How many of you have listened to a set of dipoles and realized they do some things better "conventional" speakers?
How many of you have been curious about the buzz around waveguides, and wondered if it was time to try one out in your next project?
Has anyone considered using BOTH in the same speaker?
If so, read on...
I've been experimenting with horns then waveguides for close to twenty years. Though my experience with dipoles is not as deep, I've been curious to learn more, especially after reading linkwitzlab.com.
Before we get into the details of this project, I think it's important to state a few of my biases up front. I am not someone who does DIY projects to save money or to clone an existing design. So if you want to learn how to make a Linkwitz Orion or a Geddes Summa for $500, this isn't the thread for you. Both of those speakers are spectacular, and I'd wholeheartedly recommend BUYING a set if they suit your needs.
The idea of this project is take the influences of a handful of great speaker designers, and distill them into ONE speaker.
I wanted to name this speaker after a band that's been influenced by half a dozen genres. A lot of names popped into my head, but naming the speaker "Beck" or "Prince" didn't have a good ring to it. So I'm sticking with "Big Audio Dynamite."
Here's the wikipedia entry; watch this thread for the next few months, and we'll find out if the speaker can effectively mix as many influences as the band did.
Big Audio Dynamite
Big Audio Dynamite was a British musical group formed in 1984 by the ex-guitarist and singer of The Clash, Mick Jones. The group was noted for its effective mixture of varied musical styles, incorporating elements of punk rock, dance music, hip-hop, reggae, and funk into a unique sound.
How many of you have been curious about the buzz around waveguides, and wondered if it was time to try one out in your next project?
Has anyone considered using BOTH in the same speaker?
If so, read on...
I've been experimenting with horns then waveguides for close to twenty years. Though my experience with dipoles is not as deep, I've been curious to learn more, especially after reading linkwitzlab.com.
Before we get into the details of this project, I think it's important to state a few of my biases up front. I am not someone who does DIY projects to save money or to clone an existing design. So if you want to learn how to make a Linkwitz Orion or a Geddes Summa for $500, this isn't the thread for you. Both of those speakers are spectacular, and I'd wholeheartedly recommend BUYING a set if they suit your needs.
The idea of this project is take the influences of a handful of great speaker designers, and distill them into ONE speaker.
I wanted to name this speaker after a band that's been influenced by half a dozen genres. A lot of names popped into my head, but naming the speaker "Beck" or "Prince" didn't have a good ring to it. So I'm sticking with "Big Audio Dynamite."
Here's the wikipedia entry; watch this thread for the next few months, and we'll find out if the speaker can effectively mix as many influences as the band did.
Big Audio Dynamite
Big Audio Dynamite was a British musical group formed in 1984 by the ex-guitarist and singer of The Clash, Mick Jones. The group was noted for its effective mixture of varied musical styles, incorporating elements of punk rock, dance music, hip-hop, reggae, and funk into a unique sound.
This speaker is going to be a 2-way, which can be augmented by a sub. For the high end, I will probably use a BMS neo compression driver, since it worked so well for my last project.
For the woofer our priorities are high-excursion, good power handling, and smooth response to the crossover point. The crossover point will depend on the woofer; the bigger the woofer, the lower the crossover point.
What do you guys think of these choices:
GR-Research M165X
Dan Wiggins has convinced me that XBL^2 is the way to go. I've listened to his speakers with the Extremis, and the bass response is incredible. Unfortunately, it's no longer available. The M165X looks like the next best option. According to zaphaudio.com the M165X does not have a copper shorting ring. I believe Wiggins has argued that it's not necessary. Nonetheless, the 165X is compelling, but the Extremis would be an attractive option if it was still for sale.
Peerless Exclusive 8in Woofer
The Peerless has the shorting ring that the GR lacks, but it doesn't have as much xmax. However, the increase in surface area might make up for that. The GR costs less.
Seas CA22RNX
I stumbled across this after seeing a positive review of the ER18RNX. It's not clear if the CA22 has copper in the sleeve; if it does, it may be competitive with the Peerless, for less money.
Eminence Basslite CH2010
This one may look like an odd choice - it's a guitar speaker. It's xmax is among the lowest, and it's response isn't flat. But the excursion problem is offset by the high SD, and the rising response is out of our passband. I'm also a big fan of high efficiency drivers. It has a hemp cone, which is unusual.
P-Audio SN-12MB
One of the most underrated midranges I've come across. After hearing the Geddes Summa, I bought and measured one of these in the interest of making a high efficiency two-way. My measurements show that this speaker has flat response, high efficiency, and is reasonably priced. For our application, it's low xmax and low QTS may be a deal breaker. In a conventional enclosure, it's great.
I haven't settled on a midrange (yet.) The P-Audio is probably my favorite. It's low XMAX looks like a deal breaker, until we consider it's MUCH higher SD.
Lets do the math.
P-Audio SN12MB has an SD of 530cm^2 and an xmax of .35cm. That gives us a swept volume of 186cm^3, or .1855 liters.
Peerless Exclusive 8in has an SD of 221cm^2 and an xmax of .55cm. That gives us a swept volume of 122cm^3, or .122 liters.
What does this mean to us? It means that even though the pro sound midrange looks like a lousy candidate for a dipole mid, it's surface area trumps the XMAX of a more-modern motor design.
For the woofer our priorities are high-excursion, good power handling, and smooth response to the crossover point. The crossover point will depend on the woofer; the bigger the woofer, the lower the crossover point.
What do you guys think of these choices:
GR-Research M165X
Dan Wiggins has convinced me that XBL^2 is the way to go. I've listened to his speakers with the Extremis, and the bass response is incredible. Unfortunately, it's no longer available. The M165X looks like the next best option. According to zaphaudio.com the M165X does not have a copper shorting ring. I believe Wiggins has argued that it's not necessary. Nonetheless, the 165X is compelling, but the Extremis would be an attractive option if it was still for sale.
Peerless Exclusive 8in Woofer
The Peerless has the shorting ring that the GR lacks, but it doesn't have as much xmax. However, the increase in surface area might make up for that. The GR costs less.
Seas CA22RNX
I stumbled across this after seeing a positive review of the ER18RNX. It's not clear if the CA22 has copper in the sleeve; if it does, it may be competitive with the Peerless, for less money.
Eminence Basslite CH2010
This one may look like an odd choice - it's a guitar speaker. It's xmax is among the lowest, and it's response isn't flat. But the excursion problem is offset by the high SD, and the rising response is out of our passband. I'm also a big fan of high efficiency drivers. It has a hemp cone, which is unusual.
P-Audio SN-12MB
One of the most underrated midranges I've come across. After hearing the Geddes Summa, I bought and measured one of these in the interest of making a high efficiency two-way. My measurements show that this speaker has flat response, high efficiency, and is reasonably priced. For our application, it's low xmax and low QTS may be a deal breaker. In a conventional enclosure, it's great.
I haven't settled on a midrange (yet.) The P-Audio is probably my favorite. It's low XMAX looks like a deal breaker, until we consider it's MUCH higher SD.
Lets do the math.
P-Audio SN12MB has an SD of 530cm^2 and an xmax of .35cm. That gives us a swept volume of 186cm^3, or .1855 liters.
Peerless Exclusive 8in has an SD of 221cm^2 and an xmax of .55cm. That gives us a swept volume of 122cm^3, or .122 liters.
What does this mean to us? It means that even though the pro sound midrange looks like a lousy candidate for a dipole mid, it's surface area trumps the XMAX of a more-modern motor design.
One thing someone is bound to ask, is why not use dual midranges? For me, this is a question of cost, design complexity, and directivity.
Two Peerless eights will cost more than a single P-Audio 12; a D'Appolito midrange is trickier to cross over; and a BIG woofer offers superior directivity (this is why the Summa uses a VERY expensive 15in woofer.)
I ignored one of the most popular DIY eights because metal cones are a challenge to crossover, and it has low excursion. But what if we use TWO?
Here's the math.
Dual Dayton Reference eights have an SD of 442cm^2 and an xmax of .6cm. That gives us a swept volume of 265cm^3, or 0.265 liters.
P-Audio SN12MB has an SD of 530cm^2 and an xmax of .35cm. That gives us a swept volume of 186cm^3, or 0.1855 liters.
Peerless Exclusive 8in has an SD of 221cm^2 and an xmax of .55cm. That gives us a swept volume of 122cm^3, or .122 liters.
What does this mean to us? A single Dayton Reference eight is trumped by the Peerless, but a pair of them may be the best bet, particularly since they cost less than a SINGLE 12. (PE has them on sale for $35 right now, versus $135 for the SN12MB from loudspeakersplus.)
Two Peerless eights will cost more than a single P-Audio 12; a D'Appolito midrange is trickier to cross over; and a BIG woofer offers superior directivity (this is why the Summa uses a VERY expensive 15in woofer.)
I ignored one of the most popular DIY eights because metal cones are a challenge to crossover, and it has low excursion. But what if we use TWO?
Here's the math.
Dual Dayton Reference eights have an SD of 442cm^2 and an xmax of .6cm. That gives us a swept volume of 265cm^3, or 0.265 liters.
P-Audio SN12MB has an SD of 530cm^2 and an xmax of .35cm. That gives us a swept volume of 186cm^3, or 0.1855 liters.
Peerless Exclusive 8in has an SD of 221cm^2 and an xmax of .55cm. That gives us a swept volume of 122cm^3, or .122 liters.
What does this mean to us? A single Dayton Reference eight is trumped by the Peerless, but a pair of them may be the best bet, particularly since they cost less than a SINGLE 12. (PE has them on sale for $35 right now, versus $135 for the SN12MB from loudspeakersplus.)
panomaniac said:
On the compression driver.
IMHO, the best sounding speaker at the 2007 CES was from Emerald Physics.
While the sound is excellent, the WAF factor is nil. Based on my research into psychoacoustics and a lot of experimentation, the huuuuuge size of the Emerald Physics speakers is overkill.
Is anyone interested in the technical details? If so, let me know.
In a nutshell, here's the trick:
In order for a speaker to image well, two things are required. The frequency response of the two speakers must match, and each speaker must be time aligned. The TRICK is that we are EXTREMELY sensitive to time-alignment problems in the low frequencies, yet insensitive to frequency response problems; the opposite holds true at high frequencies, where time alignment is difficult to perceive, but a FR mismatch is noticeable.
So what does this mean?
Stay tuned for design details...
With the HF/LF hinge point in the 300~500 Hz range, I presume.
One of the little tricks we did at Audionics was match all speakers against the prototype. The matching technique was interesting: the L and R speakers would be facing forward, about a foot apart from each other, and connected in reverse-phase. The measuring microphone was about 1~2 meters away, and precisely in the midline. The goal was a full-spectrum null at least 25~30 dB deep.
When you think about it, this requires both precise level and phase-matching - within a few degrees - to attain a 25~30 dB null. I can tell you, it worked, and gave far more precise pair-matching than any other method, which usually doesn't take account of phase-matching across the pair.
One of the little tricks we did at Audionics was match all speakers against the prototype. The matching technique was interesting: the L and R speakers would be facing forward, about a foot apart from each other, and connected in reverse-phase. The measuring microphone was about 1~2 meters away, and precisely in the midline. The goal was a full-spectrum null at least 25~30 dB deep.
When you think about it, this requires both precise level and phase-matching - within a few degrees - to attain a 25~30 dB null. I can tell you, it worked, and gave far more precise pair-matching than any other method, which usually doesn't take account of phase-matching across the pair.
Lynn Olson said:
If you are referring to the Emerald Physics CS1, I believe it's around 1,000 Hz (with a 15" bass/mid vs the 8 inch pair on the earlier model).
Apparently Peter Walker used to do something similar for final stage QC at Quad with the ESL 57 - he'd put the speaker face-to-face and out of phase with the reference and play square waves.
Lynn Olson said:
At low frequencies, timing is key;
At high frequencies, matched frequency response is key.
There IS a "hinge point", which is different for every person.
That frequency is equivalent to the distance between your ears. For instance, if the distance for you is nine inches, the frequency is about 1500hz.
This psychoacoustic phenomenon has a number of implications:
- At 1500hz, we are incredibly sensitive to frequency AND timing.
- Coincidentally, our hearing acuity is higher at this frequency, which makes point #1 even more critical.
- At two or three octaves below this "hinge point", we can get away with frequency response problems, but timing is EVERYTHING. This is one of the fundamental mistakes people make with subwoofers; they don't time align them.
Needless to say, this project is going to put a lot of attention into time alignment of the low frequencies.
Patrick -
That's interesting information on the crossover point where the ear's sensitivity switches from timing-supersensitive to intensity-supersensitive. I hadn't really put that together quite like you did.
I'm aware that the mechanism by which the ears localize sound has the same crossover point; it's primarily arrival time below 1.5k, and primarily intensity above 1.5k.
Have you ruled out doing a two-way? Geddes and SP Technology and yours truly all get away with it, and Emerald Physics has gone to a two-way. That might sidestep your low frequency timing issues.
Which waveguide are you planning to use? Since there aren't very many waveguides to choose from, I'm thinking the waveguide choice will inevitably narrow down the midrange driver choice, which will in turn narrow down the woofer choice. Of course if you are able to build your own waveguide, you can make the other choices first and just build the waveguide to suit.
Duke
PS - sorry Patrick I was editing while you were replying.
That's interesting information on the crossover point where the ear's sensitivity switches from timing-supersensitive to intensity-supersensitive. I hadn't really put that together quite like you did.
I'm aware that the mechanism by which the ears localize sound has the same crossover point; it's primarily arrival time below 1.5k, and primarily intensity above 1.5k.
Have you ruled out doing a two-way? Geddes and SP Technology and yours truly all get away with it, and Emerald Physics has gone to a two-way. That might sidestep your low frequency timing issues.
Which waveguide are you planning to use? Since there aren't very many waveguides to choose from, I'm thinking the waveguide choice will inevitably narrow down the midrange driver choice, which will in turn narrow down the woofer choice. Of course if you are able to build your own waveguide, you can make the other choices first and just build the waveguide to suit.
Duke
PS - sorry Patrick I was editing while you were replying.
audiokinesis said:
I can make a good waveguide, and I already have a mold for an oblate spheroidal waveguide (which is what I used for last winter's project.)
But I'd prefer to just use what you and the the Emerald Physics folks are using.
Unfortunately Assistance Audio hasn't returned the last two emails I sent them, and US Speakers is sold out.
Patrick -
Emerald Physic's current model doesn't use the DDS waveguide, or so I've been told. I'm an Emerald Physic dealer, but that's a recent development and I won't have speakers until after RMAF.
I still use the DDS waveguide, and recently received two shipments totalling about a dozen units direct from the factory. So I think DDS will start shipping to Assistance Audio and US Speaker pretty soon. Initial quality control was not real high, but the second batch was definitely better than the first.
If you keep striking out with Assistance and US Speaker, shoot me an e-mail and I'll see what I can do.
Duke
Emerald Physic's current model doesn't use the DDS waveguide, or so I've been told. I'm an Emerald Physic dealer, but that's a recent development and I won't have speakers until after RMAF.
I still use the DDS waveguide, and recently received two shipments totalling about a dozen units direct from the factory. So I think DDS will start shipping to Assistance Audio and US Speaker pretty soon. Initial quality control was not real high, but the second batch was definitely better than the first.
If you keep striking out with Assistance and US Speaker, shoot me an e-mail and I'll see what I can do.
Duke
The emerald physics design looks really really interesting, it makes a lot of sense to use a DSP based xover when doing this type of design.
My guess is that they are using the following 24$ waveguide:
http://www.partsexpress.com/pe/showdetl.cfm?&Partnumber=270-312
According to people that have actually herad this speaker it is very nice, it is kind of surpricing, since I would not bet on integrating two 15" woofers with a 1" tweeter!
I guess the success hinges on making a low cross over point (1kHz for the CS2). This is made possible by the waveguide
(see Zaph's horn conversion study: http://www.zaphaudio.com/hornconversion.html )
I would like to experiment using two 12" or 15" woofers but with a waveguide loaded B&G neo8 that is operated in dipole mode.
My guess is that they are using the following 24$ waveguide:
http://www.partsexpress.com/pe/showdetl.cfm?&Partnumber=270-312
According to people that have actually herad this speaker it is very nice, it is kind of surpricing, since I would not bet on integrating two 15" woofers with a 1" tweeter!
I guess the success hinges on making a low cross over point (1kHz for the CS2). This is made possible by the waveguide
(see Zaph's horn conversion study: http://www.zaphaudio.com/hornconversion.html )
I would like to experiment using two 12" or 15" woofers but with a waveguide loaded B&G neo8 that is operated in dipole mode.
I was BTW refering to the new emerald physics design (the CS2), much better WAF IMHO.
http://www.emeraldphysics.com./intro.htm
http://www.emeraldphysics.com./intro.htm
I took the plunge and purchased two waveguides for this project. Barring any major disaster, I'll be using the 18Sound XT1086.
While I have my own mold for an OS waveguide, I wanted this to be a project that others could try too. And I don't expect many people would be willing to learn how to make their own mold. Also, my mold isn't perfectly symmetrical. Even with an existing mold, finishing the WG is a P.I.T.A.
The DDS ENG 1-90PRO is *probably* a better waveguide, but it also cost 2X as much. More importantly, it's difficult to acquire right now.
So here are my negative expectations of the 18sound WG:
- For some stupid reason there's a diffraction device in the throat.
- I'd prefer a 90x90 coverage pattern, like the DDS
- 18Sound doesn't compensate for the 6 degree entrance angle into the WG. Then again, nobody but Geddes and Audio Intelligence do.
And my positive expectations:
- The measured response of the WG is unusually good. Predictable polar response is a great reason to go with something off-the-shelf.
- The 18Sound is cast aluminum. It might ring, but I'll treat it. Should be sturdy.
- It's much cheaper.
- Supply seems to be consistent - anyone can buy one of these today from Loudspeakers Plus.
I should have them by Thursday.
While I have my own mold for an OS waveguide, I wanted this to be a project that others could try too. And I don't expect many people would be willing to learn how to make their own mold. Also, my mold isn't perfectly symmetrical. Even with an existing mold, finishing the WG is a P.I.T.A.
The DDS ENG 1-90PRO is *probably* a better waveguide, but it also cost 2X as much. More importantly, it's difficult to acquire right now.
So here are my negative expectations of the 18sound WG:
- For some stupid reason there's a diffraction device in the throat.
- I'd prefer a 90x90 coverage pattern, like the DDS
- 18Sound doesn't compensate for the 6 degree entrance angle into the WG. Then again, nobody but Geddes and Audio Intelligence do.
And my positive expectations:
- The measured response of the WG is unusually good. Predictable polar response is a great reason to go with something off-the-shelf.
- The 18Sound is cast aluminum. It might ring, but I'll treat it. Should be sturdy.
- It's much cheaper.
- Supply seems to be consistent - anyone can buy one of these today from Loudspeakers Plus.
I should have them by Thursday.
Using Martin's worksheets, I've come up with a design for the Dayton Reference 8" in an open baffle.
The first thing you'll notice is that the reponse droops very gently, so it's going to sound "thin." In fact it's 12db down at just 100hz.
I did this by design; the slow rolloff will be compensated with a BSC filter. In conjunction with BSC, we'll have flat response from about 100hz up. (There will be a sub later in the thread.)
Here's the response.
The first thing you'll notice is that the reponse droops very gently, so it's going to sound "thin." In fact it's 12db down at just 100hz.
I did this by design; the slow rolloff will be compensated with a BSC filter. In conjunction with BSC, we'll have flat response from about 100hz up. (There will be a sub later in the thread.)
Here's the response.
Patrick Bateman said:
After modeling the Dayton 8", I modeled the P-Audio 12". I was expecting the Dayton to clobber the P-Audio, but this isn't the case. Even though the Dayton is smaller, it has a higher Q and the same FS. So it *should* play lower, but the graphs are almost identical.
The P-Audio has noticeably higher efficiency. The Dayton is much cheaper.
Patrick Bateman said:
Someone who attended the RMAF commented positively about the Lyngdorf open baffle speakers. Just for fun, I modeled the Dayton Reference 8" in a OB that's identical to what they are using. I found the dimensions on their website.
It looks like the Lyngdorf folks know what they are doing. Using a baffle half as large, their design is only 3db down at the 300hz crossover frequency. The ripple in my modeled response would be evened out a bit by their curved and tapered baffle (which can't be modeled easily.)
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.