Interesting.
Why do you think that is? (re emphasized text above)
Some time last year, I 3d printed a couple of the meta-puck back chamber @Patrick Bateman created & tried them on the SB26ADC, for which they were built. They reduced the amplitude & freq of Fs but introduced a 2nd peak in the impedance curve. I also tried these meta-chambers on the Yamaha JA0801 2.7" beryllium mid domes -- again with mixed results. For the mid-dome, the PB's meta-chamber is probably too small? I'd like to revisit this project soon & eliminate that 2nd peak.
@mikessi Your observations with the tweeter modifications are exactly in line with all of the research I conducted regarding the behavior of individual chamber resonances and those coupled together, specifically the air volume trapped behind the VC gap. That resonance is problematic for lots of reasons, but trying to eliminate (or just reduce) it can be difficult. This area is practically inaccessible after the motor asy is assembled.
Some manufacturers use a perforated VC former to lessen the peak by means of venting the cavity, but it rarely makes a significant difference. Ferrofluid also helps, mainly due to the mechanical VC dampening but that's also a bandaid. The best way is to make the chamber as odd shaped as possible which breaks up the primary standing radial wave. Thats however just a symptom of the air cavity itself regardless of how the chamber is shaped.
The reason why enlarging the back chamber worsens the secondary resonant peak is the way the chambers couple together, interacting with each other. The larger the ratio of rear chamber volume behind the dome is to the VC gap chamber, the higher the 2nd peak becomes. Why is this a problem? Most tweeters have a double amplitude hump above primary Fs which has a dip in between the two FR humps. This is where 3rd order HD will often peak above the average 2nd order HD. It also intermodulates with higher breakup frequencies of the diaphragm itself.
What's even more frustrating is some manufacturers use the double hump response to raise sensitivity down lower. Thats typical on inexpensive dome tweeters, but its not necessary if the designer isn't greedy with sensitivity despite an undersized motor.
Your metamaterial chamber uses randomly broken up smaller volumes to diffuse the resonances. It doesn't address the volume behind the VC gap, which couples stronger if the primary chamber volume can flow more freely and react faster with the VC gap air volume as a helmholz resonator. The only solution to reliably reducing the secondary peak is somehow adding dampening material behind the VC gap or redesigning the whole chamber altogether. Thats why SS came up with the Air Circ motor on their higher end tweeters. Seas used a toroidal chamber design which lowered mechanical Q to the point of reducing the reactive coupling with the VC gap volume. Venting the motor is better than venting the VC former in this case but there are mixed opinions on which is more effective.
TL style chambers with tapering cross sections don't address the secondary gap resonance. This is the primary offender in raised 3rd order HD observed on most dome tweeters. The smaller the 2nd resonance is, the cleaner the tweeter will usually sound down low. This also applies to the primary resonance in some ways. The thing to look for in a well designed tweeter suitable for low crossover points is a very clean, symmetrical Fs peak without any other visible bumps just above it. You can also easily go backwards modifying tweeters in this regard. I usually favor the toroidal chamber shape with a neo motor which inherently has a smaller VC gap volume and no sharp corners inside the back chamber. Most of my favorite tweeters have this sort of design (Audax TW025A28, Seas T35C002). The Audax actually has no dampening material inside it whatsoever and its one of the cleanest sounding smaller domes I've heard to date. The Seas T35C002 has minimal dampening, using a single piece of felt rolled up inside the rear pole cavity. It also sounds superb in many ways.
My philosophy is usually to identify excellent examples of drivers and analyze their design as to why and how they manage to sound so good.
Some manufacturers use a perforated VC former to lessen the peak by means of venting the cavity, but it rarely makes a significant difference. Ferrofluid also helps, mainly due to the mechanical VC dampening but that's also a bandaid. The best way is to make the chamber as odd shaped as possible which breaks up the primary standing radial wave. Thats however just a symptom of the air cavity itself regardless of how the chamber is shaped.
The reason why enlarging the back chamber worsens the secondary resonant peak is the way the chambers couple together, interacting with each other. The larger the ratio of rear chamber volume behind the dome is to the VC gap chamber, the higher the 2nd peak becomes. Why is this a problem? Most tweeters have a double amplitude hump above primary Fs which has a dip in between the two FR humps. This is where 3rd order HD will often peak above the average 2nd order HD. It also intermodulates with higher breakup frequencies of the diaphragm itself.
What's even more frustrating is some manufacturers use the double hump response to raise sensitivity down lower. Thats typical on inexpensive dome tweeters, but its not necessary if the designer isn't greedy with sensitivity despite an undersized motor.
Your metamaterial chamber uses randomly broken up smaller volumes to diffuse the resonances. It doesn't address the volume behind the VC gap, which couples stronger if the primary chamber volume can flow more freely and react faster with the VC gap air volume as a helmholz resonator. The only solution to reliably reducing the secondary peak is somehow adding dampening material behind the VC gap or redesigning the whole chamber altogether. Thats why SS came up with the Air Circ motor on their higher end tweeters. Seas used a toroidal chamber design which lowered mechanical Q to the point of reducing the reactive coupling with the VC gap volume. Venting the motor is better than venting the VC former in this case but there are mixed opinions on which is more effective.
TL style chambers with tapering cross sections don't address the secondary gap resonance. This is the primary offender in raised 3rd order HD observed on most dome tweeters. The smaller the 2nd resonance is, the cleaner the tweeter will usually sound down low. This also applies to the primary resonance in some ways. The thing to look for in a well designed tweeter suitable for low crossover points is a very clean, symmetrical Fs peak without any other visible bumps just above it. You can also easily go backwards modifying tweeters in this regard. I usually favor the toroidal chamber shape with a neo motor which inherently has a smaller VC gap volume and no sharp corners inside the back chamber. Most of my favorite tweeters have this sort of design (Audax TW025A28, Seas T35C002). The Audax actually has no dampening material inside it whatsoever and its one of the cleanest sounding smaller domes I've heard to date. The Seas T35C002 has minimal dampening, using a single piece of felt rolled up inside the rear pole cavity. It also sounds superb in many ways.
My philosophy is usually to identify excellent examples of drivers and analyze their design as to why and how they manage to sound so good.
About D7608: out of curiosity, does anybody know if this driver is being used in any current commercial speaker?
AFAIK this Quested monitor still uses it: https://quested.com/v3110/
AFAIK this Quested monitor still uses it: https://quested.com/v3110/
Here's a fun one I found earlier in this thread - pretty expensive but identifiable drivers all https://www.elfton.cz/?p=65 (HDS, D75, L26ROY)
In the old days, there was the cool looking Dali 15 and 700 and the Dunlavy SC-V; sure there were others.
Also, I remember someone did an extensive refurbish of an old Vifa kit 3 or 4 way with the D75; I'll try to find it cause it was awesome.
In the old days, there was the cool looking Dali 15 and 700 and the Dunlavy SC-V; sure there were others.
Also, I remember someone did an extensive refurbish of an old Vifa kit 3 or 4 way with the D75; I'll try to find it cause it was awesome.
Last edited:
The Jean Maurer 370Fs use the D7608.
The Quested V3110 have a custom version D7608 and Morel CAT308.
Last edited:
I found that thread! These bad boys were called the Vifa Basis https://www.diyaudio.com/community/threads/re-built-vifa-basis-95-2023.395894/
Interesting video! Not cheap speakers, either. I'm intrigued by the cabinet - it seems to have no internal bracing, and a sand-filled rear panel.
I noticed that they use the D7608 with what looks like the factory felt, and put it into a double-walled custom chamber, of probably at least a litre. It has a thick felt-type lining material round the walls of the chamber (though no central fill). The felt appears to sit against the outer part of the factory felt, probably covering the outer ring of holes.
Last edited:
@ianbo I've heard these speakers on a few occasions. They sound very natural, transparent but not boring for a typical 3 way. They put alot of effort into the woofer, as the midrange is crossed around 500 hz based on crude measurements and there's no sign of any coloration from the cabinet. The midrange has the unmistakable D7608 character. The chamber net volume is smaller than it appears, around 400 - 500 ml. There is a thick layer of felt lining in there. The main thing that stood out was the vocals. They sound lifelike and very balanced top to bottom with the correct lower mids.
A friend of a friend has owned a similar model of these for several years now and doesn't plan to get rid of them anytime soon. He has a small vinyl mastering studio and uses these to proof his work, which is usually older mono jazz restorarions. He also usually goes through a pair of speakers every year or two, so that does say something about them. He also has a pair of larger Harbeths, but now mostly uses his JMs instead.
A friend of a friend has owned a similar model of these for several years now and doesn't plan to get rid of them anytime soon. He has a small vinyl mastering studio and uses these to proof his work, which is usually older mono jazz restorarions. He also usually goes through a pair of speakers every year or two, so that does say something about them. He also has a pair of larger Harbeths, but now mostly uses his JMs instead.
What an absolutely amazing thread with so much knowledge shared.
Cudos to @profiguy for opening up this subject and sharing your vast knowledge on domes.
There have been so many valuable contributions in this thread, and I’m inclined to start reading all over again just to be sure, I didn’t miss anything of importance.
The thread also came to my attention with a superbly good timing, as I’m in the prototype phase of building a 4 way speaker.
I had nailed the upper midrange down to two (cone)contenders: Scan Speak 12MU and Audio Technology 15J52. So far the 12MU seems to have the upper hand, as it is so easy to work with and sound marvellous. You really have to work with 5J52 to get the best out of it. It ain’t easy.
Now I simply must include at least one, but probably two, dome contenders, as they fit in nicely in my design concept: The Scan Speak D7608 and Satori MD60N-6. Working range 800 Hz to 3-4 kHz. Estimated sensitivity 91-93 dB/2.83 V.
Oh - and they also fit another design constraint, that is of being of Danish design 😊
I will eagerly await news on the closed cavity for the D7608 to fully explore the potential of this large dome.
Cudos to @profiguy for opening up this subject and sharing your vast knowledge on domes.
There have been so many valuable contributions in this thread, and I’m inclined to start reading all over again just to be sure, I didn’t miss anything of importance.
The thread also came to my attention with a superbly good timing, as I’m in the prototype phase of building a 4 way speaker.
I had nailed the upper midrange down to two (cone)contenders: Scan Speak 12MU and Audio Technology 15J52. So far the 12MU seems to have the upper hand, as it is so easy to work with and sound marvellous. You really have to work with 5J52 to get the best out of it. It ain’t easy.
Now I simply must include at least one, but probably two, dome contenders, as they fit in nicely in my design concept: The Scan Speak D7608 and Satori MD60N-6. Working range 800 Hz to 3-4 kHz. Estimated sensitivity 91-93 dB/2.83 V.
Oh - and they also fit another design constraint, that is of being of Danish design 😊
I will eagerly await news on the closed cavity for the D7608 to fully explore the potential of this large dome.
@SMABB I appreciate the kind words and hope you're able to get the info you need from this thread.
The SS 12MU is IMO one of the best cone mids being produced and comes close to the performance of a higher end mid dome. Very hard to outdo that one with any type of midrange. It costs a fortune.
That driver also reminds me alot of the 12M4631, sort of an oddball stuck in the middle as a hybrid of both cone and dome driver. It too costs a small fortune at over $300 and has a somewhat low sensitivity figure compared to other cone mid drivers. The sound of this driver is however nothing short of amazing. It has this rich and complex harmonic midrange texture similar to a higher end fullrange but it also sounds accurate, smooth and detailed. Too bad it costs so much.
The D7608 has endured over the years and I'm trying to save it from: extinction by introducing it to a wider range of diy people. It deserves a second chance and is capable of hanging with many other excellent cone midrange drivers even though it has a bandwidth disadvantage. With the revised rear chamber it should be able to deliver the performance in the lower midrange to keep up with an excellent woofer in a 3 way design. It should do so while being capable of higher midrange resolution and accuracy than most common cone drivers can deliver. I want to prove how much performance is obtainable from this obsxure mid dome despite the somewhat underwhelming factory spec sheet holding it back.
The SS 12MU is IMO one of the best cone mids being produced and comes close to the performance of a higher end mid dome. Very hard to outdo that one with any type of midrange. It costs a fortune.
That driver also reminds me alot of the 12M4631, sort of an oddball stuck in the middle as a hybrid of both cone and dome driver. It too costs a small fortune at over $300 and has a somewhat low sensitivity figure compared to other cone mid drivers. The sound of this driver is however nothing short of amazing. It has this rich and complex harmonic midrange texture similar to a higher end fullrange but it also sounds accurate, smooth and detailed. Too bad it costs so much.
The D7608 has endured over the years and I'm trying to save it from: extinction by introducing it to a wider range of diy people. It deserves a second chance and is capable of hanging with many other excellent cone midrange drivers even though it has a bandwidth disadvantage. With the revised rear chamber it should be able to deliver the performance in the lower midrange to keep up with an excellent woofer in a 3 way design. It should do so while being capable of higher midrange resolution and accuracy than most common cone drivers can deliver. I want to prove how much performance is obtainable from this obsxure mid dome despite the somewhat underwhelming factory spec sheet holding it back.
@profiguy Thanks for mentioning the 12M4631.
I haven’t tried that particular Revelator mid, since it is not within my requirements, but I have its sibling, the 18M4631. They both have the sliced paper membrane. When pairing 18M with 12MU crossed at 725-750 Hz using LR2, you simply cannot dissect either one. They sound as one, and that sound is utterly amazing and “right” to my ears. Let’s hear if any of the two domes can outperform this combination.
I have a couple of other low midrange drivers to try out with the domes. Some specially build Audio Technology drivers in 6” and 8” and the Satori MW16P.
I watch this thread not to miss any new additions and shared knowledge.
Thank you so much to all contributors 👏🏻
I haven’t tried that particular Revelator mid, since it is not within my requirements, but I have its sibling, the 18M4631. They both have the sliced paper membrane. When pairing 18M with 12MU crossed at 725-750 Hz using LR2, you simply cannot dissect either one. They sound as one, and that sound is utterly amazing and “right” to my ears. Let’s hear if any of the two domes can outperform this combination.
I have a couple of other low midrange drivers to try out with the domes. Some specially build Audio Technology drivers in 6” and 8” and the Satori MW16P.
I watch this thread not to miss any new additions and shared knowledge.
Thank you so much to all contributors 👏🏻
@SMABB I really like your philosophy on this stuff. A good sign your drivers are integrating well is them becoming homogeneous sounding. The slit cone Revalators are beautiful drivers and probably some of the finest sounding cone drivers ever made. The 18M4631T is one of my favorite 6" midbass drivers to date. It has such smooth response up to 4k +. Also, running LR2 filters in the lower mids preserves more of the transient response. Thats something often overlooked by designers and decided against to obtain the lowest possible filter points. Not always desirable compared to some more important priorities.
Last edited:
In your own recomendation based on your knowings, what is the best lower and higher xover-points for that mid?
Is your recomendation any different for a real high SPL guy?
And the 12M4631 ? (really like you say an oddball) and sadly a little low sensitivity.
Have 2 pairs of the 18M4631T and one pair of 12M4631 and SS 7000, a "small" MTM will maby be quite good for my SPL needings. 10 liter in total !
@jawen With the 18M4631T the main limiting parameter is thermal power of the VC. Despite the relatively large VC for a 7" midbass, the titanium former doesn't do well with shedding heat. Titanium is a natural heat insulator, which may sound like a surprise, but thats the reality of it. About 200 hz 2nd order LR2 is the point where thermal (80W continuous) and excursion limits intersect. For listening levels in excess of 106 dB (per stereo pair), you'll want to derate that by 25% if operating in a small sealed box, so about 60W is what the driver can take long term thermally limited. This also keeps you in the 2% THD window. Adding a second driver covering the same bandwidth adds 3 - 4 dB for 2 drivers. So for a thermally comfortable 110 dB you'll be running 2 x mids per side in 12 - 14 liters sealed total, I'd aim for a min 200 hz LR2 HP. Thats to avoid the high Qtc rolloff of the enclosure, not corrupting the electrical HP slope. Otherwise group delay will start negatively effecting transient behavior.
The 12M4631 is good to about 300 - 325 hz LR2 HP until you start stressing the mid thermally to get to105 dB continuously. These are conservative numbers, but that 38mm Ti former VC will start thermally compressing. I always advocate staying under 2% THD in the lower mids, but some people don't care about distortion as much as others. Power compression is very audible and stresses the driver out, also shifting TSPs by quite a bit.
The 12M4631 is good to about 300 - 325 hz LR2 HP until you start stressing the mid thermally to get to105 dB continuously. These are conservative numbers, but that 38mm Ti former VC will start thermally compressing. I always advocate staying under 2% THD in the lower mids, but some people don't care about distortion as much as others. Power compression is very audible and stresses the driver out, also shifting TSPs by quite a bit.
Well, maybe in comparison to copper and aluminum. Titanium has a thermal conductivity of about 20 W/(m-K), while aluminum alloys are in the range of 150-250, and copper is in the range of 300 - 400.
Titanium thermal conductivity is about the same as steel, which is also in the range of 15 - 30 depending on the steel alloy. Solder is about 50.
For comparison, fiberglass reinforced plastic with polyester or epoxy resin is going to be about 0.3 W/(m-K). Almost any polymer is going to be less than 1.
So among metals, titanium is on the low side. But all metals have very high thermal conductivity compared to other materials, such as wood, plastics, and ceramics.
There may be other effects going on the contribute to the thermal compression in drivers using titanium voice coil formers, but I don't think that thermal conductivity by itself can explain the compression.
I appreciate your opinions very much, thank you very much!
Experience and trail and error beats everything, some book tells you.
best regards John
@hifijim The reason for concern with running titanium VC formers on the hot side is its tendency to warp and distort more than other VC former materials of similar thermal conductivity. I know about this because I learned to tig weld titanium a long time ago when it was more of a strategic material, uncommonly used in luxury goods. Its very tricky to keep titanium from physically distorting, even when exposing it to mildly higher temps.
ScanSpeak domes arrived this morning! Will use with Scan alum 8 inch woofers and Morel CAT 378 tweeters. Triamped, Mini DSP, Topping PAS2 amp on the tweeters, Hypex NILAI on the mids, Behringer A800 on woofers. Enclosures being planned now. Cant wait! Will sit atop my 12" subs for home theater duty. Crown Class D on them.
Attachments
Last edited: