Experiments in speaker design for cardioid radiation pattern

[Vineethkumar01]

Next iteration of the cardioid mid 3way speaker

 

[diyaudio]

You have changed the design, so measurements need to be taken again

 

[Vineethkumar01]

You have changed the design, so measurements need to be taken again

Basic principles remain the same though. The leaking slot behind the driver has similar area as the previous enclosure shape. The tweeter is still not affected due to the properly terminated waveguide formed by the mid driver + its edge rounded housing. The leaking slot port will cause cancellation and cardioid like radiation in the mids and hence reduce the impact from whatever enclosure is behind it.
But you are right in the sense that if one wants to know the absolute impact, one has to measure again.. However I expect it to be similar to the previous enclosure

 

[diyaudio]

In your new design the driver is naked. This creates a very small distance between the front and rear source. This will give decreased efficiency.
Ideally, the distance should be 1/4wl of the highest frequency of operation of cardioid. Thus you should increase the distance after calculating it and put a 'spacer' on the rear to achieve it

 

[Vineethkumar01]

Thanks for your suggestions..

In your new design the driver is naked.

Actually, it is not that naked.
The driver is housed in a 36mm thick circular disc with rounding at the edges where the driver frame meets the disc. This is to act as a proper termination for the waveguide formed by the mid driver for the tweeter. Here is the zoomed in pic of the relevant area of the cabinet

So the distance between the front and rear source just based on the above geometry is let's say atleast 30mm. The gap will anyway be filled with 50mm thick melamine foam rings as in by xps foam prototype.

This will give decreased efficiency.

Can you please explain what do you mean by efficiency in this context. Is it the efficiency of the implementation of the cardioid polar pattern?

Ideally, the distance should be 1/4wl of the highest frequency of operation of cardioid.

Can you let me know how the 1/4 wl rule is obtained in the context of cardiod implementation. Is that consideration to have roughly co-located sources so that when the dipole meets the monopole, we get cardioid?

Thus you should increase the distance after calculating it and put a 'spacer' on the rear to achieve it

So as per the 1/4wl rule, I will get cardioid working to approx 2800Hz. which is more than enough since
1) The driver would have started to beam by then and the polar pattern width narrows fast.
Here is a sim that shows how a flat piston approximation of the driver would behave w.r.t polar pattern width in the above type baffle.

Diffraction sim

Polar pattern (normalized) from above configuration


2) The intended crossover with the tweeter (of this coaxial driver) is around 2.5 kHz. Hence from there on, the tweeter will control directivity

 

[diyaudio]

Can you please explain what do you mean by efficiency in this context. Is it the efficiency of the implementation of the cardioid polar pattern?

Can you let me know how the 1/4 wl rule is obtained in the context of cardiod implementation. Is that consideration to have roughly co-located sources so that when the dipole meets the monopole, we get cardioid?

See, lets say your highest frequency of operation is 2.5khz then 1/4wavelengh is 3.45 cms. This should be the front to rear distance. If distance is higher than this then you will not get good polars, if distance is smaller than this then you will get higher loss due to smaller distance than needed, hence lower efficiency.

 

[Vineethkumar01]

So many months later, here is the status of this project. A lot of hard work went in. With the help & hard work of @sadik, I got my dream cabinets built. He really makes speaker dreams come true..
Some pics of the build:

 

[Vineethkumar01]

Painting pics

 

[Vineethkumar01]

And now they are home and its my turn to complete the build

To they are ready for measurements

Back of the coaxial driver filled with 55mm thick melamine foam covered with a sock cloth ring (for now). This is for the cardioid midrange

 

[Vineethkumar01]

Some starter measurements

Woofer free air vs in box impedance measurements

Woofer nearfield + impedance in box (not baffle step adjusted)

Sica mid and tweeter impedance (in box)

This much is all I have got today.. The far-field polar measurements will tell us whether the cardiod midrange works (or not)... Fingers crossed for that

Whatever happens, I really love the looks of this speaker..

 

[Kannan]

Wow...this is some build.....Salutations. Neenga Vera Level Saar

 

[Vineethkumar01]

The speakers are measured (horizontal polars 0 to 180 degrees)

Mid driver (measurements beyond 160 degrees are probably garbage)

tweeter

Woofer

Now a first attempt crossover (reference axis is 10 degrees off axis)

I think it is all working as expected..

 

[Vineethkumar01]

I did the measurements on the coax driver again due to the hint of dipole behaviour in the mids. Turned out that it was a measurement error and the actual speaker behaviour is even better in terms of cardioid midrange behaviour. Here is a linear phase crossover based on new measurements.

A passive crossover version is also in the works

The mysterious right triangle step response

 

[grindstone]

Yeah that one is lined-up better. Nice work. I know you're chasing several technical parameters but it's looking like the forward lobe is pretty close to dead-flat so that's nicely done, too.

What I've been wondering about these things Vineeth is where the mid-cone-to-cab-front length dimension shows-up (thinking it should) as a squiggle in the Z or in the FR of the tweeter. __Maybe__ (??) ~3k5 which'd be about a half-wave of 1.93" ? Or maybe you have that whole space loaded with foam or something?

Good to see it seems speakers are multiplying in your home...

Been wondering about you guys (you and @aeroash) and how the big box is going. You guys are always most dangerous when you're not posting Carry on and be well!

 

[Vineethkumar01]

Hi @grindstone
The space behind the back of the coax driver is filled with 55mm thick basotect melamine foam (under the sock cloth) like below

As you might already know, it delays the low frequencies & absorbs the highs, helping create the cardioid radiation pattern in the mids. So the backwave from the mid-cone is also most probably absorbed in the above 1.5k region I think. The sound absorption coefficient of this material is shown below. We can see that 50mm melamine has good absorption above1kHz.

Maybe that is why no impedance wiggles are seen above 1kHz due to the back wave reflection in the midcone's impedance plot shown below?



I am trying to make space for all these speakers at home.. Each one has different flavour to the sound and I like them all

The big boxes are completed and singing well at @aeroash's home. I think he is on travel, currently. Currently the speaker uses a DSP crossover. He has plans to experiment with a passive crossover as well once he gets time..

 

[Vineethkumar01]

This speaker is now alive and singing..
I tested this crossover first

Here is the reference axis (20 degrees off-axis) measurement I made of the speaker from 50 inches away.
It is almost exactly as per the predicted response in VituixCAD.
The blue line is what I took as the reference based on Vituixcad

So far I like it a lot.. But honeymoon days.. I Need to listen a lot more..
The vocals sounding special is one of the first feelings I got, while listening.
I am missing a bit of bass that I am used to with my big speakers but maybe more tuning needed..
More updates in coming days as I listen more

If anyone is wondering how the speaker looks like (sounds like ) while it is singing, and also want to get some vague idea about the cardioid directivity, see the below video. Offcourse the sound balance will just depend on what you play this video on, but still...

 

[Vineethkumar01]

Work on both speakers is complete now..
Here they are (for now)

4.2ms gated measurements of the L & R speakers (each taken at 20 degrees off axis)

Measurements of L speaker, R speaker & L+R speakers from listening position (on that chair)

Step response at listening position for L+R speakers

Frequency magnitude & phase response at MLP

ETC at MLP

On the chair, imaging is superb... It feels like sound is coming from behind the monitor in front of me
Time to seriously listen to some music now..

 

[Vineethkumar01]

Yet another flawed attempt to capture the sound using my mobile phone from the listening position.
But let this be here
Need to get a binaural mic diyed

 

[grindstone]

Thanks for the reply earlier and glad you are listening now You've covered more DIY'er ground faster than anybody I've ever seen and I can't say enough about your enthusiasm and commitment. Thanks for sharing your projects.

 

[Vineethkumar01]

A big cardioid 2way + subwoofer concept

The cardioid enclosure for a 15inch driver to house the 15PR400 driver + an EXAR 400 horn
The box is only about 26cm deep from the 'baffle surface'
I like the looks of this
A subwoofer (for crossing over below 80-100 Hz is not shown)