I agree with Magnetar's post that SPL claims based on frequency response graphs or sensitivity to pink-noise have to be treated with caution. The number of interest is the one derived from the Theile/Small efficiency, sometimes stated as a percentage efficiency, and sometimes as SPL at one meter into a half-space (on a flat baffle, in other words). This is commonly the lower efficiency figure when there are two numbers quoted.
Why does this matter? The T/S figure is valid for the flat-response, piston-band region of response, not the rising-response or peaky upper range of the driver. That region can have peaks up to 5~10 dB higher than the piston-band region - and I don't think measuring peaks and resonances is a valid method of describing true conversion efficiency.
P.S. This is a historical remnant of Fifties/Sixties practice of blasting a driver with 2.83V of pink-noise and seeing what the averaging SPL meter says. This has the marketing advantage of rewarding speakers with midrange peaks and grossly nonflat response curves - this was a common practice with "West Coast Sound" speakers, the JBL L100 being the most famous example.
Why does this matter? The T/S figure is valid for the flat-response, piston-band region of response, not the rising-response or peaky upper range of the driver. That region can have peaks up to 5~10 dB higher than the piston-band region - and I don't think measuring peaks and resonances is a valid method of describing true conversion efficiency.
P.S. This is a historical remnant of Fifties/Sixties practice of blasting a driver with 2.83V of pink-noise and seeing what the averaging SPL meter says. This has the marketing advantage of rewarding speakers with midrange peaks and grossly nonflat response curves - this was a common practice with "West Coast Sound" speakers, the JBL L100 being the most famous example.
Thanks also to JohnK's post showing what emerges from each end of a non-symmetrical acoustical environment.
This, by the way, is why I'm a little cautious about the shape of the rear baffle - if there is any tendency towards cavity resonance, that will (strongly) appear in the backwave and will also appear out the front (as small ripples). Partial boxes with recycled cotton filling seems the best approach when a quasicardiod (a la Gary Pimm) is desired.
This, by the way, is why I'm a little cautious about the shape of the rear baffle - if there is any tendency towards cavity resonance, that will (strongly) appear in the backwave and will also appear out the front (as small ripples). Partial boxes with recycled cotton filling seems the best approach when a quasicardiod (a la Gary Pimm) is desired.
And more thanks to Mige0 for the links to the Prodance site with extensive MLSSA CSD and Theile/Small measurements.
Most of the 6" and 8" drivers don't look all that good. The graphic below compares the CSD of the complete Ariel system (above) to the Eminence Beta8 (below). It's evident the Beta8 needs extensive equalization, including strong in-band notch filters, to approach the flatness of the Ariel (which has very simple filtration for the midbass drivers). I'm not too sure if the Beta8 could ever match the time-domain performance of the Ariel.
Another surprise on this site is the performance of the Selenium 18WS600. It looks like a 2nd-order low-pass filter in the 200~300 Hz region would give exemplary performance as a bass driver. This isn't something you'd expect from an 18" driver.
The additional specs on the Parts Express site shows pretty respectable 2nd and 3rd-harmonic distortion figures as well.
Another gratifying surprise is the performance of the 18Sound XT120 waveguide combined with the B&C DE25 compression driver.
Good information here! Thanks again!
Most of the 6" and 8" drivers don't look all that good. The graphic below compares the CSD of the complete Ariel system (above) to the Eminence Beta8 (below). It's evident the Beta8 needs extensive equalization, including strong in-band notch filters, to approach the flatness of the Ariel (which has very simple filtration for the midbass drivers). I'm not too sure if the Beta8 could ever match the time-domain performance of the Ariel.
Another surprise on this site is the performance of the Selenium 18WS600. It looks like a 2nd-order low-pass filter in the 200~300 Hz region would give exemplary performance as a bass driver. This isn't something you'd expect from an 18" driver.
The additional specs on the Parts Express site shows pretty respectable 2nd and 3rd-harmonic distortion figures as well.
Another gratifying surprise is the performance of the 18Sound XT120 waveguide combined with the B&C DE25 compression driver.
Good information here! Thanks again!
Attachments
Lynn,
the question that interested me most during this thread, and that remains unanswered, is Shinobiwan right when he says any tweeter, no matter how good it is, shouldn`t be crossed at 1.5 kHz. This is derived from his experiences with the single-height RAAL, and I ask myself whether the observed thin-ness is due to the rolloff below 3kHz, and will improve with the double RAAL. Any experiences so far?
Oliver
the question that interested me most during this thread, and that remains unanswered, is Shinobiwan right when he says any tweeter, no matter how good it is, shouldn`t be crossed at 1.5 kHz. This is derived from his experiences with the single-height RAAL, and I ask myself whether the observed thin-ness is due to the rolloff below 3kHz, and will improve with the double RAAL. Any experiences so far?
Oliver
None to report so far. Alexander, who ought to know, says the double-high should sound excellent down to 1.6 kHz - below which frequency, Alexander feels, no ribbon, regardless of size, should be used. During the development phase, he auditioned the new tweeter against his other tweeters, and reported it will make extraordinary demands on the quality of the widerange driver - simply because the new tweeter has much less distortion - and more headroom - than the previous tweeters. More area, more efficiency, etc. etc.
This is why I am seriously looking at the 6ND410 used in a 2,3, or 4x array, or a custom-designed front waveguide, or combined with the 10NDA610 or 12NDA520 as midbass fill-in driver. (In other words, the 6ND410 from 300 Hz to 3 kHz, and the 10NDA610 or 12NDA520 operating in parallel from 300 Hz to 1 kHz.)
The overall system could be as simple as a pair of side-by-side Selenium 18" woofers in a floor-standing open-backed cabinet with a moderate amount of recycled cotton filling, a transparent curved acrylic/plexiglass "wing" for a 10NDA610/12NDA520 with a 6ND410 on top of it, and topping the whole array, the double-high RAAL tweeter - or the 18Sound waveguide and a good 1" mylar-film compression driver. The overall system efficiency would be in the same range as Magnetar's system mentioned earlier - probably 103 dB/metre.
This is very different kind of OB than the audiophile-oriented ones that are on the market now, which are compromised dynamically and make serious demands on the amplifiers (due to heavy bass-boost EQ). What Magnetar has already built is showing the way forward for wide dynamic range OB systems.
This is why I am seriously looking at the 6ND410 used in a 2,3, or 4x array, or a custom-designed front waveguide, or combined with the 10NDA610 or 12NDA520 as midbass fill-in driver. (In other words, the 6ND410 from 300 Hz to 3 kHz, and the 10NDA610 or 12NDA520 operating in parallel from 300 Hz to 1 kHz.)
The overall system could be as simple as a pair of side-by-side Selenium 18" woofers in a floor-standing open-backed cabinet with a moderate amount of recycled cotton filling, a transparent curved acrylic/plexiglass "wing" for a 10NDA610/12NDA520 with a 6ND410 on top of it, and topping the whole array, the double-high RAAL tweeter - or the 18Sound waveguide and a good 1" mylar-film compression driver. The overall system efficiency would be in the same range as Magnetar's system mentioned earlier - probably 103 dB/metre.
This is very different kind of OB than the audiophile-oriented ones that are on the market now, which are compromised dynamically and make serious demands on the amplifiers (due to heavy bass-boost EQ). What Magnetar has already built is showing the way forward for wide dynamic range OB systems.
Driving an array of 6.5ers up to 3kHz is for sure no textbook solution.
http://www.audiodiycentral.com/resource/pdf/nflawp.pdf
http://www.audiodiycentral.com/resource/pdf/nflawp.pdf
On further reflection, it looks like the 6ND410 and 10NDA610 were designed to work together. Both have Theile/Small efficiencies of 98dB/metre, the 6" has an Efficiency Bandwidth Product (Fs/Qes = EBP) of 444, the 10" has an EBP of 371 (both quite high), the 6" has a linear Xmax of 4mm, and the 10" has a linear Xmax of 5mm (both pretty good for midrange drivers).
Fed from a voltage source and working in parallel, they should deliver 104 dB/metre from 2.83V rms. If they share a common lowpass filter set to 3 kHz, with the 10" driver having an additional 1st-order lowpass around 1~1.5 kHz with a possible notch filter at 3 kHz, the combination should have a smooth crossover to a HF radiator at 3 kHz.
The 6ND410/10NDA610 combination gets me back to my original goal of a 12" midrange driver, at least in terms of radiating area in the "power band". With the 6" driver having an Fs of 120 Hz, the 10" driver an Fs of 89 Hz, a common low-Q 2nd-order highpass filter set between 240 and 300 Hz should give extremely high power-handling, far above audiophile-exotic drivers.
P.S. As for the array, the original idea was to use one driver for full bandwidth, and all of the others for bass/mid reinforcement, with a similar choice of 1~1.5 kHz transition frequencies. The downside of the array is the fact they don't really sum all that well - when you sit to one side, or move up and down too much, the non-coincident arrivals from the multiple centers of radiation are noticeable as combing or "phasiness".
The two-driver solution is simpler and easier to tune the crossover for the desired summing action.
Fed from a voltage source and working in parallel, they should deliver 104 dB/metre from 2.83V rms. If they share a common lowpass filter set to 3 kHz, with the 10" driver having an additional 1st-order lowpass around 1~1.5 kHz with a possible notch filter at 3 kHz, the combination should have a smooth crossover to a HF radiator at 3 kHz.
The 6ND410/10NDA610 combination gets me back to my original goal of a 12" midrange driver, at least in terms of radiating area in the "power band". With the 6" driver having an Fs of 120 Hz, the 10" driver an Fs of 89 Hz, a common low-Q 2nd-order highpass filter set between 240 and 300 Hz should give extremely high power-handling, far above audiophile-exotic drivers.
P.S. As for the array, the original idea was to use one driver for full bandwidth, and all of the others for bass/mid reinforcement, with a similar choice of 1~1.5 kHz transition frequencies. The downside of the array is the fact they don't really sum all that well - when you sit to one side, or move up and down too much, the non-coincident arrivals from the multiple centers of radiation are noticeable as combing or "phasiness".
The two-driver solution is simpler and easier to tune the crossover for the desired summing action.
I hope this has nothing to do with my post about AER. I just found Magnetar's afirmation very interesting and thought how could this be achieved. For example I see an obvious choise in using one of the high spl AER drivers whizeerless, in a custom waveguide, from 300Hz up to 4Khz or so. It wasn't meant to disaprove or something.
Woofers
Gentlemen,
I'm intrigued with PE's latest sales special of Lanzar Max12" dual coil subwoofers :
non pressed paper cone
4 ohms per coil (I'd wire them for 8 ohms)
VCdia: 2"
Xmax:10mm
Fs: 33
SPL: 88dB (their single 4 ohm coil model goes to 92 dB)
Qts: 0.47
Magnet weight: 130 oz.
Please give me your opinions about these before I order any. I would only use them up to a max of 300Hz.
dobias
Gentlemen,
I'm intrigued with PE's latest sales special of Lanzar Max12" dual coil subwoofers :
non pressed paper cone
4 ohms per coil (I'd wire them for 8 ohms)
VCdia: 2"
Xmax:10mm
Fs: 33
SPL: 88dB (their single 4 ohm coil model goes to 92 dB)
Qts: 0.47
Magnet weight: 130 oz.
Please give me your opinions about these before I order any. I would only use them up to a max of 300Hz.
dobias
I like the way you are thinking these days Lynn, although you are getting into 4-way territory here. It was appealing to use multiples of the same driver rolled off at different frequencies, and still is a concept worth exploring, but it is becoming clear why it generally isn't done that way. I think that the high output required, combined with t he open baffle is the main reason it is beginning to look like it won't be the best approach.
On the other hand, to get the kind of output that you are talking about, 4 ways are not unusual in the pro world, so I guess not a surprise.
No matter whether you choose 6" 10", or 12" drivers of course there is no way to be sure that you have chosen the right drivers for you until you have listened to them
What is nice is that we know of various appealing, but very different candidates, including 18 Sound, Alnico Tone Tubbies, Behma, etc, - about 4 drivers really need to be auditioned for each size. Then for the tweeter, you have the RAAL ribbon and 2 or 3 different 1" compression drivers.
Then you are done!! So, I think that your first effort will really be a test mule, and your huge contribution to DIY will be to ferret out your favorite drivers. I suspect that some manufacturers will send some "loaners" or give you a good price. The ones you buy and don't use can surely be sold as they will all be desirable units.
On the other hand, to get the kind of output that you are talking about, 4 ways are not unusual in the pro world, so I guess not a surprise.
No matter whether you choose 6" 10", or 12" drivers of course there is no way to be sure that you have chosen the right drivers for you until you have listened to them
What is nice is that we know of various appealing, but very different candidates, including 18 Sound, Alnico Tone Tubbies, Behma, etc, - about 4 drivers really need to be auditioned for each size. Then for the tweeter, you have the RAAL ribbon and 2 or 3 different 1" compression drivers.
Then you are done!! So, I think that your first effort will really be a test mule, and your huge contribution to DIY will be to ferret out your favorite drivers. I suspect that some manufacturers will send some "loaners" or give you a good price. The ones you buy and don't use can surely be sold as they will all be desirable units.
I would be very surprised if John Harrison at Tone Tubby didnt... these dyed in the wool rockers have become fascinated by the audio DIY interest in their stuff. They are responding accordingly, and have several new drivers in the works. The 8 inch guitar cone, sadly with mud magnetics, but the FR curve, Fs/50, Qts/.70, and reasonable SPL/93db also look really good for scaled down iterations' midband driver.
Lynn, if you approached John or Charlie, they would put you onto one Jason Scheuner, who is their "speaker guy" and you could have a new collaboration, right there. Just thinking out loud here. I have spoken at length with Jason, very energetic, and interested in expanding and advancing the HempTone thang...
Lynn, if you approached John or Charlie, they would put you onto one Jason Scheuner, who is their "speaker guy" and you could have a new collaboration, right there. Just thinking out loud here. I have spoken at length with Jason, very energetic, and interested in expanding and advancing the HempTone thang...
Sounds At An Exhibition
I just came back from the Athens High End show, and I could not help but remembering Lynn. What he said about nothing being built around the concept of reproducing real sound as experienced in a hall. Just experiments on mega luxury is all I saw and heard. Commercial expensive sound has nothing to do with music. Thank God we can still wield a soldering iron and a screwdriver in this neck of woods.
I just came back from the Athens High End show, and I could not help but remembering Lynn. What he said about nothing being built around the concept of reproducing real sound as experienced in a hall. Just experiments on mega luxury is all I saw and heard. Commercial expensive sound has nothing to do with music. Thank God we can still wield a soldering iron and a screwdriver in this neck of woods.
Early on in this thread..
I seem to remember an early proposed configuration by someone that used 4 side firing woofers. Can anyone point me back to that? I'm trying to avoid rereading the first 50 pages to find it.
I'm wondering if a U shaped open baffle speaker has 16" deep sides and an 8" baffle if the effective baffle length is 40"? and how the bass performance is affected by side firing a pair of 12" woofers on each side? I know the "nulls" would be facing the lstening position but also know that the polar response in the bottom octaves isn't as affected as much as the midrange. Thoughts? I have already read John K's and SL publishings on the U frame dipoles and just wanted to hear a few people weigh in. I attached a pic in case my english isn't clear.
Thanks,
C
I seem to remember an early proposed configuration by someone that used 4 side firing woofers. Can anyone point me back to that? I'm trying to avoid rereading the first 50 pages to find it.
I'm wondering if a U shaped open baffle speaker has 16" deep sides and an 8" baffle if the effective baffle length is 40"? and how the bass performance is affected by side firing a pair of 12" woofers on each side? I know the "nulls" would be facing the lstening position but also know that the polar response in the bottom octaves isn't as affected as much as the midrange. Thoughts? I have already read John K's and SL publishings on the U frame dipoles and just wanted to hear a few people weigh in. I attached a pic in case my english isn't clear.
Thanks,
C
Attachments
Hi
Lynn, know you don't like EV and I agree on that they can be brute, but for me the EVL 12S_OEM is a speaker with most beautiful looking CSD out of that list.
Should go down to 120 Hz at 110 dB with no problem and also has an exceptional "drive" number though performing not so hot in Qms.
Greetings
Michael
Lynn Olson said:
Lynn, know you don't like EV and I agree on that they can be brute, but for me the EVL 12S_OEM is a speaker with most beautiful looking CSD out of that list.
Should go down to 120 Hz at 110 dB with no problem and also has an exceptional "drive" number though performing not so hot in Qms.
Greetings
Michael
Re: Early on in this thread..
OK for subwoofers with sharp lowpass filtering, but using through the midrange (as drawn) would be problematical. Note the magnets of the drivers are facing each other - this is a good technique for vibration cancellation (which is why we see this configuration using in closed-box woofer arrays with rods connecting the woofers to each other).
However, in a conventional closed-box configuration, this array is typically used in a 3 or 4-way system, since the tall, narrow box is prone to severe internal standing-wave modes, and off-axis listeners hear two arrivals, one for the left, and another for the right array of woofers. A sharp-cut lowpass filter reduces audibility of the box modes as well as the differential arrival times of the wavefronts from the L and R woofer arrays.
Keep in mind diaphragms are acoustically transparent, so you can "see" one driver through another. This is why Isobarik alignments must be aggressivly lowpass filtered, otherwise the cavity between the two woofers not only has non-coincident arrivals from the front and rear drivers, but standing waves formed by the cavity itself. (With no drivers mounted in it at all, an Isobarik would form a resonant cavity.)
U and H-baffles are partway back to closed-boxes in terms of standing-wave colorations, thus the requirement for lowpass filtering to minimize coloration. Moreover, the lowpass filtering has to be well below midrange frequencies, since the 300~600 Hz range is exactly where conventional closed-boxes suffer the most from standing waves (due to poor absorption of readily available damping materials at lower frequencies).
As mentioned earlier, I am not as interested in a particular polar pattern (dipole, quasicardioid, controlled 90-degree dispersion) as I am in the intrinsic sonics of the drivers themselves. I probably have a less charitable view of driver and cabinet sonics than many of my contemporaries - I'm looking for the handful of current-production drivers with low coloration, large dynamics, and low IM distortion.
These drivers are scarce, especially since I don't have the luxury of being an OEM that can pre-order several thousand custom models of prosound drivers. Rest assured that 18Sound, BMS, B&C, Beyma, Tone Tubby, Radian, et al make many OEM versions we don't get to see in the catalogs. These are proprietary, secret, we ain't gonna hear about these - and if we do, we can't buy them anyway, so no point in pursuing the matter.
So in the restricted universe we have, no unobtanium out-of-production drivers, no OEM versions, good CSD and IM distortion measurements by audiophile and professional standards, and - oh yes - it would be nice if they sounded good too.
chrismercurio said:
OK for subwoofers with sharp lowpass filtering, but using through the midrange (as drawn) would be problematical. Note the magnets of the drivers are facing each other - this is a good technique for vibration cancellation (which is why we see this configuration using in closed-box woofer arrays with rods connecting the woofers to each other).
However, in a conventional closed-box configuration, this array is typically used in a 3 or 4-way system, since the tall, narrow box is prone to severe internal standing-wave modes, and off-axis listeners hear two arrivals, one for the left, and another for the right array of woofers. A sharp-cut lowpass filter reduces audibility of the box modes as well as the differential arrival times of the wavefronts from the L and R woofer arrays.
Keep in mind diaphragms are acoustically transparent, so you can "see" one driver through another. This is why Isobarik alignments must be aggressivly lowpass filtered, otherwise the cavity between the two woofers not only has non-coincident arrivals from the front and rear drivers, but standing waves formed by the cavity itself. (With no drivers mounted in it at all, an Isobarik would form a resonant cavity.)
U and H-baffles are partway back to closed-boxes in terms of standing-wave colorations, thus the requirement for lowpass filtering to minimize coloration. Moreover, the lowpass filtering has to be well below midrange frequencies, since the 300~600 Hz range is exactly where conventional closed-boxes suffer the most from standing waves (due to poor absorption of readily available damping materials at lower frequencies).
As mentioned earlier, I am not as interested in a particular polar pattern (dipole, quasicardioid, controlled 90-degree dispersion) as I am in the intrinsic sonics of the drivers themselves. I probably have a less charitable view of driver and cabinet sonics than many of my contemporaries - I'm looking for the handful of current-production drivers with low coloration, large dynamics, and low IM distortion.
These drivers are scarce, especially since I don't have the luxury of being an OEM that can pre-order several thousand custom models of prosound drivers. Rest assured that 18Sound, BMS, B&C, Beyma, Tone Tubby, Radian, et al make many OEM versions we don't get to see in the catalogs. These are proprietary, secret, we ain't gonna hear about these - and if we do, we can't buy them anyway, so no point in pursuing the matter.
So in the restricted universe we have, no unobtanium out-of-production drivers, no OEM versions, good CSD and IM distortion measurements by audiophile and professional standards, and - oh yes - it would be nice if they sounded good too.