in the original post as well. I want to damp that down as far as possible.

Thanks.

Originally Posted by tuxedocivic

You have 3 different amplitudes for each frequency so PCD just chooses 0db or something.

Yes I am sorry there are not really standard frd. Those columns are the distortion figures. You can just open it in excel and remove them as well as other header information in the file.