• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

Introducing Directiva - An ASR open source platform speaker project

OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
1,674
Likes
2,736
Location
Stow, Ohio USA
Here are the Directiva driver impedance curves from Amir and massaged into VCAD by @ctrl. They are done at a level of 2V, but are very comparable to mine that were measured much lower...
 

Attachments

  • Directiva Woofer and Tweeter Impedance.zip
    342.5 KB · Views: 43
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
1,674
Likes
2,736
Location
Stow, Ohio USA
To be specific, we were working on correlations and the project uses shelving filters in the crossover. When the project was set to minidsp 96 kHz, the simulated filter shape did not match the one in the minidsp HD. One I set the project dsp to minidsp 48 kHz, the shelving filters matched.

One key aspect is that the VituixCAD filter had a fixed Q and the minidsp allows Q to be manipulated. To match, the filter Q for the minidsp had to be set to 0.5. Not sure I saved all the tests, but recall I tested a couple of other filter types and did not have any issues with any other minidsp setting. As was busy sorting out other measurements issues, was happy to find a solution and move on.

Hope this helps clarify. I would expect the same filter to have the same shape regardless of which minidsp is used. I did not scrutinize further but you can see this issue in the posted project. If you only change the dsp choice, you can see the sim’s frequency response change significantly.

I ran this by Kimmo Saunisto and he said the dsp diffs are common and due to frequency warping. He was not sure whether the minidsp HD was 96kHz, but acknowledged the shelving filter issue and suggested using second order ones.

He also was willing to produce a crossover for Directiva r1. Actually he has already revised it and have not had an opportunity to compare to the original one. He designed for minidsp 48 kHz. Here it is…

1629078655167.png

and here is how his sim looks…
1629078737107.png
 
Last edited:

HammerSandwich

Addicted to Fun and Learning
Forum Donor
Joined
Nov 22, 2018
Messages
966
Likes
1,180
Wow, that's very impressive. Not just the FR/DI curves but creating them with just a few filters, all low-Q... Hat tip, @kimmosto!
 

headshake

Active Member
Joined
Apr 20, 2019
Messages
293
Likes
229
I ran this by Kimmo Saunisto and he said the dsp diffs are common and due to frequency warping. He was not sure whether the misdsp HD was 96kHz, but acknowledged the shelving filter issue and suggested using second order ones.

He also was willing to produce a crossover for Directiva r1. Actually he has already revised it and have not had an opportunity to compare to the original one. He designed for minidsp 48 kHz. Here it is…

View attachment 147719
and here is how his sim looks…
View attachment 147720
Cool!

Can you generate the vert. directivity of Kimmo's XO.

It is so neat to see what is prioritized. Kimmo's XO (to my newbie eyes) looks to put phase as a low priority compared to the other graphs. As a novice, I thought the phase needed to track better.
 

Draki

Member
Joined
Jan 22, 2020
Messages
41
Likes
51
Location
Macedonia
It is so neat to see what is prioritized. Kimmo's XO (to my newbie eyes) looks to put phase as a low priority compared to the other graphs. As a novice, I thought the phase needed to track better.[/QUOTE]

The phase is excellent. Note the 125 us delay on the tweeter, applied for this purpose.
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
1,674
Likes
2,736
Location
Stow, Ohio USA
Cool!

Can you generate the vert. directivity of Kimmo's XO.

It is so neat to see what is prioritized. Kimmo's XO (to my newbie eyes) looks to put phase as a low priority compared to the other graphs. As a novice, I thought the phase needed to track better.

Will do, but am checking with Kimmo as have some remaining phase issues in my copy that do not match what he sent to me.
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
1,674
Likes
2,736
Location
Stow, Ohio USA
Here is Kimmo's vertical polar...

1629115991569.png
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
1,674
Likes
2,736
Location
Stow, Ohio USA
Thought it was time to circle back here. For those who may have been watching here and missed, we did publish the design on the main thread. The implementation with minidsp ran into wall as the Klippel cannot handle more than one signal wire. So we regrouped and @ctrl is working on a passive crossover so Amir can review it.

Will copy a link here for some of the content on the other thread, but will repost the base measurement files and any r1 future posts will be done here.
 

NTK

Major Contributor
Forum Donor
Joined
Aug 11, 2019
Messages
1,283
Likes
2,531
Location
US East
The implementation with minidsp ran into wall as the Klippel cannot handle more than one signal wire.
I wonder if we (... I mean Amir) can rig up our (his) own poor man's "multi-source superposition" module.

Amir has the complex data export module. He can measure the speaker one driver at a time, and export the impulse responses at the 70 spinorama points. Then sum the impulse responses for each of the drivers. These impulse responses can then be converted back to frequency responses, and become the FR curves for the 70 spinorama points of the combined total output from the speaker.

This is possibly not the best or simplest way to measure this particular speaker. Just an idea may be for future use.
 

TimW

Addicted to Fun and Learning
Joined
Aug 15, 2018
Messages
669
Likes
851
Location
Martha Lake, WA
Thought it was time to circle back here. For those who may have been watching here and missed, we did publish the design on the main thread. The implementation with minidsp ran into wall as the Klippel cannot handle more than one signal wire. So we regrouped and @ctrl is working on a passive crossover so Amir can review it.

Will copy a link here for some of the content on the other thread, but will repost the base measurement files and any r1 future posts will be done here.
This makes no sense, how does Amir measure other active speakers? There should be no issue measuring an active speaker that uses a minidsp as crossover if it is setup properly. It seems like we are giving up on the superior performance of an active design because someone can't figure out how to setup a minidsp?
 

TimW

Addicted to Fun and Learning
Joined
Aug 15, 2018
Messages
669
Likes
851
Location
Martha Lake, WA
Looking at the other Directiva thread, it seems the stumbling point is that you did not want the electronics sitting on top of the speaker? I also can't tell if Amir has a miniDSP available to him or not.

The electronics being a stereo amp and a minidsp could be placed away from the speaker with two sets of speaker wire connecting the amp to the drivers. The minidsp should only need one signal. I have plenty of speaker wire including some 4 conductor stuff that could be used if there is only room for one cable. I also have a few minidsp units. And I live nearby Amir and have met up with him before. If you guys need help getting this setup working I'm available.
 

McFly

Addicted to Fun and Learning
Joined
Mar 12, 2019
Messages
631
Likes
996
Location
NZ
Rick, I was the OP of that diyaudio thread. While I still battle to find the ultimate port design for that woofer (I’m close, but it requires a floorstanding cabinet), I question what you mean by the varying woofer flange and why you had different recess depths, and subsequent impedance wiggles? My woofer flanges here are nice and flat on the magnet side, and the open air side the flange varies from 3.5mm at the edge to 5mm by the surround. This makes for a nice finish with a 3mm recess only and a layer of your choice of vinyl/veneer.

Also, apologies if it was said in the other thread, can someone not just send Amir a DSP plate amp with the XO loaded on it to test the active version? I suppose Amir is too busy if the xo needed to be changed..
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
1,674
Likes
2,736
Location
Stow, Ohio USA
Looking at the other Directiva thread, it seems the stumbling point is that you did not want the electronics sitting on top of the speaker? I also can't tell if Amir has a miniDSP available to him or not.

The electronics being a stereo amp and a minidsp could be placed away from the speaker with two sets of speaker wire connecting the amp to the drivers. The minidsp should only need one signal. I have plenty of speaker wire including some 4 conductor stuff that could be used if there is only room for one cable. I also have a few minidsp units. And I live nearby Amir and have met up with him before. If you guys need help getting this setup working I'm available.

The minidsp I have works just fine and would be easy to explain to @amirm how to make it work. If there was a way to make the setup working without compromising the measurements would be happy to avoid. Shipping these boxes is not inexpensive, so was in no hurry to retrofit one with a passive crossover. If he lacked the minisdsp, am prepared to ship one to him prewired.

His stated issue is that there is no room in the Klippel wiring column to run another speaker wire so the electronics could be located remotely. Naturally would be pleased if this could be overcome, but took Amir at his word.
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
1,674
Likes
2,736
Location
Stow, Ohio USA
Rick, I was the OP of that diyaudio thread. While I still battle to find the ultimate port design for that woofer (I’m close, but it requires a floorstanding cabinet), I question what you mean by the varying woofer flange and why you had different recess depths, and subsequent impedance wiggles? My woofer flanges here are nice and flat on the magnet side, and the open air side the flange varies from 3.5mm at the edge to 5mm by the surround. This makes for a nice finish with a 3mm recess only and a layer of your choice of vinyl/veneer.

Also, apologies if it was said in the other thread, can someone not just send Amir a DSP plate amp with the XO loaded on it to test the active version? I suppose Amir is too busy if the xo needed to be changed..

A plate amp would still have to be mounted externally, so might might help, but only partly addresses. We considered it too. As I replied to @TimW, would have readily supplied a preconfigured minidsp to Amir if that resolved the issue.

Not sure what you mean about flanges? I did have some test environment issues, but they are relatively minor. I am replacing my UMIK1 with a EMM6 so I can do a dual channel measurements. Kimmo has stated that a USB mic cannot be used with REW to integrate the driver measurements (although mine are not too bad). Comparing woofer measurements between my older measurement setup and the improved one, made us wonder if there might be some difference in the his woofer vs mine. To resolve any potential issues before he reviewed, we wanted to have him compare the crossover sim results to his real measurements. This is when he revealed that he could not do the measurement due to the lack of space for the wiring.
 

mcdn

Active Member
Forum Donor
Joined
Mar 7, 2020
Messages
113
Likes
219
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
1,674
Likes
2,736
Location
Stow, Ohio USA

hdussen2

New Member
Joined
Aug 14, 2021
Messages
1
Likes
6
Hello,

I have build this speaker and loaded both crossovers in my Minidsp.
Although I was a bit surprised by its high tweeter crossover, my first impression is that the Kimmo Saunisto his crossover gives a better balance. Especially around 2Khz there is a huge difference.

The Purifi PTT6.5 is a new speaker for me and I must say that it performces very well and can play very loud in combination with my Hypex amplifiers.
 
Top Bottom