• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Introducing Directiva - An ASR open source platform speaker project

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,087
Location
.de, DE, DEU
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.
Can you please post your gated on-axis measurements here so we can compare how well the VCAD simulation matches the real measurements.

There is still a small uncertainty in Amir's NFS measurements regarding the phase alignment of tweeter and woofer.
The distance of the "virtual measuring mic" could possibly have been slightly different in the measurements of tweeter and woofer.
This could lead to a phase shift between simulation and real measurements and thus slightly change the overall frequency response.
 

kimmosto

Active Member
Joined
Sep 12, 2019
Messages
213
Likes
498
There is still a small uncertainty in Amir's NFS measurements regarding the phase alignment of tweeter and woofer.

That's right. Phase responses indicate that mic has been ~60 mm above tweeter's axis while measuring tweeter, and ~120 mm above woofer's axis while measuring woofer. Mic must be on driver's axis while measurements in order to simulate anything related to vertical directivity if data includes also vertical plane. Another option is to keep mic at the same spot for both drivers, but that requires actual measurement distance of 2 meters. In that case X,Y,Z mm of both drivers should be 0,0,0 mm in crossover simulation because vertical timing is included in phase responses of measurement data i.e. program should not calculate anything related to geometry.

Another confusing thing is that hor and ver planes are swapped with amir's data if 'Klippel NFS' is checked in Options while loading responses to Drivers tab. Planes in amir's data are logical and standard IMO, but opposite compared to test data exported from Klippel's data browser. So I will rename 'Klippel NFS' to 'Swap planes' in Options window of VCAD, and user is responsible to check which option is correct.

So my simulation has also some small errors due to swapped planes and timing errors.
 

kimmosto

Active Member
Joined
Sep 12, 2019
Messages
213
Likes
498
Phase responses indicate that mic has been ~60 mm above tweeter's axis while measuring tweeter, and ~120 mm above woofer's axis while measuring woofer.

~60 and ~120 mm with simulation to 2 m so Y offset has probably been much smaller if mic was actually at 1 m and Klippel does not reflect all details to 3 m.
 

kimmosto

Active Member
Joined
Sep 12, 2019
Messages
213
Likes
498
..hor and ver planes are swapped with amir's data if 'Klippel NFS' is checked in Options...

This was wrong conclusion. Checked 'Klippel NFS' produces more symmetrical polarmap in horizontal plane, but something else in timing confuses..
 

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,087
Location
.de, DE, DEU
Phase responses indicate that mic has been ~60 mm above tweeter's axis while measuring tweeter, and ~120 mm above woofer's axis while measuring woofer.
~60 and ~120 mm with simulation to 2 m so Y offset has probably been much smaller if mic was actually at 1 m and Klippel does not reflect all details to 3 m.

I would be a bit surprised if Amir was so wrong with the "reference point" setting of the Klippel NFS.

1629367262852.png


To get a "feel" for the possible Y-axis deviation, I used the simulations for the Directiva speaker.

Since the woofer is not centered in the vertical direction and as a reference point, the first simulation with the "virtual mic" at 2m distance centered on the woofer. As expected, the vertical frequency responses +-90° normalized to the axial frequency response are almost congruent in pairs.
1629367942059.png 1629368129977.png

Now for comparison Amir's NFS measurements of the woofer in vertical direction, also +-90° and normalized:
1629368309439.png
Up to 4kHz, the positive and negative angles of the vertical frequency responses match well. This means that Amir has actually hit the center of the woofer well in his measurements.

A bit strange are the jagged curves and deviation in the range 700-2000Hz.

If the "virtual mic" had been offset 120mm on the Y-axis at a distance of 2m, one would expect the following vertical frequency responses:
Y-axis, 120mm off
1629367830729.png 1629368981409.png
This would result in a significant shift in the vertical frequency responses. In Amir's measurements, however, the range 1.5-4kHz is almost congruent.
Therefore, the deviation of the woofer by 120mm at 2m distance on the Y-axis cannot be correct. However, this does not exclude small deviations.


The 60mm deviation (at 2m distance or 30mm@1m) on the Y axis is almost a bit much. If you look at the vertical frequency response measurements of the tweeter +-90°, the FR of the negative and positive angles are very similar in the 8-20kHz range. With 60mm@2m offset, wouldn't you expect a bigger difference there?
1629372573472.png 1629372592812.png

As example here is the same tweeter with 55mm@1m vertical offset. There the divination is clearly visible.
1629373522972.png 1629373541919.png


My guess would be that there might have been a slight shift in the Z-direction of the "reference point" of tweeter and woofer.
Or everything is okay and resolves when Rick has his new measuring equipment running.

The measurement of @hdussen2 should provide a first hint.
 

ctrl

Major Contributor
Forum Donor
Joined
Jan 24, 2020
Messages
1,616
Likes
6,087
Location
.de, DE, DEU
...but something else in timing confuses..
Tried to get a measure of how far apart the sound generation points of the drivers are in the direction of the z-axis.

If we equalize the frequency responses of woofer and tweeter to congruent FR and assume that both drivers behave like a minimum-phase system in the considered frequency range, then the phase frequency responses of tweeter and woofer should also be congruent if both driver (sound generation points) are on the same plane.

They are not, of course.
If I move the woofer 9-10mm further forward (in the direction of the listener) on the Z-axis in VCAD, then the phase frequency response fits.
1629375141732.png 1629375172445.png

That seems a bit small to me (would have expected more), but the Purifi woofer is very flat and the waveguide of the SeasDXT shifts the sound generation point to the rear.
 

kimmosto

Active Member
Joined
Sep 12, 2019
Messages
213
Likes
498
Obviously something was wrong with my investigations resulting high Y offsets. Probably sector in vertical plane was too narrow and responses up / down did not match good enough. Now everything looks okay. That's good.
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
3,518
Likes
7,028
Location
Stow, Ohio USA
If you have been following this thread, you know I have struggled with my test rig and environment. It was not bad for my own purposes, but needed to improve to do Directiva and other projects justice. I have learned a lot and will share some on that shortly on the build thread.

Am please to announce that I finally got dual channel measurements working. I really wanted to do this with my QA401 audio interface, but have spent too much time on trying to get past some of the bleeding edge shortcomings. I considered using the minidsp HD, but not sure it is much more mature. So, fell back to my Audigy FX soundcard. Am still using the QA471 mic preamp. I may try the minidsp just to see if it better than my soundcard…

At this stage, I have done some dual channel measures and can use the correlate measurements with VituixCAD filter sims using the 96k setting for it. The shelving filter is not too far off, but not as good as the other filters. So we will heed Kimmo's advice and use the 2nd order if needed. :cool:
 
Last edited:

mdsimon2

Major Contributor
Forum Donor
Joined
Oct 20, 2020
Messages
2,478
Likes
3,315
Location
Detroit, MI
@hardisj just posted NFS measurements of a JBL M2 which is a 2 channel active design using an external amplifier / DSP which means that his NFS setup is cable of testing an active design like the Directiva. If Erin is willing to test and Rick is willing to ship I'd gladly cover shipping costs.

Michael
 

PurdueDeep

Member
Forum Donor
Joined
Aug 17, 2021
Messages
42
Likes
11
If you have been following this thread, you know I have struggled with my test rig and environment. It was not bad for my own purposes, but needed to improve to do Directiva and other projects justice. I have learned a lot and will share some on that shortly on the build thread.

Am please to announce that I finally got dual channel measurements working. I really wanted to do this with my QA401 audio interface, but have spent too much time on trying to get past some of the bleeding edge shortcomings. I considered using the minidsp HD, but not sure it is much more mature. So, fell back to my Audigy FX soundcard. Am still using the QA471 mic preamp. I may try the minidsp just to see if it better than my soundcard…

At this stage, I have done some dual channel measures and can use the correlate measurements with VituixCAD filter sims using the 96k setting for it. The shelving filter is not too far off, but not as good as the other filters. So we will heed Kimmo's advice and use the 2nd order if needed. :cool:
Hey Rick. Just looking into making a purifi build. Is there a spec somewhere or if it's in the process is there a prospective timeline? Thanks for your effrots!
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
3,518
Likes
7,028
Location
Stow, Ohio USA
Minidsp HD does not appear to support using the analog inputs and the USB input together. Staying with the SB Audigy for now.

Am currently working with @ctrl on his latest crossover. I need to simulate with my measurements in VCAD to cross check, but once we have will use as a basis for the passive design AND will post the active one to the build thread.
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
3,518
Likes
7,028
Location
Stow, Ohio USA
Hey Rick. Just looking into making a purifi build. Is there a spec somewhere or if it's in the process is there a prospective timeline? Thanks for your effrots!

My pleasure! Only have the build thread for now. Should be enough to catch up to us though. If you need something more, just let me know. :cool:
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,370
Likes
234,430
Location
Seattle Area
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.
I have another company sending me an active speaker with outboard electronics. So at this rate if I can bunch the two tests together, I can first pull out all the current cables in there, fish new ones for this test and test. Then put everything back the way it was. Alternatively I can dangle the wires from the ceiling but that requires building a boom as I do not want to put holes on the ceiling to hold the wires. Either way, we will get it done.
 

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
44,370
Likes
234,430
Location
Seattle Area
@hardisj just posted NFS measurements of a JBL M2 which is a 2 channel active design using an external amplifier / DSP which means that his NFS setup is cable of testing an active design like the Directiva. If Erin is willing to test and Rick is willing to ship I'd gladly cover shipping costs.

Michael
Looks like he just dangled the wires down the fixture in which case the rotation of the robotic arm keeps pushing them around. That may work but I rather have it not contact the fixture.
 

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
Looks like he just dangled the wires down the fixture in which case the rotation of the robotic arm keeps pushing them around. That may work but I rather have it not contact the fixture.

No. I did not.
 
OP
Rick Sykora

Rick Sykora

Major Contributor
Forum Donor
Joined
Jan 14, 2020
Messages
3,518
Likes
7,028
Location
Stow, Ohio USA
I have another company sending me an active speaker with outboard electronics. So at this rate if I can bunch the two tests together, I can first pull out all the current cables in there, fish new ones for this test and test. Then put everything back the way it was. Alternatively I can dangle the wires from the ceiling but that requires building a boom as I do not want to put holes on the ceiling to hold the wires. Either way, we will get it done.

Thanks Amir!

Directiva fans should know that this may take a while. Aside from having to modify the Klippel…

EDIT of EDIT: Amir will need to validate the crossover against his initial measurements. This originally will be done by me as the attempt to recall the speaker shipment was not successful. :confused:
 
Last edited:

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
Guys, the Klippel doesn't have to be modified to use 2 pair of speaker wires. The NFS comes with a long 4-conductor wire which has a 4-pole neutrik speakON connector on one end and connector clips on the other end. IOW, it is already wired for this.

All Amir has to do is remove the large connecting clips from the wires on the NFS and … ta-da.... 4 bare wires (red, white, blue, black). The speakON connector end has the wires clearly labeled as to what color is what channel (i.e., Channel 1+, Channel 1-, Channel 2+, Channel 2-).

Assuming you have a modern two channel pro-amp, the other end (at the amplifier) simply goes to Channel 1 of the amplifier because if it is a 2-channel amp it will have speakON connector that has the ability to use a 4-pole connector (channel 1 and channel 2 are powered individually from their respective amplifier channels but via the same speakON connector). However, if the amp doesn't have a speakON connector, just get an adapter and run the 4 bare wires to the amplifier output channels. Done. Here's the adapter you need ($4):
https://www.markertek.com/product/n...MI1cvasLrM8gIVja_ICh1lswgBEAQYBCABEgIxAvD_BwE


Super easy. And no modifications needed. This is how I am able to test active 2-way speakers like the M2, previously mentioned.
 
Last edited:

sigbergaudio

Major Contributor
Audio Company
Forum Donor
Joined
Aug 21, 2020
Messages
2,639
Likes
5,397
Location
Norway
Sica 6.5" was measured at Dibirama. Much more usable FR than most pro coax drivers, for sure. Tweeter THD appears to be the major weak area.

You just have to cross over high enough. Those measurements look a bit weird too, compared to what we've got. I've measured the 5.5" coax at 109,5dB@1m outdoors (both midbass and tweeter active), that got me 1,48% THD @5khz. So my impression is that you can freely play as loud as realistically possible indoors with this tweeter as long as you cross over at 2500-3000.

We've also played 111dB continously at the listening position indoors with no hint of audible compression.
 

hardisj

Major Contributor
Reviewer
Joined
Jul 18, 2019
Messages
2,907
Likes
13,908
Location
North Alabama
I made a quick video to illustrate what I mean above. This should get you squared away.

Edit: Just make sure to put the spring-loaded connectors back on when you’re done. I believe (I may be wrong) that the additional channels are part of the feedback for the closed loop when measuring a passive speaker. But in the case of active 2-way - where output signal is routed to an external amp(s), you don’t need that.


Wiring up the Klippel NFS for active 2-way measurements
 
Last edited:
Top Bottom