Directiva r2 project: market requirements gathering

[RickS]

Unless you use a very wide baffle no speaker will have much directivity below 1000Hz. The wavelength at that frequency is longer than the baffle width hence it doesn’t block the driver. Speaker is in a 4Pi space.

True with a traditional monopole...

As mentioned earlier is the thread, there are others that are deploying dipoles and other techniques (Kii and B&O use multiple drivers). They do not follow the traditional path either. I have seen some good extended directivity results from gainphile's OB designs. These often depoy OB woofers too.

Am still busy with r1 and the upcoming wedding. Will drive a decision some time next week. This is comparable to r1 at this stage, so am not worried. We'll figure it out.

 

[JRS]

there is no real life example (maybe the Tom Danley Synergy horn without the horn?) I'm just curious if this is a good way to go to have multiple woofers in a baffle without having it be a floor stander.

I understand that the industry shifted away from having wide baffles for a reason that i don't remember right now, but that was in the 80s, we have technology now that they didn't have access to back then.

Mostly about having a slim footprint that wouldn't dominate the living space, along with the popularity of the D'Appolito configuration and the like. Call it WAF but I prefer the look myself, especially in a small moderate sized room--and yes, improvements in midbass drivers was key.

 

[fluid]

Unless you use a very wide baffle no speaker will have much directivity below 1000Hz. The wavelength at that frequency is longer than the baffle width hence it doesn’t block the driver. Speaker is in a 4Pi space.

This isn't entirely right as the entire enclosure does have an effect on the mid/low frequency directivity. 3D simulations with ABEC show the depth and width of the enclosure combined can have more effect on the directivity control. You can see it the measurements of Revel and other bookshelf speakers here. The effect is not as strong as the baffle alone but it is there.

any of these in particular you recommend over the others?

This book from Newell and Holland is good for getting started on understanding speaker design
https://www.taylorfrancis.com/books...9202/loudspeakers-philip-newell-keith-holland

If you want to see more combinations of drivers simulated and arrayed then have a look at this thread, it is long and started about something else but pretty much every conceivable option has been tried with a focus on creating smooth narrow vertical directivity
https://www.diyaudio.com/forums/ful...-array-wall-corner-placement-post6772349.html

 

[Lbstyling]

Unless you use a very wide baffle no speaker will have much directivity below 1000Hz. The wavelength at that frequency is longer than the baffle width hence it doesn’t block the driver. Speaker is in a 4Pi space.

I believe he is semi referencing cardioid which potentially fixes this.

Your point reinforces my last message on the exact wording for a poll being important

 

[fluid]

I believe he is semi referencing cardioid which potentially fixes this.

From the words used it doesn't seem so to me

Your point reinforces my last message on the exact wording for a poll being important

Yes poll results could easily be skewed by misunderstanding or a lack of clarity in the description. One of the downsides of design by popular consensus

 

[RickS]

I believe he is semi referencing cardioid which potentially fixes this.

Your point reinforces my last message on the exact wording for a poll being important

Correct. The poll would refer to what feature is desired - improved directivity. How to get improved directivity is a design team task.

 

[Tangband]

Correct. The poll would refer to what feature is desired - improved directivity. How to get improved directivity would is a design team task.

Interesting project .

Have you already decided which drivers to be used ?
A very good loudspeaker can have :
SB 26adc with a waveguide.
SB NBAC 15 as a mid.
Seas R0y as woofer.

Dsp crossovers would be like 250 Hz and 2,2 KHz in a floorstander.
The Seas R0y ( low Q version ) should be in a closed box with LT . The response inroom with LT would reach down to 15 Hz.

The baffle widh should be about 32-34 cm wide , The whole loudspeaker about 110 cm high, thus using minimal baffle step correction down to 250 Hz .

If you’re gonna use coaxials is my question: why ? There is not a lot of good alternatives and you have to cross it higher than 400 Hz to avoid IM distortion when the waveguide acts as a variable waveguide for the tweeter. And the optimal shape of a tweeter waveguide is not exactly the same as a coaxial , circle construction.

I guess you know all this , and maybe its temting to go the Genelec coaxial route but there is really no good drivers available in the market to buy.
Revel has already made a speaker like this….

 

[TimVG]

It is not possible to have a cardioid design while using a dome/cone tweeter. High frequency sound waves will always be blocked by the enclosure.

Instead of using potentially confusing terms we could use somewhat more objective terminology, such as this case beamwidth control down to ~500Hz.

 

[Trdat]

I would suggest a true 4 way with the lowest frequencies designed for the subwoofer. So, essentially the 8/7inch dual woofer can play to 40hz but with a closed roll off. I'm thinking emulating a Revel or Elsinore type speaker waveguide tweeter, 2 mids and 2 upper bass woofers, this will allow us to add a sub making it a pristine speaker.

Definitely controlled dispersion and possibly narrow directivity, lets also consider a respectable sized horn

And DSP please.

 

[temps]

Interesting project .

Have you already decided which drivers to be used ?
A very good loudspeaker can have :
SB 26adc with a waveguide.
SB NBAC 15 as a mid.
Seas R0y as woofer.

Dsp crossovers would be like 250 Hz and 2,2 KHz in a floorstander.
The Seas R0y ( low Q version ) should be in a closed box with LT . The response inroom with LT would reach down to 15 Hz.

The baffle widh should be about 32-34 cm wide , The whole loudspeaker about 110 cm high, thus using minimal baffle step correction down to 250 Hz .

If you’re gonna use coaxials is my question: why ? There is not a lot of good alternatives and you have to cross it higher than 400 Hz to avoid IM distortion when the waveguide acts as a variable waveguide for the tweeter. And the optimal shape of a tweeter waveguide is not exactly the same as a coaxial , circle construction.

I guess you know all this , and maybe its temting to go the Genelec coaxial route but there is really no good drivers available in the market to buy.
Revel has already made a speaker like this….View attachment 158976

I agree, DIY options for coaxial drivers are simply too poor. Many major OEMs struggle with their coaxials, with really only a couple notable exceptions. DIY we stand no chance...

tbh the most appealing idea seems to be a W371 style bass extender to go along with a Directiva R1 (or really, any bookshelf). You already have an excellent design -- why throw it out? Iterate, don't reinvent.

R2 could be the bass tower - then a hypothetical R3 could be a revision of R1 with less of the BoM allocated to the woofer, and more to the tweeter, as R2 + R3 would have much lighter demands on the woofer, eliminating the requirement for the ultra-dollar Purifi woofer while still equaling its low distortion performance.

 

[JRS]

Also agree that we should pass on the coax. Granted there are benefits, but this project is now well beyond throwing together a two way w/ separate bass bin. So I'm more inclined to go with some readily available over performing conventional drivers. Big fan of SB (and Satori if we need a 5" or 6.5" driver).

 

[Mudjock]

If you're looking for value in a woofer, the Dayton SD215A-88 is something to think about. A quad is $150. There have been several recent instances where they have been used in well-received 2-way designs. I used it's big brother, the SD270, in my Indium 7 design and it significantly exceeded expectations.

 

[Scgorg]

Unless you use a very wide baffle no speaker will have much directivity below 1000Hz. The wavelength at that frequency is longer than the baffle width hence it doesn’t block the driver. Speaker is in a 4Pi space

Edit: I just saw the post that TimVG wrote about confusing terminology, and I agree.

I'll do my best to clear up what seems to be a slight disconnect/confusion about the "cardioid" talk (from both sides, not just you).
Directivity control to a low frequency is achieveable by using a cardioid design, either by woofers wired out of a phase with a time delay on the side and/or back, or by passive means using a leaky slot. I am assuming you know this, but it's worthwhile context for other readers.

You have correctly noted that directivity control this way is challenging (impossible) to do high in frequency, primarily because of the width of the baffle. There are limits to how slim you can make the baffle, and the baffle necessarily needs to be small relative to the wavelength of the frequency we wish to cancel.

However; directivity control by the means of a waveguide, if looked at with a polar plot, will have a shape that is not too unlike a cardioid solution: low pressure 180 degrees from the front, and gradually sloping before that as you move off-axis. Attached is an image of a polar plot of the gradient helsinki 1.5, as measured by Princeton University (scroll down). This is a design with a passive cardioid midwoofer and a waveguided tweeter. At 5khz (the lowest frequency the tweeter was measured at before being crossed over to the woofer) we can see that the tweeter, as you move very far off-axis, drops by close to 20dB.

But a 20dB drop is not an infinite drop, which is what a theoretically ideal cardioid should have 180 degrees off-axis. "Luckily" for us, we won't get a perfect cardioid performance from the actual cardioid aligment either, due to various losses. I believe the designer of the dutch and dutch has mentioned that he can get a passive cardioid down to about -20dB attenuation at the back. More is possible with active solutions of course, and the passive gradient seems to have about 30dB of attenuation at 500hz (no measurements below that).

The end result, if we combine these two techniques to control directivity, is a speaker that has a "cardioid-ish" response over a wide bandwidth. In the case of dutch and dutch, about 150 to 20khz. Technically it isn't a cardioid across the whole bandwith (or, if we're being pedantic, at any point), but the behaviour is similar enough. You don't have to take my word for it, Erin of Erin's audio corner has measured it, as can be seen in this polar map:

Gradient measurements by Princeton U:

 

[briskly]

I would like to see a closer investigation of the coincident drivers from Sica for MF/HF duty. At the same time, I don't know of any retailers carrying those drivers in the United States.

Something to keep in mind is that these methods to increase LF directivity will compromise the low-frequency efficiency, particularly the dipole arrangement. The phase shift from path lengths and simple delays will eventually decrease at LF, but the loss in acoustical output from a damped vent will force a higher x-over than a standard closed alignment.

 

[D!sco]

I think I've said before that one of my favorite aspects of the R1 build is the use of a prefabricated box. There are lots of clever adjustments that can be made to existing designs, one of my favorite examples being the Nephila, which uses the Denovo Tower and lifts the baffle for a dipole tweeter. If this goes in the direction of Cardioid, are there any existing boxes that would accommodate this?

 

[TimW]

I think I've said before that one of my favorite aspects of the R1 build is the use of a prefabricated box. There are lots of clever adjustments that can be made to existing designs, one of my favorite examples being the Nephila, which uses the Denovo Tower and lifts the baffle for a dipole tweeter. If this goes in the direction of Cardioid, are there any existing boxes that would accommodate this?

I was going to say that the Directiva R1 cabinet is the Denovo 0.56 cu. ft. box and it will perfectly match the Denovo 1.16 cu. ft. box you linked to but actually the bookshelf is 1/2" wider unfortunately. So they might not look very appealing when stacked.

I think someone already mentioned this but I like the idea of the 1.16 cu. ft. cabinet made into a bass module for the 2-way bookshelf R2 design to sit on top of. Something like the W371 or Kii BXT would be ideal. I don't know if the dimensions of this cabinet would be conducive to a W371 style design and I imagine a BXT style design would be more expensive and complicated so maybe it doesn't make sense.

In any case I like the design direction I have seen expressed a few times in this thread which seems to be a bass module, possibly with directivity control down to "x" frequency, along with a 2-way bookshelf design similar to R1 but with a less expensive woofer.

 

[TimVG]

Doing a bass module will do nothing for the directivity of the R1 unless one actively modifies the R1 into a 3-way and manipulate it's pattern a la Kii3 - but then it won't be an R1 anymore. As one can see, the R1 goes from +/- 60° to 120° abruptly at ~750Hz

 

[TimW]

Doing a bass module will do nothing for the directivity of the R1 unless one actively modifies the R1 into a 3-way and manipulate it's pattern a la Kii3 - but then it won't be an R1 anymore. As one can see, the R1 goes from +/- 60° to 120° abruptly at ~750Hz

I was just realizing that. So how does the Genelec system work? Is there a directivity mismatch between the W371 and the monitor on top of it?

 

[TimVG]

I was just realizing that. So how does the Genelec system work? Is there a directivity mismatch between the W371 and the monitor on top of it?

I would assume this depends on which monitor is sitting on top, the orientation of said monitor - and on the crossover frequency.

 

[temps]

I was just realizing that. So how does the Genelec system work? Is there a directivity mismatch between the W371 and the monitor on top of it?

Anybody spending that kind of money is assuredly using GLM to measure and integrate the full system. They have extension to 500hz and have an operating mode specifically to maintain constant directivity when used with Ones with that high crossover point. Presumably the monitors are high passed to match. We think of the W371 as a subwoofer but they are much more than that.