OP
- Thread Starter
- #21
It is not about me but whether the conversation applies to the thread. For the sake of others, probably should start another thread. It’s free and easy!
-
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!
Directiva r1 speaker build
- Thread starter Rick Sykora
- Start date
It is not about me but whether the conversation applies to the thread. For the sake of others, probably should start another thread. It’s free and easy!
I deleted the post. I was trying to help someone who asked a question in this thread.For the sake of others, probably should start another thread. It’s free and easy!
OP
- Thread Starter
- #23
With the front baffle on and some wood filler on some chipped corners, we are on the homestretch. Here is a pic of the speakon wiring. I used different color wire for the woofer and the tweeter. Tweeter is wired to the #1 terminals and the woofer to the #2 terminals on the speakon…
You may also notice that there are some bare spots on the inside of the front baffle. If you have some Sonic Barrier left, you could apply it there, but should stay away from the woofer as it can interfere with air flow.
Also, if you bought the better speakons, you will need some pan head screws to fasten it to the cabinet.
You may also notice that there are some bare spots on the inside of the front baffle. If you have some Sonic Barrier left, you could apply it there, but should stay away from the woofer as it can interfere with air flow.
Also, if you bought the better speakons, you will need some pan head screws to fasten it to the cabinet.
Last edited:
OP
- Thread Starter
- #24
Wired and mounted the drivers. Before mounting the woofer, should drill out the left and right screw holes as they go into the brace. This will prevent the brace from possibly cracking with the screw pressure. The woofer uses 6 #8 screws and tweeter needs 4 #6 screws.
EDIT: Should emphasize here that the tweeter needs some gasketing tape as the stock driver lacks.
Turned over and added the final pieces of Sonic Barrier. Used a scrap piece on the back of the brace so the tweeter wire has a cushy landing spot...
Finally added about 2 ounces of Acousta-stuf…
EDIT: It may appear that there is more than 2 ounces of damping material. One former team member suggested that the Acousta-Stuf needs to be teased a bit. No major measurement validation on this, but I did follow his advice.
and mounted the passive radiator with (4) #8 screws…
Now off to do impedance measure to see whether this really is a twin to the one Amir has!
EDIT: Should emphasize here that the tweeter needs some gasketing tape as the stock driver lacks.
Turned over and added the final pieces of Sonic Barrier. Used a scrap piece on the back of the brace so the tweeter wire has a cushy landing spot...
Finally added about 2 ounces of Acousta-stuf…
EDIT: It may appear that there is more than 2 ounces of damping material. One former team member suggested that the Acousta-Stuf needs to be teased a bit. No major measurement validation on this, but I did follow his advice.
and mounted the passive radiator with (4) #8 screws…
Now off to do impedance measure to see whether this really is a twin to the one Amir has!
Last edited:
OP
- Thread Starter
- #25
So am going to share a couple of impedance plots with you. One that shows my initial measurement and what I later achieved. As it turns out, I had made a change for cosmetic reasons and it had a negative performance impact. I hope that by sharing my mistake, it will help the less experienced with how to diagnose and resolve comparable build issues.
So here is the initial impedance plot...
I did not expect to see kinks in the curve. This smaller capture actually looks better than my larger view. Anyway, it was not that far off, so what did I do wrong?
If you look back at post #14, I mention that the woofer has a curved flange. This lead me to cut a deeper recess closer in. Appearance-wise, it was all good. My best guess is it ended up allowing slight leaking all around the woofer gasket. Once I cut the outer ridge away, I got a much smoother plot...
With the recess deeper. I now have a sharp edge at the woofer outer perimeter. I have dealt with comparable edges by lightly rounding them with sandpaper. As for matching the first build, it is pretty close and actually a bit better as there is less of a blip at the box tuning frequency. Am not going to overlay. but here is the earlier build...
Note the horizontals scales are significantly different, so the blip get stretched out a bit more. Given he close frequency matches as well as the magnitudes, am pretty confident that they match well.
PS: I realized later one indicator of how different the seal was before and after I leveled it. Before, when I needed to remove it, the woofer would readily come out. After, I had to pound on it to break the seal and remove from the baffle.
So here is the initial impedance plot...
I did not expect to see kinks in the curve. This smaller capture actually looks better than my larger view. Anyway, it was not that far off, so what did I do wrong?
If you look back at post #14, I mention that the woofer has a curved flange. This lead me to cut a deeper recess closer in. Appearance-wise, it was all good. My best guess is it ended up allowing slight leaking all around the woofer gasket. Once I cut the outer ridge away, I got a much smoother plot...
With the recess deeper. I now have a sharp edge at the woofer outer perimeter. I have dealt with comparable edges by lightly rounding them with sandpaper. As for matching the first build, it is pretty close and actually a bit better as there is less of a blip at the box tuning frequency. Am not going to overlay. but here is the earlier build...
Note the horizontals scales are significantly different, so the blip get stretched out a bit more. Given he close frequency matches as well as the magnitudes, am pretty confident that they match well.
PS: I realized later one indicator of how different the seal was before and after I leveled it. Before, when I needed to remove it, the woofer would readily come out. After, I had to pound on it to break the seal and remove from the baffle.
Last edited:
OP
- Thread Starter
- #26
Initial quick FR check looks good! Currently have speaker apart to taper the woofer recess a bit...
OP
- Thread Starter
- #27
So, while I have been trying to round off the sharp edge of the woofer recess, @amirm has been busy doing some measurements on the r1 prototype. Ofc, his efforts were much more complex and were complicated by a Klippel software update. @ctrl has been working with Amir‘s output and massaging it into VituixCAD. At one point, after having to rework the measurement templates, Amir sent me his measurement of the bass performance for me to review. It looked like this…
It was late Saturday night and my initial reaction was ok, but Amir cornered me and, after digging up my measurement, this one looks pretty ragged. Bear in mind, there is no crossover or eq involved here. So, I went to sleep and awoke in the middle of the night to find Amir had rerun the test and gotten the same results. So, I sent him my measurements (green line) and went back to sleep…
At this point, Amir is now asleep but I awake and am pondering why his results seem so different. Our vertical scales are close and then I notice he had measured much lower frequency than I had requested. This is the part I call eyes, damn eyes and measurements. After trying to eyebalI the comparison with this new revelation, still thought the graphs were too different. At this point , I could have used VCAD’s trace tool to more precisely compare the graphs, but just did a quick and dirty rescaling of his results in a graphics editor. Now, I was able to see that his plot was almost the same as mine (given the resolution of the Klippel is much better than my gated measurement).
Anyway, all of us are multitasking other projects, but am now back in the critical path. Once I get the speaker back together, will remeasure it to see if my filing and sanding made any difference and then will load the new configuration into my minidsp. @ctrl has already designed 3 crossover variants of various slopes and crossover frequencies. Will do a bit of listening and some tweaking and then plan to share the initial VCAD project and measurements this weekend. Stay tuned!
It was late Saturday night and my initial reaction was ok, but Amir cornered me and, after digging up my measurement, this one looks pretty ragged. Bear in mind, there is no crossover or eq involved here. So, I went to sleep and awoke in the middle of the night to find Amir had rerun the test and gotten the same results. So, I sent him my measurements (green line) and went back to sleep…
At this point, Amir is now asleep but I awake and am pondering why his results seem so different. Our vertical scales are close and then I notice he had measured much lower frequency than I had requested. This is the part I call eyes, damn eyes and measurements.
After trying to eyebalI the comparison with this new revelation, still thought the graphs were too different. At this point , I could have used VCAD’s trace tool to more precisely compare the graphs, but just did a quick and dirty rescaling of his results in a graphics editor. Now, I was able to see that his plot was almost the same as mine (given the resolution of the Klippel is much better than my gated measurement).Anyway, all of us are multitasking other projects, but am now back in the critical path. Once I get the speaker back together, will remeasure it to see if my filing and sanding made any difference and then will load the new configuration into my minidsp. @ctrl has already designed 3 crossover variants of various slopes and crossover frequencies. Will do a bit of listening and some tweaking and then plan to share the initial VCAD project and measurements this weekend. Stay tuned!
Last edited:
OP
- Thread Starter
- #30
Lol, no but would need @amirm to explain...
You are correct about Purezza comparison. Without a crossover, Directiva r1 bass subsystem should be almost identical performance to Purezza. Expect Directiva to be better or equal depending on crossover implementation. Once I get the speaker back together, will get a better read on a more accurate sensitivity value.
OP
- Thread Starter
- #31
Depending on crossover implementation, sensitivity is in the 85-90 dB range.
If you approach the minidsp crossover design as you would a passive crossover, you get the lower sensitivity.
If you approach the minidsp crossover design as you would a passive crossover, you get the lower sensitivity.
OP
- Thread Starter
- #32
Here is sneak peek at our progress after Amir's measurements...
This is one of the 3 initial crossovers created by @ctrl (red is on-axis at 1m and green is reverse null, output level is set to protect my hearing). After debugging my test rig (it was causing rippling in the frequency response), as @ctrl reported, we are getting good matching to sims he had done. Still have a bunch of listening to do and potential voicing that my occur as a result.
This is one of the 3 initial crossovers created by @ctrl (red is on-axis at 1m and green is reverse null, output level is set to protect my hearing
). After debugging my test rig (it was causing rippling in the frequency response), as @ctrl reported, we are getting good matching to sims he had done. Still have a bunch of listening to do and potential voicing that my occur as a result. 
TimW
Major Contributor
Looks a little bright.Here is sneak peek at our progress after Amir's measurements...
View attachment 135063
This is one of the 3 initial crossovers created by @ctrl (red is on-axis at 1m and green is reverse null, output level is set to protect my hearing
OP
- Thread Starter
- #34
Yes, but not done with voicing and, as you probably know, bass is a bit tricky with the Purifi woofer. I have done a crossover that takes advantage of more gain that is available in an active design. It is easy to push that too far and drive the woofer to excursion limits…
Another consideration is how off-axis will rolloff as well. We will have to decide on some slope before Amir tests, but my goal was to produce a stereo speaker that could be adapted to the room and to one’s taste. In the minidsp HD, you have room for 4 different crossover confiigurations, so the hobbyist can experiment beyond our base crossover. There is also still room for more PEQs if one wanted to use with REW to do adapt the speaker to their room.
Not really looking to market a finished product, but that same flexibility could be used to provide different timbres or bass rolloffs to support different placements.
Last edited:
Well the flatness of the on and off axis is only aspect and bright might only if you listen in the nearfield with speaker pointing directly at you like a laser beam. I'm sure Rick and ctrl follow CTA2034 and would be designing for a balance between on-axis and power response or directivity index or whatever you want to call it.
Advanced designers also don't aim for a flat anechoic response. eg. Revel always have a drooping response below 80Hz, even in their largest speakers to better match the room. Typical in-room response will go all the way down to the bottom octave/low bass or whatever you want to call it
(20-40Hz)
@Rick Sykora
Do you have a turntable to take 10, 20, 30, 40..90 degrees measurements?
At the time I didn't so I only took 40 and 80 degrees, but I should probably redo them at at the very least 15/30/45/60 and 90
And then of course depending on your room, you may need to adjust treble levels to your liking. And bass tilt/rolloff depending on how far you place the speaker out into the room eg. 2 ft vs close to wall. This is one of the advantages of an active design IMHO like the miniDSP with it's 4 presets that you select with a remote from your listening positioning.
Advanced designers also don't aim for a flat anechoic response. eg. Revel always have a drooping response below 80Hz, even in their largest speakers to better match the room. Typical in-room response will go all the way down to the bottom octave/low bass or whatever you want to call it
(20-40Hz)
@Rick Sykora
Do you have a turntable to take 10, 20, 30, 40..90 degrees measurements?
At the time I didn't so I only took 40 and 80 degrees, but I should probably redo them at at the very least 15/30/45/60 and 90
And then of course depending on your room, you may need to adjust treble levels to your liking. And bass tilt/rolloff depending on how far you place the speaker out into the room eg. 2 ft vs close to wall. This is one of the advantages of an active design IMHO like the miniDSP with it's 4 presets that you select with a remote from your listening positioning.
OP
- Thread Starter
- #36
I bought a turntable, but needs some work to turn into something more useful. My challenge is no dedicated space and so it has to be portable. This is why I just ran into some test stand resonance issues and we are still not sure all the demons are exorcised. Did not make sure all was tight after I rebuilt recently and @ctrl pointed out the problem. Now that my measurements are more scrutinized, plan to start setting a baseline to calibrate to known reference. Still not sure about sensitivity to standard levels as is uncomfortably loud even in my fairly large family room.
Last edited:
OP
- Thread Starter
- #37
I had mentioned previously that @ctrl pointed out a potential resonance issue in my test rig but thought it might be useful to share how I might have realized sooner. Whereas, I had not compared my new measures to prior ones, he was comparing them to the sims. So the sims served as baseline for knowing what was a good measurement, but they are just sims. Fortunately, he has the experience to know how well they should be aligning to real measures.
I mentioned earlier about the importance of establishing a know good baseline for reference. For those who are less experienced, thought it would be helpful to show how the measurements compared when the test rig was corrupting them vs. after I fixed...
Red trace is the earlier measure and the blue one is once I fixed the test rig. There may still be some issues to address over the 10k Hz range, but you can see the obvious waviness in the earlier measure.
I mentioned earlier about the importance of establishing a know good baseline for reference. For those who are less experienced, thought it would be helpful to show how the measurements compared when the test rig was corrupting them vs. after I fixed...
Red trace is the earlier measure and the blue one is once I fixed the test rig. There may still be some issues to address over the 10k Hz range, but you can see the obvious waviness in the earlier measure.
What was the resonance issue? Currently in the testing phase of my own turntable and am curious where to look for problems.
