Building a 2-way small active speaker with software crossover

[pma]

On-axis frequency response with the bass extension. The sound is the sound of a typical mini-monitor, neutral, detailed, non-coloured. Still, it would need a subwoofer. I will not bother with drilling a port hole. The vented box has different drawbacks, like port midrange leakage resulting in distortion and coloration.

 

[pma]

Further development .... the speaker have not had properly solved the baffle step issue yet, the correction used was insufficient, resulting in a bit "thin" sound, though clean. The Edge baffle simulator suggested this:

so I have inserted another shelf filter in the EKIO low pass section, which resulted in this:

The SW crossover was checked by the loopback DAC-EKIO-ADC method:

Now, the merged frequency response (bass nearfield+baffle step, with farfield)

and in-room response like this:

Just for those who might like to see more plots, this is the step response measured at the mid-woofer:

And this is the far-field response to 1kHz CEA burst:

 

[dualazmak]

Thank you indeed for sharing your continuing intensive works/experiments with us.

Just let me have a simple question, please, only if acceptable for you.
Do you have plan to add sub-woofer, even one unit only, with your present 2-way experimental SP system?

I really looking forward to learning from your further (experimental?) efforts in EKIO XO/EQ configurations with sub-woofer which would cover around 15 Hz to 80 Hz.

 

[pma]

Thank you indeed for sharing your continuing intensive works/experiments with us.

Just let me have a simple question, please, only if acceptable for you.
Do you have plan to add sub-woofer, even one unit only, with your present 2-way experimental SP system?

I really looking forward to learning from your further (experimental?) efforts in EKIO XO/EQ configurations with sub-woofer which would cover around 15 Hz to 80 Hz.

I am thinking about it, but I have not made a final decision. My space for audio is quite limited, so I will see.

 

[pma]

Comparison of in-room measurement (brown) with merged near+baffle simulation+farfield (blue)

 

[pma]

I really looking forward to learning from your further (experimental?) efforts in EKIO XO/EQ configurations with sub-woofer which would cover around 15 Hz to 80 Hz.

I have added a subwoofer with this FR:

It adds one octave down to this small 2-way speaker. As my testing room/workshop is very small, like 4m x 2m, this extension is just about right and it helps the small speaker a lot.

 

[dualazmak]

Just wonderful! Now you can go down to around 40 Hz with almost flat response and (hopefully) in very low distortion; I understand well about your present room limitations.

Just for your reference, at least in my case, I need and would like to go down to 16 Hz as shared here and here; it sometimes causes, however, the low-Fq (air conditioning and/or other) recording QC noise issues as shared here though...

 

[pma]

Just wonderful! Now you can go down to around 40 Hz with almost flat response and (hopefully) in very low distortion; I understand well about your present room limitations.

Thank you, the resulting frequency response is reasonably flat (at low frequencies of course depending on placement of the speakers) with acceptable distortion from 40Hz up.

Interestingly enough, tuning gain between subwoofer and the speaker "by ear" resulted in the response later verified as "flat enough" by measurements.

Normally I do not listen to music with much content below 40Hz. I understand the issues you may have with audibility of recorded air con and similar noises and rumble.

 

[pma]

This is the interconnection schematics picture. Please note some important notes - to prevent ground loop induced noises.

 

[dualazmak]

Interestingly enough, tuning gain between subwoofer and the speaker "by ear" resulted in the response later verified as "flat enough" by measurements.

Indeed, I also often experienced similar situation. Yes, we should and we can trust our ears and brain at the final stage(s) of our audio exploration, then our final measurement(s) would hopefully just confirm and support your subjective final tune-up.

And,,, the flat (or near flat) Fq response would not always give us best subjective sensation and/or best subjective sound quality since subjective impression/feeling greatly varies depending on characteristics of SP drivers and our complex room modes at listening position.

 

[pma]

Electrical measurements of filter responses at outputs of amplifiers used (speakers connected)

 

[dualazmak]

Very nice to see the result; thank you for sharing.

I too measured Fq response at various "stages" of multichannel setup; EKIO's digital output level (by digital loop back), DAC8PRO's analog out level, amplifiers' SP out level before and after the protection caps, and finally real air sound level at listening position (ref. summary here).

I believe these Fq response measurements you are doing (and I have done) at various "stages" would be really nice and worthwhile giving invaluable information/insights for fine house-curve tuning of our multichannel audio setups.

 

[weme]

I have read through the entries on your project and liked it very much:
Your project notes for people who have - at least theoretically - built a speaker box with a simulation program, the use of the DSP program EKIO for control.
The funny comments from @restorer-john on your "Two Eggs matches". Be careful, restorer-john always finds something in photos!
And the marketing expert who always manages to draw attention to his projects.

The bass Magnat MW 111 CP 754-I was or still is built into the Magnat shelf speaker Monitor Supreme 102:
Power handling: (RMS / Max.) 60 / 120 Watt(s)
Frequency Response: 42 – 36 000 Hz
SPL: (2.8V/1m) 89 dB
Dimensions: (WxHxD) 149 x 250 x 190 mm

Have you measured Xmax and if so how?

If you want to use your two-way speakers without an additional bass box:
A smaller distance between the speakers and your ears and a deeper bass tuning with the current speakers or with better speakers.
The front can be cut off very easily with a suitable circular saw and the new front(s) could be screwed in. With screw-in nuts you could use (black painted) cylinder head screws.

For comparison, the maximum possible output of a 4" bass speaker: SPL measurement of the Neumann KH 80 DSP in its bass reflex cabinet.

Source: https://www.neumann.com/globalassets/digizuite/17630-en-sodru_sr_092017_t_neumann_kh_80_eng.pdf

 

[dualazmak]

Hello OP @pma and friends,

Just for your possible interest and reference regarding ASIO I/O routing into/from DSP EKIO...

I recently started test and evaluation of ASIO (and other digital audio) routing software VB-Audio MATRIX;

- VB-Audio MATRIX (VASIO64A setup) as system-wide ASIO and other audio I/O routing center: Replacement for ASIO4ALL and VB-Audio Virtual/HiFi Cables: #851
- VB-Audio MATRIX (VASIO64A setup) as system-wide ASIO and other audio I/O routing center: Replacement for ASIO4ALL and VB-Audio Virtual/HiFi Cables: Part-2_Further optimization in conformity with present configuration of DSP "EKIO": #854

We also have the dedicated thread entitled "VB Matrix" kindly started by @boxerfan88.

VB-Audio Matrix is so far so nice stable and robust in my multichannel setup for ASIO routing in 44.1 kHz, 48 kHz, 88.2 kHz, 96 kHz, 176.2 kHz and 192 kHz together with JRiver MC and EKIO.

 

[pma]

I have re-built this small speaker a bit. The Magnat small woofer MW111 is replaced by Visaton W100 X

https://doc.soundimports.nl/pdf/brands/Visaton/W-100-X/W-100-X-datasheet.pdf

I have used passive xover + rePhase filter to equalize bass. The resulting in-room on axis response looks like this now:

 

[voodooless]

What was the motivation for the change? The Visaton W-series are generally quite excellent for what they cost, so no complaints there

 

[ppataki]

I have re-built this small speaker a bit. The Magnat small woofer MW111 is replaced by Visaton W100 X

https://doc.soundimports.nl/pdf/brands/Visaton/W-100-X/W-100-X-datasheet.pdf

I have used passive xover + rePhase filter to equalize bass. The resulting in-room on axis response looks like this now:

View attachment 383930

View attachment 383931

This might have been discussed already but what was the reason for not going full active? (like using hardware or software based digital crossover)