• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Building a 2-way small active speaker with software crossover

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,506
Likes
2,544
Location
Sweden
The stereo system errors again. With respect to the timbral errors and EQ, one can dampen the peaks a bit 1-1.5 dB. Regarding the dips, just make sure there are no dips of the reflective sound. Rather a bit more energy 1-2 kHz.
 
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,649
Likes
10,860
Location
Prague
First measurements -

Near field woofer
2-waybox_onstand.JPG



On-axis frequency response
2-waybox_EKIO_onaxis.png


Note: with the LR 48dB/oct filters it is much smoother and more flat than with LR 24dB/oct filters. Q-notch at 6500Hz is used.
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,506
Likes
2,544
Location
Sweden

DanielT

Major Contributor
Joined
Oct 10, 2020
Messages
4,894
Likes
4,877
Location
Sweden - Слава Україні
I've designed numerous analog active and passive speaker crossover as day job and never, absolutely never, any of the highpass or lowpass filters was one of the textbook filters. Textbook filters are only useful, in very rare cases, for final acoustical target... in practice you need variable Q. You know, the textbook filter XO phase response falls apart the moment you have another slope in the vicinity, be it a natural roll-off (LF system corner frequency) or be it the XO to an adjacent way. In a two-way, these additional slopes are sufficiently far away from the XO point, but from three ways up (notably including any subwoofer XO) you cannot use any textbook XO filter function anymore as it stops summing correctly because phases are are not matching anymore. And that's just for the final target functions, mind you.

For example, if you settle for a target for low-to-mid 3rd-order XO with a constant 60 degree phase offset (because 90 degree standard 3rd order Butterworth would give more lobing issues and less efficient summing etc), how you are going to do that without free access to filter Q? As mentioned, you then must use additional PEQ and shelves to bend your fixed filter's magnitude and, much more importantly, phase to the required target. Again, that's for the target function alone, not the actual required electrical filter which is pretty much arbitrary and thus needs arbitrary control over parameters.
But doesn't it make it easier if you choose a suitable distance between the speaker elements and then play around with various settings? You should probably calculate and simulate the whole thing first, but still. This is what Kimmosto wrote:

Minimum c-c is 1.0 x wave length and maximum about 1.4 x wave length at XO frequency assuming that design is conventional uni-directional box (not open baffle) with phase matching (acoustical 4th order) slopes. Good and quite flexible initial/design value for c-c is 1.2 x wave length at XO, giving smooth combination of power and early reflections i.e. balanced sound without significant power dip at XO due to bump in DI and dip in vertical early reflections. In other words, this concept aims lobe nulls to directions which are the least significant for power response and vertical early reflections - and listener sitting in sweet spot of course.


Incidentally, isn't Kimmosto the creator of VituixCad? The program that PMA (and many others) use. PMA, #1 in this thread:

For acoustical SW measurements, I am using REW software in combination with VituixCad , VituixCad is also used for crossover and acoustical simulations.

Edit
What I quoted from Kimmosto above was with 4th order slopes. So with other order slops, those conditions give other that result. Kimmosto mostly shows the difficulty or even the utopia of having C-C 1/4 wavelength at XO as a design principle. It is practically physically very difficult to build such speakers.
 

Attachments

  • Screenshot_2023-09-21_143155.jpg
    Screenshot_2023-09-21_143155.jpg
    644.4 KB · Views: 44
Last edited:

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,281
Likes
1,469
Location
Budapest
Just a very basic comment/question:

I currently do not need any crossovers since all my systems are using 1-way full range drivers but when I had my FAST system I needed a crossover for the subs
What I did back then (and what I would still do) is that I put my UMIK-1 in the main listening position (since that is what I am only interested in), started an RTA measurement and kept changing the crossover settings until I got the best results. I know this is not very scientific but relatively quickly I got what I wanted.
I kept the frequency fixed (80Hz) and was playing around with the slope (12-18-24-30-36-42-48dB/oct.) and the phase type (minimum-mixed-linear)
I ended up with a mixed phase type filter and was occasionally using linear phase too and the sound was just amazing
(I did not touch the Q value at that time, probably I would do that too if I had to do it again - but I reckon now I would just start with the linear phase type right away)

Couldn't this approach work for this project too?
 
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,649
Likes
10,860
Location
Prague
The optimal distance between the speakers (centers) on the front baffle is 5/4 lambda (lambda at crossover frequency). With small diameter speakers like here it is possible to get.
 

dualazmak

Major Contributor
Forum Donor
Joined
Feb 29, 2020
Messages
2,898
Likes
3,126
Location
Ichihara City, Chiba Prefecture, Japan
This project is not intended to create the SOTA speaker, it is rather about learning and evaluation of SW xover and DSP approach.

We should not forget about this important message from OP @pma; here SW is not subwoofer but software!;)

I know this is ASR Forum, and, as well expected, we already had on this thread much and intensive discussions on "theoretical, mathematical, audio-engineering-physics oriented" approaches and suggestions on XO/EQ.

I myself, however, really would like look forward to the "real air sound measurement data" which will be steadily given by OP @pma in the near future, even whatever intensive theoretical discussions would be given here on this thread.

I well know (and have actually experienced) that very detailed/intensive theoretical variation/tuning of XO/EQ setting(s) in DSP (with "theoretical" Fq and phase curves on GUI screen) would not always be properly reflected/reproduced in real air sound measurement, at near field and/or at our listening position. I dare to say that SP drivers and room air are much lazy analog devices...

I believe our ultimate goal in "home audio exploration" is to enjoy music (not physical sound) in our audio listening room environment where room acoustic really does "matter". The possible final audio setup in our home audio environment, therefore, always would be inevitably settled/decided as "acceptable compromisation" of all the parameters including room tuning. And at least I myself always prefer proper combination of objective measurements and subjective assessments, as well as combination of digital domain tuning plus analog domain tuning (especially for relative gain control, i.e. tone tuning over SP drivers).

I assume that now OP @pma just started his exploration having his present motivation of "learning and evaluation of SW (software) xover and DSP approach" using simply and very cleverly designed 2-way SP system.

I do much respect and hope the steady step-by-step exploration of OP @pma aiming towards "his ideal" music listening audio setup in his room acoustic/environment.
 
Last edited:
  • Like
Reactions: pma
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,649
Likes
10,860
Location
Prague
Hi PMA, that looks great.
Perhaps I missed it, what is the crossover point you ended up with?

Thank you. The reply to your question is not quite simple, as the resulting acoustical characteristics are a mix of electrical and mechano-acoustical transfer functions.

However, the electrical filter scheme I used as a 1st iteration is this:

HT88+MW111_SWxover var1 XO-schema-1.png


... and the corresponding EKIO filters:

EKIO_ver2.png


This is the measured spectrogram:
2-waybox_EKIO_spectrogram.png


This is definitely not the end of the game. Low frequencies below 40Hz have to be cut sharply, as the miniature woofer has hard times when there are such frequencies, which results in great cone excursion + distortion. Sweeps from 20Hz are now impossible, I have to start at 50Hz, even without the 150Hz/6dB shelf filter. I also plan to cut everything above some 15-18kHz, as the horn tweeter behaviour deteriorates there quickly.
 

MAB

Major Contributor
Joined
Nov 15, 2021
Messages
2,179
Likes
4,936
Location
Portland, OR, USA
Thank you. The reply to your question is not quite simple, as the resulting acoustical characteristics are a mix of electrical and mechano-acoustical transfer functions.

However, the electrical filter scheme I used as a 1st iteration is this:

View attachment 313775

... and the corresponding EKIO filters:

View attachment 313779

This is the measured spectrogram:
View attachment 313778

This is definitely not the end of the game. Low frequencies below 40Hz have to be cut sharply, as the miniature woofer has hard times when there are such frequencies, which results in great cone excursion + distortion. Sweeps from 20Hz are now impossible, I have to start at 50Hz, even without the 150Hz/6dB shelf filter. I also plan to cut everything above some 15-18kHz, as the horn tweeter behaviour deteriorates there quickly.
Thanks, very clear. Makes sense, I think I see what you are doing.
Not to throw too many random suggestions... If you have any residual noise with that sensitive tweeter, you can replace your 1st-order DSP filter with a capacitor (5kHz in your 1st iteration example). It would knock down any remaining noise, and provide some measure of protection against issues and accidents.
Not sure what your end purpose is, but if it's desktop/computer the limited bass isn't the worst problem, I even find the vibrations distracting.
I love these software crossovers. Very intuitive and flexible and really quick!
 
Last edited:
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,649
Likes
10,860
Location
Prague
Thanks, very clear. Makes sense, I think I see what you are doing.
Not to throw too many random suggestions... If you have any residual noise with that sensitive tweeter, you can replace your 1st DSP order filter with a capacitor (5kHz in your 1st iteration example). It would knock down any remaining noise, and provide some measure of protection against issues and accidents.
Not sure what your end purpose is, but if it's desktop/computer the limited bass isn't the worst problem, I even find the vibrations distracting.
I love these software crossovers. Very intuitive and flexible and really quick!
Yes, I was first thinking about capacitor instead of the 1st order HP in the scheme, hopefully my amp has very low noise, so there is no issue with the noise from horn tweeter.

I am attaching the individual acoustical responses from the woofer and the tweeter, together with them playing both.

2-waybox_EKIO_acousticfilters.png


You are asking what is the purpose of this box. As mentioned in the post #1, the purpose is to learn to work with SW crossover and to try the horn tweeter. Definitely not intended as my main speaker system :))), it will probably end as a desktop speaker near my PC.

I will also measure the directivity, but the pain is that my project room is very small and rotating the speaker on the stand results in very different reflections, depending the angle of rotation. No Klippel here :D. So they will be only approximate measurements.
 

DanielT

Major Contributor
Joined
Oct 10, 2020
Messages
4,894
Likes
4,877
Location
Sweden - Слава Україні
Thank you. The reply to your question is not quite simple, as the resulting acoustical characteristics are a mix of electrical and mechano-acoustical transfer functions.

However, the electrical filter scheme I used as a 1st iteration is this:

View attachment 313775

... and the corresponding EKIO filters:

View attachment 313779

This is the measured spectrogram:
View attachment 313778

This is definitely not the end of the game. Low frequencies below 40Hz have to be cut sharply, as the miniature woofer has hard times when there are such frequencies, which results in great cone excursion + distortion. Sweeps from 20Hz are now impossible, I have to start at 50Hz, even without the 150Hz/6dB shelf filter. I also plan to cut everything above some 15-18kHz, as the horn tweeter behaviour deteriorates there quickly.
Good job PMA. :D

I have nothing to add on the subject itself, but I am interested in your tinkering. Interesting solution you are working on.:)
 

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,533
Likes
18,607
Location
Netherlands
This is the normalized horizontal directivity, measured under my limited conditions. We can see that the horn tweeter is pretty directional.

View attachment 313809
I think this is as messy (or clean ;) ) as can be expected. I would have expected that the little woofer would be a bit wider at 2 kHz. You may try to lower the crossover point towards 2 kHz to improve things a bit.
 
  • Like
Reactions: pma
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,649
Likes
10,860
Location
Prague
I think this is as messy (or clean ;) ) as can be expected. I would have expected that the little woofer would be a bit wider at 2 kHz. You may try to lower the crossover point towards 2 kHz to improve things a bit.
Unfortunately, resonant frequency of the horn tweeter is 2.5kHz and recommended crossover frequency is 4kHz (datasheet). It does not work nice near 2kHz (distortion). Both drivers need very sharp cut - the woofer has break-ups above 4kHz and tweeter is messy below 3kHz.
 

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,533
Likes
18,607
Location
Netherlands
Unfortunately, resonant frequency of the horn tweeter is 2.5kHz and recommended crossover frequency is 4kHz (datasheet). It does not work nice near 2kHz (distortion).
Ah, do you have the distortion plot? Compression drivers can sometimes be used below its resonant frequency, but it depends highly on the particular design.
 
OP
pma

pma

Major Contributor
Joined
Feb 23, 2019
Messages
4,649
Likes
10,860
Location
Prague
Ah, do you have the distortion plot? Compression drivers can sometimes be used below its resonant frequency, but it depends highly on the particular design.
Yes, I have it, as a part of standard REW measurements ;).

HT-88 direct.png

Tweeter FR, no xover. Avoid area below 3kHz.
 
Last edited:

voodooless

Grand Contributor
Forum Donor
Joined
Jun 16, 2020
Messages
10,533
Likes
18,607
Location
Netherlands
Yes, I have it, as a part of standard REW measurements ;).
Obviously ;) But can you show it? I'm anyway interested in what these cheap drivers can do distortion-wise.

Monacor made some very good low-distortion drivers, like those:


But it seems nowadays they are quite expensive... I'm sure I didn't pay that much about 20 years ago.

Looking at the linearity of the little woofer, it might be quite good.

As far as tweeters go, the DT-300 has a bit of fame as I recall. It's not much more expensive than the one you have right now.
 

Thomas_A

Major Contributor
Forum Donor
Joined
Jun 20, 2019
Messages
3,506
Likes
2,544
Location
Sweden
This is the normalized horizontal directivity, measured under my limited conditions. We can see that the horn tweeter is pretty directional.

View attachment 313809
One "problem" I see here is the off-axis dip 1-2 kHz vs peaking at 3-4 kHz. This will exaggerate the stereo system errors since reflections in the 1-2 kHz range is a way to "fill in" the stereo system error dip. It is no solution to the problem but it is a way to milder it. I would suggest listening in stereo in a normal room and play with a bit of increase 1-2 kHz and a small dip at 3-4 kHz, but still being within +/- 1.5 dB on-axis or so.
 
Top Bottom