ASR Directiva Open Source Speaker Review

[Rick Sykora]

The woofer is superb, but the tweeter simply isn't.
Directivity is not that good, and one of my least favoured tweeters because of the way it sounds, I genuinely hate that thing if I'm honest.

So, choosing "as a product" by the measurements and my personal impression of the tweeter, I still will choose one of Neumann over Directiva.
They sound neutral at least, and there is a clear difference between the loudspeaker which is designed by professionals who can design the directivity properly or not.
If I spend on that woofer in my DIY project, I will choose (subjectively) a far preferable tweeter and design a better waveguide for it.

Appreciate your opinion, but think you may have missed the intent of the project. The r1 design targeted just being good and not an endgame speaker. For that matter, the whole project is built around the invitation to do different (hopefully better) designs. While I am not a professional speaker designer, many other contributors to the r1 design are. For that matter, @ctrl has demonstrated technical skills that I am sure exceed many "professional" designers.

If you have a better design in mind, please create a thread and share it. Thanks!

 

[Rick Sykora]

With a steep HP filter below the PR:s tuning frequence you can limit the excursion somewhat.

Yes, the risk of over-excursion below the tuning is why the design includes a HP filter.

It may be need to be tuned further in a larger room (like Amir's). Am pretty sure the clicking he mentions is over-excursion. At least one of his bass test tracks strained the bass output at loud listening levels. My tuning was meant to provide solid bass extension in an average room with the speakers away from the walls. This tuning is not for every room or speaker placement, but did not compromise bass extension for most music just because a test track pushed the design limits.

 

[Godataloss]

Did you read the review? It clearly states: "My understanding is that a traditional port design is difficult for the Purifi driver and hence the use of the radiator."

There's a lot of vague/generic design criticism in the last couple of pages with no work done to ensure they're realistic or good ideas. "just add a port", "just design your own waveguide for a different tweeter", not very reasonable or simple suggestions.

The Purifi data on Madisound specifies a port. I assumed the difficulty here with the port was fitting the specified port length in the extremely small 'off the shelf' cabinet.

"Box Suggestions:

.20 -.30 cf sealed for an F3 of 87Hz, F6 of 59Hz, F10 of 41Hz
.40 cf with a 2” dia. x 8” long port for an F3 of 44Hz
.50 cf with a 2” dia. x 7” long port for an F3 of 41Hz
.60 cf with a 2” dia. x 7” long port for an F3 of 38Hz
.70 cf with a 2” dia. x 7” long port for an F3 of 35Hz
.80 cf with a 2” dia. x 8” long port for an F3 of 33Hz"

Forgive me for questioning design choices that were made in the spirit of understanding more about this project. If you choose to take them as criticism, so be it, but isn't this how we get to the important discussions in hifi? To me the exciting thing about this project is not the off the shelf box and driver, but instead the implementation of the amplification and DSP crossover. What a revelation for prototyping! As such, the natural next question would be how good can this driver be over a wider FR? The results from such a miniscule box are extraordinarily promising.

 

[hex168]

I just wanted to take a moment to express my appreciation for an extraordinarily well done project. Thank you.

 

[eddantes]

Appreciate your opinion, but think you may have missed the intent of the project. The r1 design targeted just being good and not an endgame speaker. For that matter, the whole project is built around the invitation to do different (hopefully better) designs. While I am not a professional speaker designer, many other contributors to the r1 design are. For that matter, @ctrl has demonstrated technical skills that I am sure exceed many "professional" designers.

If you have a better design in mind, please create a thread and share it. Thanks!

"Professional" or otherwise is not a matter of concern here - we have measurements. This project is a resounding success. Why even acknowledge this slight?

If we can get a passive xover that can do this speaker justice - it would be pretty darn close to an endgame speaker for many of us. (Perhaps @Dennis Murphy would have a go?)

 

[sarumbear]

"Professional" or otherwise is not a matter of concern here - we have measurements. This project is a resounding success. Why even acknowledge this slight?

If we can get a passive xover that can do this speaker justice - it would be pretty darn close to an endgame speaker for many of us. (Perhaps @Dennis Murphy would have a go?)

I don't think it is feasible to have a passive crossover on this set-up (which is excellent BTW). There are various FR corrections applied not to mention level matching the drivers.

 

[abdo123]

I don't think it is feasible to have a passive crossover on this set-up (which is excellent BTW). There are various FR corrections applied not to mention level matching the drivers.

aren't tweeters historically attenuated in all passive crossovers?

there are only like 4 PEQ filters used, I'm pretty sure it won't be super complicated.

 

[sarumbear]

there are only like 4 PEQ filters used, I'm pretty sure it won't be super complicated.

If you know it to be a simple job why don't you have a go at it?

 

[TimVG]

Please let me preface by saying I've never actually built a passive network, and I'm sure I've made mistakes. But after getting to know Vcad a bit I came up with the following results for a passive network. If anyone who has experience with passive networks wants to take a look and correct my work.. by all means have a go.

The vertical lobe can be improved but it would require physically moving the woofer in the Z axis to the front with an extra 3/4" baffle and adjusting the crossover after.

 

[Arc Acoustics]

One of the points of this was to make it from available parts (hence shelf-available waveguide and box).
If it wasn't I believe it'd be a 3 way instead (or a large 2 way like jbl m2, but no purifi woofer then, so 3 way). Maybe with a smaller 4 inch purifi cone.

Appreciate your opinion, but think you may have missed the intent of the project. The r1 design targeted just being good and not an endgame speaker. For that matter, the whole project is built around the invitation to do different (hopefully better) designs. While I am not a professional speaker designer, many other contributors to the r1 design are. For that matter, @ctrl has demonstrated technical skills that I am sure exceed many "professional" designers.

Yes, that "one point" or "intention" was a huge limitation in the design process.
Thus, as I said at first in this conversation, choosing "as a product" by the measurements and my personal impression of the tweeter, I still will choose one of Neumann over Directiva.
If you want to make a really good product, you can have a retailed loudspeaker driver, but the waveguide is not, because it should have a dedicated design for its goal.
Is that a strange opinion?

"Professional" or otherwise is not a matter of concern here - we have measurements. This project is a resounding success. Why even acknowledge this slight?

Yes, you have measurements that showing me a clear disadvantage of rather unprofessional design.
Perhaps my tolerance is a bit more strict or my standard is too high.

 

[headshake]

I'm sure I've made mistakes

The directivity is no mistake! Nice!

 

[beagleman]

Yes, that "one point" or "intention" was a huge limitation in the design process.
Thus, as I said at first in this conversation, choosing "as a product" by the measurements and my personal impression of the tweeter, I still will choose one of Neumann over Directiva.
If you want to make a really good product, you can have a retailed loudspeaker driver, but the waveguide is not, because it should have a dedicated design for its goal.
Is that a strange opinion?


Yes, you have measurements that showing me a clear disadvantage of rather unprofessional design.
Perhaps my tolerance is a bit more strict or my standard is too high.

I am not sure if you intend this, but your comments are coming off a "Tad" harsh.

A lot of thought went into this design, and maybe wait till you get to hear it, or a revised version (assuming that happens) before being too critical.

I realize you seem to have very high standards, but maybe this design is not meant to be the ultimate speaker ever.

 

[sarumbear]

Please let me preface by saying I've never actually built a passive network, and I'm sure I've made mistakes. But after getting to know Vcad a bit I came up with the following results for a passive network. If anyone who has experience with passive networks wants to take a look and correct my work.. by all means have a go.

The vertical lobe can be improved but it would require physically moving the woofer in the Z axis to the front with an extra 3/4" baffle and adjusting the crossover after.
View attachment 157874
View attachment 157875
View attachment 157876
View attachment 157877
View attachment 157878

I would love to see this implemented and the result measured. The baffle adjustment is however is a must. When we designed Silver 5L we went through an optimisation process using LEAP, which took days as the then available CPUs were very slow. (Not to mention that the floating point unit was an optional extra chip!) There was no solution to the delay problem. We had to use a stepped baffle. See it here.

I love this forum and the enthusiasm of its members

 

[sarumbear]

Yes, that "one point" or "intention" was a huge limitation in the design process.

If that is the case why don't you start a new design and we will talk about it?

Yes, you have measurements that showing me a clear disadvantage of rather unprofessional design.

This is not a professional design, nobody is earning money, so please use proper language.

 

[Colonel7]

Such as D2904/71000*, or even T330 or 2905/970000 oldies sound far better for me.
Of course they're pricier, but it's listenable at least.

Thanks for sharing. These are pretty pricey so the BOM would be 2K-$2500, assuming an Eostar could even be purchased through normal channels.

 

[Zvu]

@TimVG I'd like to see the impedance to be able to tell you if it is viable.

 

[TimVG]

@TimVG I'd like to see the impedance to be able to tell you if it is viable.

Here's the vcad file in case you want to mess around.
I did two attemps, this was version #2.

 

[TimVG]

I would love to see this implemented and the result measured. The baffle adjustment is however is a must. When we designed Silver 5L we went through an optimisation process using LEAP, which took days as the then available CPUs were very slow. (Not to mention that the floating point unit was an optional extra chip!) There was no solution to the delay problem. We had to use a stepped baffle. See it here.

I love this forum and the enthusiasm of its members

I very much agree. If the baffle adjustment is not an option I'd make sure that you stay on the acoustical axis, or slightly above it - if that means flipping the speaker upside down, also good. From 10° degrees below axis things tend to get compromised quite severely

 

[sarumbear]

I very much agree. If the baffle adjustment is not an option...

As this is a DIY project I urge the designer to consider to offset the drivers. An inch (or so) thick overlay panel for the low frequency driver is very simple to implement but it will help the FR dramatically.

 

[TimVG]

As this is a DIY project I urge the designer to consider to offset the drivers. An inch (or so) thick overlay panel for the low frequency driver is very simple to implement but it will help the FR dramatically.

3/4" (or 18mm for the Europeans here) panel is enough to fix the biggest issue, 25mm (or an inch) works slightly better even. It only needs a very minor crossover adjustment as well (one inductor, one cap, one resistor)