Introducing Directiva - An ASR open source platform speaker project

[ctrl]

I think most people would be interested in at least getting ~100 dB to 105 dB of peak output @ 1 meter with realistic amounts of amplification (~100 to 300W with [email protected]).

This should be possible.

According to the manufacturer, the SB15NBAC-8 delivers about 85-86dB, the 4 ohm version about 88dB in infinite baffle @2.83V above 200Hz.

As @TimVG said, the baffle step, for the speaker top, is about -4dB around 200Hz compared to the infinite baffle. So in the worst case the driver in the top speaker cabinet reaches about [email protected]@1m@>=200Hz.


This number also depends somewhat on the baffle size of the bass module. With a 0.26m wide bass module (0.8m height), the sound pressure around 200Hz increases by a little more than 1dB.

Thus, the baffle step loss @200Hz, compared to the infinite baffle, should be a little less than 3dB.

This would put us at around 83dB@200Hz@[email protected] for the eight ohm driver and around 85dB for the four ohm driver.
If the bass module is designed powerful enough, the sound pressure level can also increase by one or more dB, depending on the crossover tuning (slight low bass hump).


In addition, there is another effect which, according to the simulation, is caused by the lateral slots.
The slots radiate the phase-inverted sound from the rear of the driver. Due to the damping and the detour caused by the construction, a frequency-dependent phase shift of the phase-inverted sound occurs.

This leads to the fact that, according to the simulation, the sound pressure level of the driver and slots, in certain frequency ranges, is higher than the driver alone.
Here is the simulation of the speaker top axis frequency response of woofer plus slots:
blue - woofer alone
green - slots alone
red - summed sound pressure level of the speaker top

At 300Hz, the sound pressure level is increased by about additional 1.5dB compared to the woofer alone.

The achievable peak sound pressure level is therefore, depending on the bass module, on the level of large floorstanding speakers, such as a Revel F328be.


Excursus:
You can also see (in the graphic above) very well why the crossover frequency should not fall below 200Hz, because there is sound cancellation due to the phase shift between the driver and slots and therefore the harmonic distortion will increase significantly (when the sound pressure level of the speaker top is linearized below 200Hz).
At 100Hz the cancellation is over -4dB compared to the woofer alone.

It could be that some slotted speaker manufacturer may have overlooked this "feature"

 

[Rick Sykora]

Would it be an idea, to built a bass module more or less similar the Genelec W371?

It would, but not likely in the current plan unless somebody signs on to do.

This is why Directiva r2 was done is modular fashion. We expect the community will continue to modify and improve over time.

 

[D!sco]

I really like the curved front baffle. Beautiful work.

Can’t wait to see the explanation for cardioid dispersion once the design is finalized. It would be nice if @kimmosto could apply cardioid enclosures to the designer in VCAD somehow.

 

[dwkdnvr]

Wow. I haven't been paying close attention to the ongoing work, but this definitely has me intrigued. I'd 'need' a 200Hz xover since I have rythmik FM8's and I think that's as high as they can go, but if I could somehow, miraculously, find the time to undertake the projects comparing my Kef R3's with the M105 clone and with this cardiod V2 design would be really very interesting.

Are you using the 4ohm or 8ohm SB15?

 

[TimVG]

Before we go on, I want to give a shoutout to @ctrl. Without his work behind the scenes, there's no way this prototype would have turned out as it did . It's truly amazing how close the simulations came to the measurements, given that as far as I'm aware, he figured out how to simulate a passive cardioid system on his own, after I sacrificed a small cheap speaker. It usually takes a whole of trial & error.

 

[abdo123]

Before we go on, I want to give a shoutout to @ctrl. Without his work behind the scenes, there's no way this prototype would have turned out as it did . It's truly amazing how close the simulations came to the measurements, given that as far as I'm aware, he figured out how to simulate a passive cardioid system on his own, after I sacrificed a small cheap speaker. It usually takes a whole of trial & error.

@Rick Sykora @ctrl It is indescribable how amazing this turned out in such a short amount of time and with such a limited number of people. Congratulation to everyone involved, really, because this might just be a new industry standard in terms of price to performance.

The best part about all of this is that I think the directivity will even measure better on a high resolution system like the NFS simply because everything has been modeled to perfection. really astounding work honestly!

 

[Madjalapeno]

Certainly looks impressive - well done. I can't wait to try it. I'm guessing this will need a 4 channel amp to drive the top section?

 

[Vineeth Kumar V]

Excursus:
You can also see (in the graphic above) very well why the crossover frequency should not fall below 200Hz, because there is sound cancellation due to the phase shift between the driver and slots and therefore the harmonic distortion will increase significantly (when the sound pressure level of the speaker top is linearized below 200Hz).
At 100Hz the cancellation is over -4dB compared to the woofer alone.

Hi ctrl,
Could you explain a bit more about why harmonic distortion increases significantly below 200 Hz with SPL linearization in this particular case. Is it because of the driver's inherent harmonic distortion specs relatively bad below 300 Hz or is it something else?

Thanks

 

[Rick Sykora]

Certainly looks impressive - well done. I can't wait to try it. I'm guessing this will need a 4 channel amp to drive the top section?

The answer will likely be up to you. The plan is that the monitor section will utilize a passive crossover.

 

[abdo123]

Hi ctrl,
Could you explain a bit more about why harmonic distortion increases significantly below 200 Hz with SPL linearization in this particular case. Is it because of the driver's inherent harmonic distortion specs relatively bad below 300 Hz or is it something else?

Thanks

If i understand correctly the slots are kind of behaving like a port tuned at 200Hz so output above that frequency is boosted and output below it is reduced.

So if the response is equalized to be flat from 100Hz up the sensitivity will be in the realm of [email protected] or the distortion will be considerably high at 100Hz to 200Hz because of the amount of boost used.

 

[fluid]

If i understand correctly the slots are kind of behaving like a port tuned at 200Hz so output above that frequency is boosted and output below it is reduced.

So if the response is equalized to be flat from 100Hz up the sensitivity will be in the realm of [email protected] and the distortion will be considerably high at 100Hz to 200Hz because of the amount of boost used.

Kind of, a gradient source created from cancellation or any other means has a peak in the response before a steeper rolloff starts, just like a dipole.

Image of a generalised first order cardioid response to demonstrate

Equalising that rolloff through boost causes excursion to rise which results in higher levels of distortion from the driver.

The driver itself is not very large and has a fast rise in distortion below 200Hz so more excursion is not wanted there.

 

[Rick Sykora]

@Rick Sykora @ctrl It is indescribable how amazing this turned out in such a short amount of time and with such a limited number of people. Congratulation to everyone involved, really, because this might just be a new industry standard in terms of price to performance.

The best part about all of this is that I think the directivity will even measure better on a high resolution system like the NFS simply because everything has been modeled to perfection. really astounding work honestly!

Thanks! It was a team effort, but @TimVG took what was proposed and worked the slotted design magic. He is the lead for it and deserves much of the credit. I will make a simplified copy of it and then the bass module(s) will be integrated. The plan is to design 2 bass modules as stated earlier. I expect to build one or more here. Tim has plans to build the smaller one if it fits his decor.

Since this is a sneak peek, plans may change. The team has already considered using the 4 ohm version of the woofer and so am planning to use it for my build. As @ctrl mentioned, should give the monitor about another 2-3 dB more sensitivity. Since the monitor crossover for it is not yet final, we still have see whether it is fully passive or a hybrid. As with r1, we will likely have options, but only plan to have Amir to test one.

 

[D!sco]

Are there plans to submit a passive R1 to Amir for testing?

 

[Rick Sykora]

Are there plans to submit a passive R1 to Amir for testing?

not any planned, but will do a full spin myself in the next few months.

 

[Zvu]

. As with r1, we will likely have options, but only plan to have Amir to test one.

I'm sure that Erin wouldn't say no if you were to ask him to test the other variants of directiva.

 

[mcdn]

Oh wow, this looks like pure genius! I love how you've taken all the community feedback and created something so interesting and capable, and in such a short time. Best of all it's not something available on the open market at any price. I can't wait to sit this on top of my Linkwitz LXSub4 dipole woofers...

 

[TimVG]

I'm sure that Erin wouldn't say no if you were to ask him to test the other variants of directiva.

I'm pretty sure Amir would test different variations if Rick would ask, however I believe what Rick implies is that it's simply unnecessary from a measurement pov, since the raw drivers already get measured on the NFS in the correct cabinet. It's been established once you get driver measurements which are correct, the simulations are very reliable. Basically it comes down to having to spend more $$ simply to get Amir's listening impressions.

 

[Rick Sykora]

I'm sure that Erin wouldn't say no if you were to ask him to test the other variants of directiva.

Perhaps, but only if Amir approves. In an outdoor setting, my measurements are pretty close to Amir’s. Along with shipping cost, Erin has comparable issues for testing too. He is busy enough that we would be displacing the opportunity to test another speaker too.

 

[Zvu]

Perhaps, but only if Amir approves. In an outdoor setting, my measurements are pretty close to Amir’s. Along with shipping cost, Erin has comparable issues for testing too. He is busy enough that we would be displacing the opportunity to test another speaker too.

But wait, this is an open source project right ?

One day someone will make a version not measured by Amir and he will send it to Erin for measurement if Amir wouldn't want to do it. For an open source project it is inevitable. Having full set of measurements for all Directiva variants should be the goal in my opinion.

I'm not interested so much in spinorama done by Klippel NFS (napilopez proved quite a few times that you can do that with great success without nfs) but all other measured data that is hard to do and requires calibrated system (distortion, csd, compression etc.). I'm predicting quite the difference comparing fully active vs fully passive option.

 

[abdo123]

But wait, this is an open source project right ?

One day someone will make a version not measured by Amir and he will send it to Erin for measurement if Amir wouldn't want to do it. For an open source project it is inevitable. Having full set of measurements for all Directiva variants should be the goal in my opinion.

I'm not interested so much in spinorama done by Klippel NFS (napilopez proved quite a few times that you can do that with great success without nfs) but all other measured data that is hard to do and requires calibrated system (distortion, csd, compression etc.). I'm predicting quite the difference comparing fully active vs fully passive option.

I have to disagree, the resolution in the mid-range is non-existent for napilopez's spins, and that's where our hearing is most sensitive.