First VituixCAD Design - Please check

[Chris Roemer]

3 "notes":
1st - I don't model BSC (I just show it to you in my FR - it's "generic") or diffraction ('cause I'm lazy and don't know YOUR baffle layout).
2nd - Diffraction (which you can SEE in "Edge" - and apparently "model" in VCad) is worst when a tweeter is exactly the same distance from the 3 (closest) adjacent baffle edges. If my baffle is 8" wide (OD), then avoid having it ALSO 8" down (or 4" down, which is a harmonic). I'd put it: 8" / sqrt(2), OR 8 / 1.414 (which is around 5-1/2"). OR you could try 8" / 3 (which is NOT a power of 2 - which is an octave) - about 2-3/4". How close you can get it to the top depends on your tweeter's faceplate dia. Funny thing you see is if someone would take the tweeter, like 1.5" off-center (so ... 2.5" from the L edge, and 5.5" from the right), then they'll ALSO position it 2.5" down from the top, having it the SAME distance down and over from the corner. NOT GOOD.
3rd - I prefer the look of a centered tweeter, but ... the "wiggles" you see are there "on-axis" - many will leave their speakers pointing straight out from the (front) wall instead of toeing them IN (aiming them at their listenting position). This means that they're listening to them OFF-AXIS, and off-axis, a lot of those "wiggles" go away (at least somewhat). Many times you can get the diffraction "wiggles" substantially reduced by just moving a tweeter (or tweeter and mid - if a 3-way) over off-center by just about 1/2".

If you DID trace the woofer plot - I'd like to see what you got (errors can be made setting tracing limits - or adjusting the scale incorrectly).
You could show us the unfiltered response, or just post your .frd file. My "notch" filter (across the big coil) should pull that huge peak @ 2.4kHz down over -30dB.

[dcibel]

zinger084 Recommendation for accurate design data in VituixCAD is always to use real measured data. However, to understand the tools and response processing, I've created a document that describes how to process manufacturer data with VituixCAD tools that roughly follows a similar process that would be completed with real measurements. It should be helpful to you to get started with VituixCAD with something that at least resembles reality.

[fpitas]

What rpb said. I'll add, peaks are usually a lot more annoying than dips. Everybody has their own standard, but I try to keep the peaks to 1dB or less viewed using 1/12th octave smoothing. Also be aware you will hear big peaks from a driver even in its stopband, even though they don't do much to the summed response.

[rpb]

very interesting. the Fc looks close to 1400 or so... which is getting down there i feel like. any issues with that?

Also, how would i know to add so many configurations of R\L\Cs to get this so flat? just trying to learn as much as i can prior to pulling the trigger on the crossover components...

something i keep coming across but can't find the answer... when should i be worried about the peaks in impedance like we're seeing here at 20 and 1k? how high is too high? when should i flatten (and how, frankly...)

EDIT: my baffle is 40x14.125 with tweeter at 36 and woofer at 26-25/32 (i made them a few weeks back waiting on drivers and deciding on crossovers... which began this rabbit hole ;)

The xo complexity is optional, A lot depends on your intended use for the speaker, and your "audiophile" type pickiness. Peaks and dips may or may not bother you at all if you are cranking some rock, and roll at high levels, or playing music videos of limited quality. On the other hand, if you intend to listen to all the little details in complex music, then a better xo is in order. There is such a thing as overkill for a x-over. DeZZar, and myself would likely lean towards a complex xo, mostly because we can. On the other hand, xo parts aren't free. For the extra money you might spend on the xo, you might have purchased a more expensive driver that would not need such an elaborate xo. There's always trade offs!

[DeZZar]

very interesting. the Fc looks close to 1400 or so... which is getting down there i feel like. any issues with that?

Should be fine with this tweeter - its about as low as you would want to go but the third order network keeps it pretty safe.

Also, how would i know to add so many configurations of R\L\Cs to get this so flat? just trying to learn as much as i can prior to pulling the trigger on the crossover components...

I don't know if I have a great answer for you on that one. What you are looking at here is a third order damped elliptical network on the woofer - not that common - just something the data asks for - its experience/practice. This type of network really slams the door shut on the woofer - as its a 10" stamped steel frame high moving mass woofer I wouldn't want it to make a peep over 2K - this network suppresses all woofer breakup/beaming etc to over 80db down.

something i keep coming across but can't find the answer... when should i be worried about the peaks in impedance like we're seeing here at 20 and 1k? how high is too high? when should i flatten (and how, frankly...)

In this case you can add an RC in parallel to the tweeter at the beginning of the network to tame the crossover impedance spike. I don't think it presents an issue though.

EDIT: my baffle is 40x14.125 with tweeter at 36 and woofer at 26-25/32 (i made them a few weeks back waiting on drivers and deciding on crossovers... which began this rabbit hole ;)

Revised design based on your baffle and layout. The choppy tweeter response is because of your chosen location - nothing you can really do to fix that in the crossover. I haven't allowed for any round over on your baffle - if you have at least a 1" round over that would help a lot (soften those peaks/dips by nearly 3db)

[zinger084]

zinger084 - just a note if you are looking carefully at Dezzar's components and want to use it as a starting point....some capacitors are uF (which is normal) and some like C5 are nF. Similarly, some inductors are mH and others like L1 are uH. Some resistors are Ohms and others mOhms.

thanks for the heads up. as an engineer by trade, i try to keep an eye on units... but sadly i'm not a sparky, but a gearhead... or i think this stuff would come to me much quicker!

[a4eaudio]

zinger084 - just a note if you are looking carefully at Dezzar's components and want to use it as a starting point....some capacitors are uF (which is normal) and some like C5 are nF. Similarly, some inductors are mH and others like L1 are uH. Some resistors are Ohms and others mOhms.

[zinger084]

....and to illustrate how important tweeter position is - here's an amended result based on a slight reposition of the tweeter for a better diffraction pattern....

very interesting. the Fc looks close to 1400 or so... which is getting down there i feel like. any issues with that?

Also, how would i know to add so many configurations of R\L\Cs to get this so flat? just trying to learn as much as i can prior to pulling the trigger on the crossover components...

something i keep coming across but can't find the answer... when should i be worried about the peaks in impedance like we're seeing here at 20 and 1k? how high is too high? when should i flatten (and how, frankly...)

EDIT: my baffle is 40x14.125 with tweeter at 36 and woofer at 26-25/32 (i made them a few weeks back waiting on drivers and deciding on crossovers... which began this rabbit hole ;)

[DeZZar]

....and to illustrate how important tweeter position is - here's an amended result based on a slight reposition of the tweeter for a better diffraction pattern....

[DeZZar]

Yes, you definitely need to model the tweeter here in the same way. The output files are the new 'anticipated' frequency responses for the drivers in your application. These are the files you should now use as the FRD files for the drivers in your crossover. There is also an option on the first tab to generate 'minimum phase' from the files for use in the crossover design.

[DeZZar]

Just based on a guestimate baffle size and speaker positioning. Potentially some off axis issues.

[rpb]

The Peerless DA32TX00-08 Tweweter didn't have frd or zma files that I could readily find... maybe I'm missing someones secret???

Your post read like you traced the Dayton files. The Peerless tweeter is virtually perfectly flat. It should be easy to make a filter for it. Make the roll-off gradual,and smooth. Try to do the same with the Dayton. I'd shoot for around 1,7k, depending on your intended use of the speaker.

[zinger084]

This will give you a full set of diffraction modelled response files for your woofer. Start with the FRD file for on-axis (0 degrees), setup your baffle and driver position as follows...

OK, I didn't know I needed to hit "new" after entering all the info (I thought that would create a new model and delete all my info...smdh. So, I modeled the woofer, but does the tweeter need added (I assume so)??? What do I do with the exported .txt file?

I really with there was some training videos on this software... anyone know where I may find that?

[zinger084]

Is there some reason you could not use the frd, and zma files without tracing?

The Peerless DA32TX00-08 Tweweter didn't have frd or zma files that I could readily find... maybe I'm missing someones secret???

[DeZZar]

Chris has included BFC in his recommendation. But as you can see this is a classic example of why real measurements and proper modelling are required in the theoretical domain before you even decide to purchase drivers and build a box, in order to achieve the right results.

No online calculator is able to anticipate the misbehaved woofer here and its lumpy 1K characteristics - or the choppy bottom end of the tweeter particular when diffraction is modelled in.