I also wanted to mention glues. Glue technology has advanced by leaps and bounds over the years and there is no reason that the glue should be a limiting factor in any loudspeaker design. That is not o say that it isn't important, it is, but its not like one cannot get glues that do the job correctly, not that everyone does.
We made speakers that would survive in an exhaust system running at 200 degree C for 100,000 miles. Glues were an issue, sure, but one that we solved.
What I learned about glues is that most problems are with the substrate not the glue itself. No glue can stick to an oily/dirty substrate. And there is dirt and oil in the air especially in a factory. Most failures come from not having a good surface on the parts.
We made speakers that would survive in an exhaust system running at 200 degree C for 100,000 miles. Glues were an issue, sure, but one that we solved.
What I learned about glues is that most problems are with the substrate not the glue itself. No glue can stick to an oily/dirty substrate. And there is dirt and oil in the air especially in a factory. Most failures come from not having a good surface on the parts.
Last edited:
Aurora,
Many of the older adhesives and glue are limited in specific locations due to environmental reasons, just about anything goes in Asia though. The reality is that most will only be using single component materials that require no mixing and are meant to be as fast as possible to keep that from being a bottleneck in a production line. Single component systems with automatic dispensers that are either UV light cured or CA type of adhesives that just need some moisture in the air to cure. Epoxy and other adhesives that may be used in a very high heat application are just to slow for most production lines and require a two component dispenser costing more and requiring much greater control over the dispenser. This is purely economics on the scale of most speaker production lines.
Many of the older adhesives and glue are limited in specific locations due to environmental reasons, just about anything goes in Asia though. The reality is that most will only be using single component materials that require no mixing and are meant to be as fast as possible to keep that from being a bottleneck in a production line. Single component systems with automatic dispensers that are either UV light cured or CA type of adhesives that just need some moisture in the air to cure. Epoxy and other adhesives that may be used in a very high heat application are just to slow for most production lines and require a two component dispenser costing more and requiring much greater control over the dispenser. This is purely economics on the scale of most speaker production lines.
Earl,
the physical requirements of the adhesives are one thing but not the end all of this conversation. Yes you need to make sure the surfaces are clean and free of any mold release or dust but with something like a polypro cone you would also want to use a corona discharge type of treatment to increase the surface bonding strength. Of course that surface treatment has a time factor as the surface will revert to a natural state and so that has to be an inline process.
Outside of the physical requirements the actual acoustic requirement are often just ignored if it will slow the line speed. I think one of the most common adhesives used today would be the Locktite Black Max rubber modified CA adhesives, but they are not necessarily the best thing to use between the cone and surround, they just happen to be very fast, the basic reason that they are in such great usage.
High temperature epoxies are typically slow curing, the higher the heat distortion point the slower the reactivity, fine for a pro audio driver for a voicecoil attachment but not for a high speed production line. Other two components systems also have other processing parameters that just don't fit with most high speed lines and the difficulty of mixing and dispensing, some are very environmentally sensitive to processing that affects the dispenser and the shelf life of the materials. It is not as simple as just the adhesive bond and the Tc rating of the materials for the best sounding devices.
the physical requirements of the adhesives are one thing but not the end all of this conversation. Yes you need to make sure the surfaces are clean and free of any mold release or dust but with something like a polypro cone you would also want to use a corona discharge type of treatment to increase the surface bonding strength. Of course that surface treatment has a time factor as the surface will revert to a natural state and so that has to be an inline process.
Outside of the physical requirements the actual acoustic requirement are often just ignored if it will slow the line speed. I think one of the most common adhesives used today would be the Locktite Black Max rubber modified CA adhesives, but they are not necessarily the best thing to use between the cone and surround, they just happen to be very fast, the basic reason that they are in such great usage.
High temperature epoxies are typically slow curing, the higher the heat distortion point the slower the reactivity, fine for a pro audio driver for a voicecoil attachment but not for a high speed production line. Other two components systems also have other processing parameters that just don't fit with most high speed lines and the difficulty of mixing and dispensing, some are very environmentally sensitive to processing that affects the dispenser and the shelf life of the materials. It is not as simple as just the adhesive bond and the Tc rating of the materials for the best sounding devices.
The question should be : finding a new P13 5.1/4 wide range driver ?
Some Audax are always manufactured
Some PHL Audio deserve a listening
Supravox has a very good sounding 13 cm & 16 cm which could surprise many here with their musical qualities
Some Scan Speak in the Revelator line can mate maybe (not so width in FR ?) the P13 Scan-Speak
PRV Audio seems to have a good wide mid-woof 5" which measures like a FR : PRV Audio 5MR450-NDY 5" Midrange Neodymium Woofer 8 Ohm
Focal in the sandwich line has Something a trade off à la Audio-Technology in less expensive !
The EMS speaker are good but expensive, but developped to glue to the ears of the previous Creator of FERTIN speakers.
Here it's a list from developping countries, but I'm sure there ar a lot more elswhere !
What else ?
Some Audax are always manufactured
Some PHL Audio deserve a listening
Supravox has a very good sounding 13 cm & 16 cm which could surprise many here with their musical qualities
Some Scan Speak in the Revelator line can mate maybe (not so width in FR ?) the P13 Scan-Speak
PRV Audio seems to have a good wide mid-woof 5" which measures like a FR : PRV Audio 5MR450-NDY 5" Midrange Neodymium Woofer 8 Ohm
Focal in the sandwich line has Something a trade off à la Audio-Technology in less expensive !
The EMS speaker are good but expensive, but developped to glue to the ears of the previous Creator of FERTIN speakers.
Here it's a list from developping countries, but I'm sure there ar a lot more elswhere !
What else ?
Last edited:
AuroraB,
In pro audio things are very different than in a consumer application. You are never going to pump the power levels into a voicecoil that would be so common in a real pro audio driver, that will require different materials to handle those real extreme temperature levels and the forces on a large cone driver. At the same time your not going to see a polypro cone on a professional cone driver meant for a PA application, I think you would see meltdown at the voicecoil former to cone junction. This is not an apples to apples comparison, a true pro audio driver has very different operating conditions.
Boutique speaker builders can do whatever they want, they are not under the same constraints of a pro audio device or a mass produced speaker that is much more cost sensitive. If your only producing a few hundred or even a few thousand speakers per year you can use the best materials available if you are truly after the ultimate in performance. Paper cone drivers are still the standard by which all other speakers are compared.
In pro audio things are very different than in a consumer application. You are never going to pump the power levels into a voicecoil that would be so common in a real pro audio driver, that will require different materials to handle those real extreme temperature levels and the forces on a large cone driver. At the same time your not going to see a polypro cone on a professional cone driver meant for a PA application, I think you would see meltdown at the voicecoil former to cone junction. This is not an apples to apples comparison, a true pro audio driver has very different operating conditions.
Boutique speaker builders can do whatever they want, they are not under the same constraints of a pro audio device or a mass produced speaker that is much more cost sensitive. If your only producing a few hundred or even a few thousand speakers per year you can use the best materials available if you are truly after the ultimate in performance. Paper cone drivers are still the standard by which all other speakers are compared.
I was ranting about the smooth polycone Vifa P13WH used in the Ariel not being available any more, but I see this US-available Aurum Cantus AC-130F1 5-1/4" Woofer driver that does much the same:
Aurum Cantus AC-130F1 5-1/4" Woofer
Jeff Bagby uses that one in his little BBC LS3/5A homage Continuum. Lynn, what's stopping you using that modern and available driver in the Ariel Mk. 2?
Nothing to stop anyone from developing a modern successor to the Ariel, using drivers in current production. For myself, I'm not interested, because I have a personal stash of replacement Scan-Speak tweeters and the very last Vifa 5.5" drivers. (I contacted Madisound before they were sold out. By the way, I don't get a discount from them, or from Vifa.)
If you want an Ariel-like sound, here are the items I would concentrate on:
* The midbass driver should be musically pleasing without any help from the crossover or equalization. In other words, it should be similar to a fullrange driver but without the 2~5 kHz artifacts and roughness in the breakup region. This is a key requirement.
* When looking at the CSD of the tweeter or midbass, there should be as little energy storage in the 1~10 kHz region as possible. This is also a key requirement, and the area where many exotic-cone drivers fall down pretty badly.
* A primary goal of the system is to have an impulse response decay characteristic as good or better than well-regarded electrostats. The cabinet must have large-radii left and right edges in order to get this kind of impulse response, and the bass enclosure must be free of standing waves.
* Although 1st-order crossovers offer the possibility of accurate square-wave reproduction (provided the tweeter is recessed, which brings its own set of problems), excursion control for the tweeter is quite poor. (Excursion for direct-radiators increases at a rate of 12 dB/octave as frequency is lowered, so 1st-order crossovers actually result in increasing excursion below the nominal highpass crossover frequency. It takes a 12 dB/octave highpass to merely keep excursion constant, and 18 dB/octave to reduce excursion.)
I find that low-Q 2nd-order electrical networks offer a reasonable compromise between group-delay error and control of tweeter excursion. A tweeter that is moving too much will sound raspy (or blurred, depending on design) when the program material has a dense spectrum and starts to get loud. That's part of the reason exhibitors at hifi shows like to use demo material that has a sparse HF spectrum ... audiophile blues, for example. Many high-end speakers collapse when you give them full-spectrum symphonic or choral music at concert levels.
* The system is optimized for transient response, but not leading-edge transients nor accuracy with square waves. The optimization is for the most rapid decay characteristic, and one that is free of resonance as it decays to zero. Most drivers have problems with this. A few drivers have good decay characteristics, and subjectively, as a group, they have a "sweet" sound, and are not zingy, raspy or mechanical sounding.
* Part of the purpose of the large-radii cabinet edges is to be able to see the favorable decay characteristics of the driver when looking at the CSD. If the edges are sharp, this alone generates so much clutter in the time domain that it's difficult to see what the driver is doing.
* Since my days at Audionics, I design crossovers so the acoustical phase match between drivers (at crossover) is 5~10 degrees or better. This can be confirmed by temporarily reverse-phasing the tweeter and looking for a null (at 2 meters distance) that is at least 25 dB deep.
* The overall spectral curve is smooth with a 1~2 dB down-tilt from 100 Hz to 10 kHz. In most rooms, this gives a subjectively flat sound with pink-noise excitation.
* For quick subjective assessment, I primarily use pink-noise, with orchestral and choral music as a cross-check. Rock music and movie soundtracks can provide a quick impression of dynamics and "snap", but this is heavily influenced by outside factors like the recording and the state of tune of the rest of the system (DACs, phonograph, power amps, etc).
There! You can design your own Ariel, if you like.
Last edited:
Thanks for your being so generous with your hard-won conclusions, Lynn.
In the small speaker (perhaps desktop) category, there is a little Fostex (GX100 Limited) that has caught my attention for reasons all associated with the criteria you list in my sampling above.
In this demo video of the difference between the cone of the Limited version versus the also-available cheaper aluminum-coned version, the 'cones' are dropped onto a hard surface; first aluminum, then magnesium. The (settling) noise the aluminum one makes is exactly what I would like not to have accompanying my music! Cone drop demo starts at 1:35:
https://www.youtube.com/watch?v=wrdYofPYD4M
Finally, I was recording a choral concert last week and took along a pair of respected pro grade headphones (Beyers) to compare to my long-used Etymotic Research ER4Ss, because, if they worked well they would be more convenient than the in-ears. The Beyers sound very good on most material, but in comparison to the Etymotics, when using the live sound as the comparison, they were astonishingly clouded and cluttered through the entire frequency range, and it sounded like nothing other than slow, chaotic settling behaviour. Putting the Etymotics back in was subjectively somewhat like there was nothing in my ears, so close to the real thing was the sound.
I have not seen CSD graphs for these in-ear phones; indeed data showing settling behaviour is too rare for headphones (and microphones!).
The Madisound Speaker Store
This Seas Curv poly cone seems to be well damped, but is ~7" and has a rising response (more crossover parts...). But I don't know about this driver's CSD. Or if it complies with all prerogatives for making some Ariel-based speaker.
H1520-08 U16RCY/P
The 5" version don't have the smoothness like the 7" version, too bad...
Only my 2 cents
(or meybe one cent... )
This Seas Curv poly cone seems to be well damped, but is ~7" and has a rising response (more crossover parts...). But I don't know about this driver's CSD. Or if it complies with all prerogatives for making some Ariel-based speaker.
H1520-08 U16RCY/P
The 5" version don't have the smoothness like the 7" version, too bad...
Only my 2 cents
(or meybe one cent... )
I have always found the in-the-ear headphones to work best. I used the Etymotics for years then I did a huge study of headphones for a client. There were some better ones like the unbelievably inexpensive Visang R02. I now use the Visang instead of the Etymotic. The belief in the industry is that Killions target curve is not correct.
Earl,
That seems to be the million dollar question, what should the target response be of the in ear and over ear headphones. I have yet to see a definitive answer to that question. The test setup has still not been qualitatively standardized as far as I can follow. It seems that price as in so many other areas has little to do with the sound quality of these devices, it is a personal selection at this point. I see no real way to compared these devices except for actually listening to them to make a choice. trying to narrow down the selections by technical means seem pretty useless.
That seems to be the million dollar question, what should the target response be of the in ear and over ear headphones. I have yet to see a definitive answer to that question. The test setup has still not been qualitatively standardized as far as I can follow. It seems that price as in so many other areas has little to do with the sound quality of these devices, it is a personal selection at this point. I see no real way to compared these devices except for actually listening to them to make a choice. trying to narrow down the selections by technical means seem pretty useless.
This was studied by Sean Olive at Harman. He found what I found as well, that flat was still the best overall approach as judged by a panel of listeners. The test setup is standardized, its the response at the ear drum using a real ear impedance device (like the Zwislocki that I used.) I compared dozens of in-ear phone and found that price, as usual, had no relevance to performance. Many of the very expensive units used two transducers (some even three,) which to me, is overkill and just a waste of money. Makes for good marketing though.
Earl,
That is the curvature that I would have assumed was best, at the same time perhaps a slight downward slope as with normal speakers would also be acceptable. I will have to look up the Zwislocki device, I haven't seen that name before. It does seem there is still much disagreement about how to test over ear headphones with a head form and how to make accurate measurements.
I found the Zwislocki device on Google and will take a look at those papers. Thanks for that reference.
That is the curvature that I would have assumed was best, at the same time perhaps a slight downward slope as with normal speakers would also be acceptable. I will have to look up the Zwislocki device, I haven't seen that name before. It does seem there is still much disagreement about how to test over ear headphones with a head form and how to make accurate measurements.
I found the Zwislocki device on Google and will take a look at those papers. Thanks for that reference.
