A 3-way active crossover design study

I am sorry, I just followed what Kannan mentioned.

Yes, you are correct about diffraction. A vital aspect to address in baffle design.
Yes. Baffle design is very important at higher frequencies.
I am trying to show through these plots that both the cabinet width and height and depth are also important in the overall directivity. :)
 
Yes. Baffle design is very important at higher frequencies.
I am trying to show through these plots that both the cabinet width and height and depth are also important in the overall directivity. :)
I am no expert but at lower frequency range baffle has no or very little effect on directivity. If you study 3 way speaker measurements the lower range is omnidirectional. Cant be specific about at what hzs frequency range though. Probably below 1 K ?
Am only guessing with half knowledge so cant be sure.
regards
 
I am no expert but at lower frequency range baffle has no or very little effect on directivity. If you study 3 way speaker measurements the lower range is omnidirectional. Cant be specific about at what hzs frequency range though. Probably below 1 K ?
Am only guessing with half knowledge so cant be sure.
regards
Yes you are right. Baffle has less of an effect at the lower frequencies. Simulations show that the overall cabinet dimensions, ie width, depth, height etc do have an impact on directivity. :) however minimal it is.. :D
 
What I was trying to show with those plots are the effects of cabinet diffraction. Normally diffraction effects are studied only at the higher frequency range, specifically around the tweeter radiation range.
aah...that is s totally different aspect. Sorry for misinterpreting
 
Hi.. You are right.. :D
This is not at all about cabinet resonances. This is all about how the outer shape of the cabinet affects midrange and bass/directivity of radiation at these frequencies.
Regarding tweeters, I generally find that the baffle design usually takes care of edge diffraction with soft dome tweeters and flush mounting mitigates it further. However simple tricks like adding foam around the baffle reduces sibilance to an extent especially with metal dome tweeters.
Measurements may throw up some peaks depending on the on/off axis performance and the crossover points.
 
Regarding tweeters, I generally find that the baffle design usually takes care of edge diffraction with soft dome tweeters and flush mounting mitigates it further. However simple tricks like adding foam around the baffle reduces sibilance to an extent especially with metal dome tweeters.
Measurements may throw up some peaks depending on the on/off axis performance and the crossover points.
Yes. Foam on the baffle helps a lot. Actually this is one preferred way to reduce diffraction of frequencies radiated by tweeter if looks don't suffer :D;). Used even in very high end Avalon speakers.. :) Foam is cleverly made part of the grill.
Other options are assymetrical placement of the tweeter in the baffle (distance to each edge being different) chamfers, large round overs, use of waveguides etc. It may never go away completely, but get distributed to a larger range of frequencies, some get absorbed etc and result in reduction of amplitudes of main peaks and dips to much more tolerable ranges.

It is often said that if we want the speakers to disappear, one of the aspects to take care is in minimizing diffraction/secondary reflections from cabinet. :)
 
Regarding tweeters, I generally find that the baffle design usually takes care of edge diffraction with soft dome tweeters and flush mounting mitigates it further.
Which baffle? Are you referring to the tweeter housing baffle? If yes, it deals with it but then that housing baffle is mounted on the cabinet baffle which add it's artifacts - namely edge diffraction.
However simple tricks like adding foam around the baffle reduces sibilance to an extent especially with metal dome tweeters.
In my opinion, it has nothing to do with the type of tweeter. The waves radiating from any tweeter mounted on a cabinet baffle will be subject to cabinet edge diffraction. The sibilance that you refer to will not be mitigated with dealing with diffraction. It might help with it though.
There are materials much better than foam. But, there has to be a method to the madness and here is where measurements cannot be replaced. Whether what is measured is audible is a different matter. I refer you to David Ralph's experiments on this. In my opinion, superbly done. Here: https://www.speakerdesign.net/

Towards the bottom are the links to diffraction. Extremely interesting & fascinating stuff!
 
Last edited:
The link that @keith_correa gave above is a comprehensive explanation of diffraction from a more time domain perspective. It is awesome and is a must read in my opinion :D

Here is a bit more simplified explanation which also gives perspectives about acoustic impedance variation encountered by a wave which is guided along a cabinet and eventually encounters freespace.

From a signal processing point of view, every reflection of a broadband signal that is going to interfere with the original signal causes the resulting signal to look like the original signal has passed though a comb filter (literally filter response looks like a comb in frequency domain. :D). The comb filter characteristics are decided by the amount of time delay that the reflected signal takes to interfere with the original signal (which changes its phase characteristics and decides whether amplification or attenuation happens at different frequencies).
More info about simplified comb filters here: https://www.tonmeister.ca/wordpress/2021/07/15/intuitive-z-plane-part-1/
(The site also has lots and lots more other info about audio and signal processing. I love it.. :D)

Audibility of diffraction is a controversial topic. At least, for me I can sleep better if I don't see undesired comb filters acting on signals.. :D :p
 
Wondering about a re-usable bass module

I am wondering about making this project a small two way/CD+horn with a re usable bass module. The motivation for this comes from ABEC simulations that FM fluid on diyaudio did for me.
Pic-1 shows the dual woofer configuration.
Pic-2 shows the normalized horizontal polar plot the same
Pic-3 shows the wavefront expansion at 100 Hz frequency. It shows that in the farfield, Both woofers will act similar to a single single driver but with smoother overall directivity and higher SPLs with the right configuration. Each driver will have its own amp and will be DSP controlled.

The motivation behind this is to have a top 2 way module like in my previous posts about the foam prototype. Or a cardiod 2 way like the one going on in ASR as shown in pic 4. I am thinking that the bass module would also suit a horn for the higher frequencies since the bass drivers are good. Idea something like what is shown in Pic 5. The current state of the art in DIY able horns are available here: https://at-horns.eu/ST260.html
thanks to mabat on diyaudio
We can also have custom designed horns to suit our objectives using ATH software that he has developed.

I also have some ideas having the bass module with more directivity control by constructing the enclosure such that it will have cardioidish radiation pattern at the bottom end. Inspiration is Kimmostos projects here: https://web.archive.org/web/2016040...t/CardSub/CARDSUB.html#Cardioid_constructions

sample pic from here are attached as Pic 6 & 7. These are resistance enclosure based cardiodosh designs. ASR forum also has a group project which targets controlled directivity woofer module. I will watch that space also for more ideas.. :D
 

Attachments

  • DW Obs Front 300Hz (1).png
    DW Obs Front 300Hz (1).png
    92.4 KB · Views: 22
  • Img_0858.jpg
    Img_0858.jpg
    54 KB · Views: 18
  • Img_0865.jpg
    Img_0865.jpg
    90.3 KB · Views: 19
  • Screenshot_20220116_211316_com.android.chrome.jpg
    Screenshot_20220116_211316_com.android.chrome.jpg
    574.6 KB · Views: 21
  • Screenshot_20220118_111930.jpg
    Screenshot_20220118_111930.jpg
    526.1 KB · Views: 21
  • DW Side 100Hz.png
    DW Side 100Hz.png
    86.4 KB · Views: 21
  • DW H Polar.png
    DW H Polar.png
    18.5 KB · Views: 19
To keep this thread alive and to push forward to the (hopefully) eventual completion of this project, here are some recent experiments and updates.. :D

Seeing all the xps foam experiments going on on diyaudio forum and impatient with the delay in box building part for the project, I made a 51L sort of braced, sealed box out of some left over foam I had, filled it generously with some polyfill and placed a Satori WO24P woofer in it. The contraption looks like this:
1649340143660.png
There might be tons of resonance issues etc. Still I took a nearfield measurement and got the below response (1/6 octave smoothed).
1649340171676.png
With no expectations, I made some basic filters on EQ APO as below and applied it to the woofer
1649340192807.png
The main aim of the filters were to notch the break up peaks of this driver around 3k and to apply a slight baffle step compensation along with a little bit of excursion protection.
Then I played some songs.. Straightaway, I liked it. It is an awesome driver.. It has good bass. I never expected good midrange from it. But it does well in that part too.. All this in a foam box (at relatvely low SPL levels but sufficient enough at home).. Now I am impatient to see 4 of these in action along with proper mids and tweeter in a proper cabinet.. :D
And I found out I like sealed boxes.
Even more fun is hearing the DSP live in action. I can make changes in EQApo on the fly and listen to the changes.. :D

Here is the impedance plot of the driver in the box (The impedance magnitude peak and hence the box tuning looks to be at 40Hz along with the zero crossing of phase):
1649340377250.png
I can see some blips in the plot above, probably resonances at the lower frequencies along with that driver's inherent blip around 1kHz. I was lazy to not do anything more to the foam box in trying to treat the resonance problem.

Some one might say/feel that this is kind of gross violation of the rights/capabilities of this driver.. Well I agree :D but this is such an awesome driver to keep it in storage :p

Hopefully I will get some real boxes built for these drivers (2 Satori WO24P-8s per box each driver having sealed 35L of cabinet volume to its own, individual DSPed, and powered :cool:)

Hopefully, the final boxes for these drivers will look like this.. :)
1649341246361.png

Thanks
Vineeth
 
To keep this thread alive and to push forward to the (hopefully) eventual completion of this project, here are some recent experiments and updates.. :D

Seeing all the xps foam experiments going on on diyaudio forum and impatient with the delay in box building part for the project, I made a 51L sort of braced, sealed box out of some left over foam I had, filled it generously with some polyfill and placed a Satori WO24P woofer in it. The contraption looks like this:
View attachment 68571
There might be tons of resonance issues etc. Still I took a nearfield measurement and got the below response (1/6 octave smoothed).

Yes-yes-yes foam/etc. is the spirit and you keep moving! It is also two other gifts. You now have ear-time and data on it. Well-done!
 
Some more experiments with WO24P and SB Audience Rosso65CDN-T on Faital Pro LTH142 horn

Still, waiting for the final boxes. In the meanwhile, to study taking more polar measurements and familiarize more with VituixCAD FIR filter generation capability, I tried to do polar measurements for a speaker set up like below :D
1649612049208.png
Measurement in Progress:
1649612024582.png
Driver WO24P-8 horizontal polar:
1649611996245.png
Rosso 65CDN-T horizontal polar:
1649611972732.png
Rosso 65CDN-T vertical polar:
1649611955968.png

Sample VituixCAD FIR crossover:
1649611939275.png

EQ Apo filter implementation:
1649611925091.png

As the next exercise, I am planning to study implementing an IIR crossover for this setup :D :D
The idea behind all this was to learn how a readily available horn will sound like before proceeding with customised/more complicated horns in addition to killing some free time :)

Thanks
Vineeth
 
Last edited:
Updated FIR crossover:
I updated the crossover to the following :D
1649781487467.png
Resposes: of the overall speaker:
1649781536609.png

Predicted step response:
1649781591618.png

Listening impression:
Awesome.. :D :D
I have permanently become a fan of horn-based systems. If I can achieve this kind of sound with a single speaker with a foam box, I am never going back.. :D :D

Too lazy to pick up the mic and measure the final responses.. Just enjoying music.. :D
 
I don't know how this sort of thing will work but here are two videos of the above system playing at relatively low volume levels (Captured on my not so good mobile phone. Hence it sounds nowhere near it sounds in person but maybe able to get some sort of idea):

Please note the the woofer and tweeter may not be level matched. In short the highs could be hotter than they should be.
I have to measure and adjust this and refine the setup further. Filters are the FIR ones above.. :)
Oh.. And I think I am playing 128kbps or so MP3 file.. :D
 
Last edited:
With whatever little time I got yesterday, I tried the following IIR-based crossover on the speaker. I also level matched the system my measuring and adjusting the amplifier gains. There was around 4dB difference between levels of the two drivers around the crossover point region.
Technically it is still FIR crossover in implementation but the target transfer functions are generated using IIR filter targets... :D
I have also included a 150uF capacitor in the tweeter branch to protect it from DC but it does not affect the below transfer functions much and is not shown in diagram.

1649911364595.png
1649911379760.png

Shown as overlays in the Power & DI plot are the results from previous FIR-filter based implementation posted above.
It looks similar I think. But to my ears, it felt like a lot of difference.

Hearing impressions-wise, suddenly it seem much more enjoyable..
Earlier I felt it was a bit hot in the highs and vocals-region and details were being thrown to face but but now the balance seems like it has fallen into place, sort of and system seems more effortless to the ears.. Again all this might be due to 'N' number of factors and I am certainly not going to say pointless statements like 'IIR filter slopes-based crossovers are better than FIR slopes with perfectly flat ON-axis' etc. :)

In fact I have no motivation to change from this crossover at least immediately so that I can find its weaknesses and learn from it before trying a next one.. :)

From the system point of view, the compression driver is awesome, I think.. I should say that I am hearing little nuances in songs that I never heard before.. :D I can hardly turn up the volume control on it before I go uncomfortably loud. There is a little bit of audible 'hiss' if I put my ears into the horn mouth even when nothing is playing. Maybe the amplifier noise is on the higher side compared to the best? (I am using this cheap amp for driving the CD: https://www.allo.com/sparky/volt-plus-amp.html )

Will try more things and post impressions.. :)
 
It will be easier to dig-out the noise/hiss and other issues when you get from the 88dB to the 98dB woofer and re-level everything.

If I might suggest anything, it's to not completely ignore how the driver sounds in the foam and w/o lining/damping/filling. Over time, those experiences will add-up to help you recognize patterns. You have so much going-on that acoustics hasn't yet entered the discussion, which is fine. There is recklessness available to you when in foam that will no longer be available, too. Turn the whole thing 90 deg (change CTC), add "wings" or bevels, add "foam bracing" etc as time/inclination permits. Fill the whole thing with polyester pillow stuffing. Whatever you choose, all of it will add-up in your head to patterns/learning over time :) Cardboard/foam/cheap proto cabs are a license for experimentation.
 
Last edited:
The Marantz PM7000N offers big, spacious and insightful sound, class-leading clarity and a solid streaming platform in a award winning package.
Back
Top