Okay one more question!
If using one Neo8 and a tweeter with a 6KHz Xover, what is the best arrangement for the drivers.. to have them placed side by side or one on top of the other?
Most people, when using small numbers of the Neo8 seem to place the tweeter on top or between the two drivers vertically.
It seems to make more sense to me though, to place them as close to each other as possible horizontally since that is where the wider dispersion is and thus the sound will integrate better. Or is that not so?
I don't suppose anyone has polar response for vertical off-axis of the Neo8s do they?
If using one Neo8 and a tweeter with a 6KHz Xover, what is the best arrangement for the drivers.. to have them placed side by side or one on top of the other?
Most people, when using small numbers of the Neo8 seem to place the tweeter on top or between the two drivers vertically.
It seems to make more sense to me though, to place them as close to each other as possible horizontally since that is where the wider dispersion is and thus the sound will integrate better. Or is that not so?
I don't suppose anyone has polar response for vertical off-axis of the Neo8s do they?
This is a measurement I made some time ago, maybe it can be useful to you:
http://www.rlorello.net/rldelta45/Neo8_boxh_mlsw_1m_060430.png
The Neo8 was mounted in a 1.7 liters closed box placed horizontally. The 3 measurements were made at 0°, 30° and 60° off axis. The mic was placed at 1 meter from the speaker. The measurement is not valid below 300 Hz since it was time-windowed.
As you can see mounted this way it has not a wide dispersion, similar to a 90 mm cone speaker.
BG notch filter
Hey Guys,
I was just playing around with a pair of NEO8 drivers tonight and discovered a few things about the published notch filter in the BG datasheet.
I mounted the NEO8 in a 9"x 12" baffle and took some raw measurements. After that I connected the suggested notch filter and took another measurement. From what I can see the notch filter is nowhere near the frequency it needs to be at. It had a mild effect between about 5-10kHz, but did nothing for the 12K bump.
I started from scratch and designed a parallel notch for 12kHz with a passband of 5kHz. The values I came up with are as follows:
R = 15 ohm
C = 2.4 uF (two 4.7uF in series)
L = 0.06mH (used 0.56)
I took a third measurement with the new filter and it was spot on. The 12kHz bump is gone completely, and the rising output starting at 2kHz was also tamed somewhat. I'll attach the three measurements for reference. All measurements were gated and were made in SW with and ECM8000 microphone and M-audio sound card. No smoothing was applied.
Has anyone else noticed this about the published notch? To be honest, I have no idea how they came up with the values listed in the whitepaper. If anyone has used the BG notch successfully and taken measurements, I'd love to see them to compare. The ribbon sounds about 10 times better with the proper notch on it, the aggressive top end is much more relaxed now, and the ssss hissing on females vocals is gone too. There might be hope for these things yet!
All the best,
Owen
Hey Guys,
I was just playing around with a pair of NEO8 drivers tonight and discovered a few things about the published notch filter in the BG datasheet.
I mounted the NEO8 in a 9"x 12" baffle and took some raw measurements. After that I connected the suggested notch filter and took another measurement. From what I can see the notch filter is nowhere near the frequency it needs to be at. It had a mild effect between about 5-10kHz, but did nothing for the 12K bump.
I started from scratch and designed a parallel notch for 12kHz with a passband of 5kHz. The values I came up with are as follows:
R = 15 ohm
C = 2.4 uF (two 4.7uF in series)
L = 0.06mH (used 0.56)
I took a third measurement with the new filter and it was spot on. The 12kHz bump is gone completely, and the rising output starting at 2kHz was also tamed somewhat. I'll attach the three measurements for reference. All measurements were gated and were made in SW with and ECM8000 microphone and M-audio sound card. No smoothing was applied.
Has anyone else noticed this about the published notch? To be honest, I have no idea how they came up with the values listed in the whitepaper. If anyone has used the BG notch successfully and taken measurements, I'd love to see them to compare. The ribbon sounds about 10 times better with the proper notch on it, the aggressive top end is much more relaxed now, and the ssss hissing on females vocals is gone too. There might be hope for these things yet!
All the best,
Owen
Attachments
opc..
I don't think BG was concerned with the top octave response on the driver.
Consider that the driver beams quite a bit in that passband, the solution then is to listen somewhat off of the driver's primary axis (..horizontally).
I think the reason for their filter has more to do with its linesource character at more traditional listening distances - say 3 meters where the that upper freq. response (above 4 kHz) will still operate as a linesource (and is less directional), but the freq. response below this will not. That elevates the response considerably at the listening position.
In fact I believe its this character that seemingly makes the driver particularly difficult to deal with unless its used in multiples to make a long line array. Its also a significant factor in the differing response measurements posted by others.
I think for this driver to work you'll have to do:
1. Use in multiples for a long line array, or
2. Use it as a mid driver between say 600 - 3 kHz with steep filters (..I believe Martin Logan and BG do this), or
3. "Indulge" in the art of loudspeaker design by measuring at the listening position while adjusting the horizontal axis to achieve a better "balance", or
4. Horn load the thing and adjust for horizontal axis.
I don't think BG was concerned with the top octave response on the driver.
Consider that the driver beams quite a bit in that passband, the solution then is to listen somewhat off of the driver's primary axis (..horizontally).
I think the reason for their filter has more to do with its linesource character at more traditional listening distances - say 3 meters where the that upper freq. response (above 4 kHz) will still operate as a linesource (and is less directional), but the freq. response below this will not. That elevates the response considerably at the listening position.
In fact I believe its this character that seemingly makes the driver particularly difficult to deal with unless its used in multiples to make a long line array. Its also a significant factor in the differing response measurements posted by others.
I think for this driver to work you'll have to do:
1. Use in multiples for a long line array, or
2. Use it as a mid driver between say 600 - 3 kHz with steep filters (..I believe Martin Logan and BG do this), or
3. "Indulge" in the art of loudspeaker design by measuring at the listening position while adjusting the horizontal axis to achieve a better "balance", or
4. Horn load the thing and adjust for horizontal axis.
Hello,
Tenson: There was no EQ applied at all, but with a gated measurement like that I wouldn't pay any attention to anything below 500 or 600 Hz. The NEO8 was mounted in what basically amounts to a shoebox shape without the back. Maybe a reflection off the back wall caused a small boost at 500Hz, when really it should have been rolling off. If you look closely at the graphs you'll see that there weren't many data points taken below 1000Hz, which may account for the higher levels.
Scott: To quote directly from the BG paper: "In some applications where flat on-axis response is desirable, the use of a notch filter at 12kHz is recommended; in others correction may not be necessary, since spatial averaging will provide smoother power response without the 12kHz peak reduction. . . The notch filter schematic is given at the end of this document"
In my application I'm using the ribbons fairly close range as computer monitors, so on axis response is my first priority. I would imagine the whole thing would be a non-issue with several ribbons in a line array, and several meters of distance, but I'm not running them that way.
I found them quite terrible sounding without the notch while listening near-field. I would imagine that there are at least a few other people using them in a similar setup to my own, so if that's the case, the notch filter I mentioned above might be a better choice. It does measure much better, and I prefer the sound very much.
I do like the horn loading idea, but I'm not sure how it would do with such a near-field situation. I'm only about 3 feet from each ribbon, and they're placed directly at ear height and angled in for on-axis listening. I find the high freq. roll-off too severe in any other position, so this is where I am now. Do you think a horn would provide any benefit in such tight quarters?
Anyhow, back to the measurements to see if I can get a TB W5 woofer to play ball with these ribbons and fill in the gap below 600 Hz. I'm into active crossovers and bi-amplification from here on in.
Regards,
Owen
Tenson: There was no EQ applied at all, but with a gated measurement like that I wouldn't pay any attention to anything below 500 or 600 Hz. The NEO8 was mounted in what basically amounts to a shoebox shape without the back. Maybe a reflection off the back wall caused a small boost at 500Hz, when really it should have been rolling off. If you look closely at the graphs you'll see that there weren't many data points taken below 1000Hz, which may account for the higher levels.
Scott: To quote directly from the BG paper: "In some applications where flat on-axis response is desirable, the use of a notch filter at 12kHz is recommended; in others correction may not be necessary, since spatial averaging will provide smoother power response without the 12kHz peak reduction. . . The notch filter schematic is given at the end of this document"
In my application I'm using the ribbons fairly close range as computer monitors, so on axis response is my first priority. I would imagine the whole thing would be a non-issue with several ribbons in a line array, and several meters of distance, but I'm not running them that way.
I found them quite terrible sounding without the notch while listening near-field. I would imagine that there are at least a few other people using them in a similar setup to my own, so if that's the case, the notch filter I mentioned above might be a better choice. It does measure much better, and I prefer the sound very much.
I do like the horn loading idea, but I'm not sure how it would do with such a near-field situation. I'm only about 3 feet from each ribbon, and they're placed directly at ear height and angled in for on-axis listening. I find the high freq. roll-off too severe in any other position, so this is where I am now. Do you think a horn would provide any benefit in such tight quarters?
Anyhow, back to the measurements to see if I can get a TB W5 woofer to play ball with these ribbons and fill in the gap below 600 Hz. I'm into active crossovers and bi-amplification from here on in.
Regards,
Owen
opc:
I'm a little surprised that an off-axis position doesn't work for you around 12 kHz. My question is not that your filter works for you, but rather are you sure that the filter is what's doing it for you at that freq.? Odd question, but my thinking is that it still needs padding down lower in freq. and that IF that were done AND if the driver was off-axis (in this case "firing" behind you), that you might not ALSO perceive a better balanced sound AND have a marginal advantage in increased soundstaging and "air". (..also, when its utilized this way lower in freq. you tend to have imaging outside of the driver's "boundaries", recording dependent.)
As for the horn - why bother in your case? Don't fix it if it its not broken!
Actually, (depending on the air load) you can attenuate uppper freq.s and boost lower freq.s - but to get it correct/flat for that driver would require a lot of building trial and error or utilizing a VERY complex modeling package (with of course lots of trial and error processing). (..'bout the only software I can think of that will allow you very free-form specific horn profiles would be Akabak - and apparently it isn't easy to learn to use the program.) On the other hand if you WERE to do this then at least you could do away with passive filter (if it IS a passive filter, if its active (digital in particular) then there is even less reason to attempt it), and additionally you would lower non-linear distortion at lower freq.s (..but if you didn't do it well you would almost certainly increase linear distortion). Also, the additional mass at lower freq.s and the decreased dispersion would also likely make for a more "dynamic" (but also "forward") sound (..respectively).
I'm a little surprised that an off-axis position doesn't work for you around 12 kHz. My question is not that your filter works for you, but rather are you sure that the filter is what's doing it for you at that freq.? Odd question, but my thinking is that it still needs padding down lower in freq. and that IF that were done AND if the driver was off-axis (in this case "firing" behind you), that you might not ALSO perceive a better balanced sound AND have a marginal advantage in increased soundstaging and "air". (..also, when its utilized this way lower in freq. you tend to have imaging outside of the driver's "boundaries", recording dependent.)
As for the horn - why bother in your case? Don't fix it if it its not broken!
Actually, (depending on the air load) you can attenuate uppper freq.s and boost lower freq.s - but to get it correct/flat for that driver would require a lot of building trial and error or utilizing a VERY complex modeling package (with of course lots of trial and error processing). (..'bout the only software I can think of that will allow you very free-form specific horn profiles would be Akabak - and apparently it isn't easy to learn to use the program.) On the other hand if you WERE to do this then at least you could do away with passive filter (if it IS a passive filter, if its active (digital in particular) then there is even less reason to attempt it), and additionally you would lower non-linear distortion at lower freq.s (..but if you didn't do it well you would almost certainly increase linear distortion). Also, the additional mass at lower freq.s and the decreased dispersion would also likely make for a more "dynamic" (but also "forward") sound (..respectively).
I am looking for a commercial loudspeakers, where the RD-series from Bohlender Graebener like those from follow URLs are in use.
http://daveayers.com/rd_ribbons.pdf
http://www.top-audio.de/service/hl-refreshing/referenzen/alle-marken/refreshing-gen25/bohlender3.jpg
http://www.top-audio.de/service/hl-refreshing/referenzen/alle-marken/refreshing-gen25/bohlender2.jpg
http://www.top-audio.de/service/hl-refreshing/referenzen/alle-marken/refreshing-gen25/bohlender1.jpg
Perhaps one of the members kow this.
And a second question.
Is the United States website from German's version
BOHLENDER-GRAEBENER | SPEAKER
the follow URL ?
BG Radia - Home Theater Speakers
http://daveayers.com/rd_ribbons.pdf
http://www.top-audio.de/service/hl-refreshing/referenzen/alle-marken/refreshing-gen25/bohlender3.jpg
http://www.top-audio.de/service/hl-refreshing/referenzen/alle-marken/refreshing-gen25/bohlender2.jpg
http://www.top-audio.de/service/hl-refreshing/referenzen/alle-marken/refreshing-gen25/bohlender1.jpg
Perhaps one of the members kow this.
And a second question.
Is the United States website from German's version
BOHLENDER-GRAEBENER | SPEAKER
the follow URL ?
BG Radia - Home Theater Speakers
Last edited:
Question 2: yes.
Question 1: Answered by question 2 (i.e. from BG corp itself).
Additionally Wisdom Audio initially was a "high-end" spin-off corp from BG. I *think* Wisdom Audio's drivers are a bit different now, if still produced by BG.
Genesis Loudspeakers likely uses drivers from BG as well.
Yeah, the planars used in the higher end Wisdom products are VERY different from the stock BG models. I know that they were manufactured by BG still several years ago, but not sure now...
Check out these babies! Puts my little Neo10 to shame.
High performance in-wall and on-wall architectural products and subwoofers with Audyssey Room Correction - Wisdom Audio
Greg
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.