I have no problem with much of the essence of your discussion, however I would be surprised if a "similar quality" 15" driver was only 2x the mass of a 8" driver.
Most larger drivers are intended for higher power ratings and as such have heavier voice coils. To prevent cone break-up and maintain rigidity they can't be 'thiner' than an 8" driver and then there is that Pi.r^2 thingy ??????
Perhaps one of the learned ones with some spec sheets could give us some realistic numbers ???
cheers,
mark
Mefinnis:
You want real numbers? You got real numbers.
Peerless 15" PA speaker, SLS line. Moving mass: 105.7 grams.
Peerless CSX 217 8" speaker. Moving mass: 30.5 grams.
Three times the weight instead of twice. S3 is not outrageously off.
Can I be considered one of the learned ones now? 😉
You want real numbers? You got real numbers.
Peerless 15" PA speaker, SLS line. Moving mass: 105.7 grams.
Peerless CSX 217 8" speaker. Moving mass: 30.5 grams.
Three times the weight instead of twice. S3 is not outrageously off.
Can I be considered one of the learned ones now? 😉
I consider just about everybody apart from myself around here "learned" ..... 'tis safer, as I'm just a simple "medico".
3x is slightly more believable, there just had to be more than a simle 2x penalty, unless they were making the 15" cone from tissue-paper 😎
cheers
mark
3x is slightly more believable, there just had to be more than a simle 2x penalty, unless they were making the 15" cone from tissue-paper 😎
cheers
mark
Mefinnis:
The plot thickens.
Eminence Kappa 15 15" PA loudspeaker. Possibly the most popular 15 inch PA speaker in the world.
Diaphragm mass, including airload, (and probably voice coil): 74 grams.
About twice what the Peerless 8" diaphragm is. So maybe S3 is onto something.
The plot thickens.
Eminence Kappa 15 15" PA loudspeaker. Possibly the most popular 15 inch PA speaker in the world.
Diaphragm mass, including airload, (and probably voice coil): 74 grams.
About twice what the Peerless 8" diaphragm is. So maybe S3 is onto something.
Re: OK, my first hairbrained theory!!!
economics & profit. To get that low requires a big speaker (whether many little ones or 1 big one), and big speakers cost more to ship...
Also at the NW DIY today someone told me something i thot worth thinking about. We were talking about box resonances. The fellow said that manufactures like to build ported boxes because being pressure relieved the boxes are cheaper to build...
And IMO to get really low you need a speaker that matches well with the room gain -- usually a slow roll-off sealed box (or aperiodic or certain TLs) has a better chance than a fast roll-off ported box.
If you want to go this low, the speaker cannot be separted from the room -- eliminating 99% (or more) of commercial offereings.
I have seen this discussed to no end on the BASSlist and the lambda list. Size does not matter when it comes to starting & stopping (at least when we are talking woofers). Size is only a single factor in many of the speakers design & performance. One only needs experience the Lamda 15s...
dave
surf said:
economics & profit. To get that low requires a big speaker (whether many little ones or 1 big one), and big speakers cost more to ship...
Also at the NW DIY today someone told me something i thot worth thinking about. We were talking about box resonances. The fellow said that manufactures like to build ported boxes because being pressure relieved the boxes are cheaper to build...
And IMO to get really low you need a speaker that matches well with the room gain -- usually a slow roll-off sealed box (or aperiodic or certain TLs) has a better chance than a fast roll-off ported box.
If you want to go this low, the speaker cannot be separted from the room -- eliminating 99% (or more) of commercial offereings.
I have seen this discussed to no end on the BASSlist and the lambda list. Size does not matter when it comes to starting & stopping (at least when we are talking woofers). Size is only a single factor in many of the speakers design & performance. One only needs experience the Lamda 15s...
dave
EVM-15L: 57.2 g moving mass, Fs=43Hz, 103dB sensitivity.
Cone is rigid enough to be able to be used in front loaded horns with reasonable compression ratios.
dave said
<b>Size does not matter when it comes to starting & stopping (at least when we are talking woofers). Size is only a single factor in many of the speakers design & performance. One only needs experience the Lamda 15s...</b>
Choosing a driver determined on one specification isn't wise. To get decent bass, you need to move a lot of air, and little drivers can't do it alone.
Cone is rigid enough to be able to be used in front loaded horns with reasonable compression ratios.
dave said
<b>Size does not matter when it comes to starting & stopping (at least when we are talking woofers). Size is only a single factor in many of the speakers design & performance. One only needs experience the Lamda 15s...</b>
Choosing a driver determined on one specification isn't wise. To get decent bass, you need to move a lot of air, and little drivers can't do it alone.
I am trying to start debate!?!
Thanks everyone for the input.
I do not want to start a discussion on subwoofers, as I know this subject has been heavily debated elsewhere, but rather put forward these thoughts as a general foundation for where I am going with my speaker design and current design philosophy of speakers in general.
I have been swamped with life, work, children, spouse, and about 20 more 20 drivers, so I am not getting this posted as quickly as I would like.
I am still concentrating on the title of this thread, building a pair of reference speakers with 3 drivers, a subwoofer covering up to 75, wide-range - 75 - 8000, and tweeter - 8000 and up with minimal crossovers.
Lynn aka Surf, Sun & Sound
Thanks everyone for the input.
I do not want to start a discussion on subwoofers, as I know this subject has been heavily debated elsewhere, but rather put forward these thoughts as a general foundation for where I am going with my speaker design and current design philosophy of speakers in general.
I have been swamped with life, work, children, spouse, and about 20 more 20 drivers, so I am not getting this posted as quickly as I would like.
I am still concentrating on the title of this thread, building a pair of reference speakers with 3 drivers, a subwoofer covering up to 75, wide-range - 75 - 8000, and tweeter - 8000 and up with minimal crossovers.
Lynn aka Surf, Sun & Sound
i am also looking at the same problem. it makes sense to simplify this and lets just get a simple 2 way going first the woofer can be added later. The Bass List tried to do something similar some time back with a speaker called the Bass List reference speaker using a Audax Hm170Z0 and Morel MDT33. Mike McCall can reveal more.
bass can be added via multiple smaller woofers or a big woofer. bass will depend on the room resoances, location of speaker, WAF etc..
Does this seem like a plausible solution?
If we just list the best tweeters and bass midranges we can find we can simplify this.
the bass midrange must cover 70-4000
the tweeter can then be leaf/ribbon, ring radiator, etc..
the bass mid but handle 100W rms and be 91db or so its operating range. we can assume other drivers will compensate for difraction step etc.. since a real woofer/bass reinforcement section will likely be added later.
the reason for the high power high sens is that the speaker can then be driven by almost any amp. be it a SE tube or a FET/BJT or a AV Receiver.
a 91db db woofer will be about 89-90 db after the XO is added. usinga 10W SE one can still get a respectable 100db. Using a 50W AV reciever or power amp once can get 105db. After this compression will set in and hence the need for a good bass reinforcement section.
bass can be added via multiple smaller woofers or a big woofer. bass will depend on the room resoances, location of speaker, WAF etc..
Does this seem like a plausible solution?
If we just list the best tweeters and bass midranges we can find we can simplify this.
the bass midrange must cover 70-4000
the tweeter can then be leaf/ribbon, ring radiator, etc..
the bass mid but handle 100W rms and be 91db or so its operating range. we can assume other drivers will compensate for difraction step etc.. since a real woofer/bass reinforcement section will likely be added later.
the reason for the high power high sens is that the speaker can then be driven by almost any amp. be it a SE tube or a FET/BJT or a AV Receiver.
a 91db db woofer will be about 89-90 db after the XO is added. usinga 10W SE one can still get a respectable 100db. Using a 50W AV reciever or power amp once can get 105db. After this compression will set in and hence the need for a good bass reinforcement section.
I certainly agree with the objectives set by Lynn, I would love to have a set like this but I dont think I have knowledge and exp needed to accomplish all those goals. Even though I have built several 2 and 3 ways during the last 30 or more years. Please dont take wrong, since I don't want to be a party pooper, after saying this I will shear my 2 cents experiance with midrange and wide range speakers.
On most speaks to get to 4000 Hz or more one needs at least an FS of 100 and of course you wont be able to go down to 70Hz XO. I have tryed to cross the mid- bass mentioned by kelticwizard (Focal 5N412 DBL) at 200 and there were a lot of tradeoffs, even in a D'apollito config with parallel or serial connections, with LF compensations and so on with no good results. ( I have 4 of this speks, unable to use them). Also made a lot of effort trying to cross a Focal 7NXXX midrange at 200 Hz at no avail, I now have them in a TL but not even that worked. The lowest I have been able to cross them with good results is at 500 hz! with a limited high frec of arround 2000 or 2500 Hz and that with zobel network.
The trouble with mids and wide range we are trying to serch for this project is that most are very weak below 200Hz, if you try to go lower than that you'l get flimsy midbass and wont be able to cross it nicely to a powerfull sub.
One of the maybe few speakers that might fulfill Lynn's objectives is the Fortin (I have never heard it) There is a separate thread working with this speaker trying to accomplish a simmilar task. Ofcourse this is an expensive transducer and a big one, I think is a 12 incher! I don't think this is what it's in the minds of others in this thread.
In a more pedestrian mood I also think that some Fostex FR could come close to the desiered goal, but I have also heard that they are somewhat weak below 200Hz (and so happens with the overpriced Lowther, any model, even back horn loaded).
I would say that the first task we have is narrowing the search and test of a few alternatives of FR speks, the other 2 drivers (wof and tweet) can be more easily decided upon and there should be a big number of them that could be used for this project.
From my point of view we should stay away from D'apollitos and pasive radiators. I would definetly consider any TL (specialy), bass reflex and closed box in any of the 3 drivers. I Would also propose the goal that any of the speaks should cost less than $150 (better yet half of that). And on the XO issue to be first order when needed.
My first proposition (and only one for the moment) for FR speaker is the Fostex FE 167E at a cost of around $60 or $70 (group purchase any?). This speak might even allow not to use a tweet which of course will lower the total system cost and most important fewer XO parts and will provide all other important issues related to a single point driver.
(Have a look at the specs and graph at Madison.com)
This is a 6 inch unit which has a very smoth response and a benning impedance (at least what is shown on the manuf plot) and based on the graph we could be able to cross it at the desiered frequency. With this proposition I think the only XO components needed will be related to the low pass filter. This would be a suitable coil and maybe a zobel. This is a driver that should be tested. Anyone has heard this driver? ( I haven't)
As I have heard, this spek will work well on a bass reflex or TL.
Drawbacks: Uses a wizzer and not very good HF dispersion as on most 6" drivers.
What do you think? Any other candidate for FR?
On most speaks to get to 4000 Hz or more one needs at least an FS of 100 and of course you wont be able to go down to 70Hz XO. I have tryed to cross the mid- bass mentioned by kelticwizard (Focal 5N412 DBL) at 200 and there were a lot of tradeoffs, even in a D'apollito config with parallel or serial connections, with LF compensations and so on with no good results. ( I have 4 of this speks, unable to use them). Also made a lot of effort trying to cross a Focal 7NXXX midrange at 200 Hz at no avail, I now have them in a TL but not even that worked. The lowest I have been able to cross them with good results is at 500 hz! with a limited high frec of arround 2000 or 2500 Hz and that with zobel network.
The trouble with mids and wide range we are trying to serch for this project is that most are very weak below 200Hz, if you try to go lower than that you'l get flimsy midbass and wont be able to cross it nicely to a powerfull sub.
One of the maybe few speakers that might fulfill Lynn's objectives is the Fortin (I have never heard it) There is a separate thread working with this speaker trying to accomplish a simmilar task. Ofcourse this is an expensive transducer and a big one, I think is a 12 incher! I don't think this is what it's in the minds of others in this thread.
In a more pedestrian mood I also think that some Fostex FR could come close to the desiered goal, but I have also heard that they are somewhat weak below 200Hz (and so happens with the overpriced Lowther, any model, even back horn loaded).
I would say that the first task we have is narrowing the search and test of a few alternatives of FR speks, the other 2 drivers (wof and tweet) can be more easily decided upon and there should be a big number of them that could be used for this project.
From my point of view we should stay away from D'apollitos and pasive radiators. I would definetly consider any TL (specialy), bass reflex and closed box in any of the 3 drivers. I Would also propose the goal that any of the speaks should cost less than $150 (better yet half of that). And on the XO issue to be first order when needed.
My first proposition (and only one for the moment) for FR speaker is the Fostex FE 167E at a cost of around $60 or $70 (group purchase any?). This speak might even allow not to use a tweet which of course will lower the total system cost and most important fewer XO parts and will provide all other important issues related to a single point driver.
(Have a look at the specs and graph at Madison.com)
This is a 6 inch unit which has a very smoth response and a benning impedance (at least what is shown on the manuf plot) and based on the graph we could be able to cross it at the desiered frequency. With this proposition I think the only XO components needed will be related to the low pass filter. This would be a suitable coil and maybe a zobel. This is a driver that should be tested. Anyone has heard this driver? ( I haven't)
As I have heard, this spek will work well on a bass reflex or TL.
Drawbacks: Uses a wizzer and not very good HF dispersion as on most 6" drivers.
What do you think? Any other candidate for FR?
should that be madisound?
if the FE167 has a whizzer cone will that not interfere with the tweeter. i have not found many whizzer cone drivers that have good dispersion.
i have also played with a 7" Focal 7N303 and it was un uasable below 300Hz atleast at the SPLs I wanted.
another alternate is to use a ESL for the top end and a TL sub below. i know it has been done and is nothing new.
a 12" wide x 36" high ESL on top of a 18" high box using a 12" TL sub XOed at 200Hz is what I am hinting at.
The ESL will be 16" wide if you include a 2" wide wood frame and box below it will then be 16" wide, 18" high and about 24" deep giving an internal volume of about 60 liters. A small TL or a aperiodic damped box or a short heavily damped TL/port could work.
beore i get flamed for the last suggestion I once use a 12" (pioneer car woofer) in a 1 cu. ft. box with a 4" dia 8" long port that was heavily damped (I assume resultant loading was aperiodic). I was surprised by the speed and tightness of the resultant bass.
the enviroment was a car and eventually we ended up using 3 of these side by side driven by a 200W car amp. one of those lucky days i guess.
if the FE167 has a whizzer cone will that not interfere with the tweeter. i have not found many whizzer cone drivers that have good dispersion.
i have also played with a 7" Focal 7N303 and it was un uasable below 300Hz atleast at the SPLs I wanted.
another alternate is to use a ESL for the top end and a TL sub below. i know it has been done and is nothing new.
a 12" wide x 36" high ESL on top of a 18" high box using a 12" TL sub XOed at 200Hz is what I am hinting at.
The ESL will be 16" wide if you include a 2" wide wood frame and box below it will then be 16" wide, 18" high and about 24" deep giving an internal volume of about 60 liters. A small TL or a aperiodic damped box or a short heavily damped TL/port could work.
beore i get flamed for the last suggestion I once use a 12" (pioneer car woofer) in a 1 cu. ft. box with a 4" dia 8" long port that was heavily damped (I assume resultant loading was aperiodic). I was surprised by the speed and tightness of the resultant bass.
the enviroment was a car and eventually we ended up using 3 of these side by side driven by a 200W car amp. one of those lucky days i guess.
The Next Installment
One of the parameters I have laid down for these speakers is that they must be efficient enough to produce 100db+ with low-powered amps. I believe the more efficient a driver is the less inherent distortion there is for the whole system of sound reproduction. OK, before you slaughter me, I know this is not that simple, and this part of my theory is just my brainwave.
So let’s think about this abit, firstly, we can use amplifiers that are inherently more linear, such as Class A, whether tube or solid state. Also we do not have to work the voice coil as hard thus reducing heat and distortion caused by over-excursion. I remember the Mr. Nelson Pass saying he would rather be able to produce 100db with 1 watt than 100 watts. So I have this hunch that all things being equal the more efficient a speaker the greater its potential of having less distortion. Just an observation….
One of the laws of speaker design is the smaller the driver, the less efficient. So while drivers such the small Tagband’s, Jordan’s. Bandor’s and such fit the bill quite nicely in terms of response, on and off-axis, the price paid is quite low efficiency, most are under 90db/watt. So for efficiency we have to look at larger drivers, but then we are faced with very poor off-axis response, which is one thing we want as little as possible of.
A side product of the low efficiency of these smaller drivers is to get the spl out of then we want, with any amount of power, we hit a wall of serious over excursion on the lower end, just insert the numbers for any of the above mentioned drivers into any box plot, and you will see that they start to run out of linear excursion about 125Hz. We need approx. 8-10mm of linear excursion out of a small driver, 3” – 5”, to be able to hit 90Hz. The efficiency of a driver like this would be even lower. So we can see that a small driver is not the answer, not efficient enough and does not have enough linear excursion. The laws of physics are cruel……
So to satisfy the 2 criteria above, we have to move to larger drivers, and if you look at the larger Fostex’s, Jordan’s, Bandor’s, Tagband’s, and the like you can see they can be very efficient and quite comfortably cover the range we want. But….. we have to deal with very poor off-axis response which is not acceptable.
So where does the answer lie? There have been many attempts, and some extremely successful, and I will not run the list here, go to The Single Driver Site and you will see an entire site dedicated to this cause.
Well, here is what I envision, we take a larger FR/WR driver, between 6” –10”, and load it into a sealed box of appropriate size and make up-firing, against a dispersion plate mounted with our tweeter down-firing into the “mid-range” driver. We would then use a large down-firing woofer, 15” –18”, mounted in a sealed enclosure on the bottom firing on to a dispersion plate on the floor.
There it is folks, call me a complete lunatic, but I have been listening to 2 sets of speaker for the last 6 months, a pair of direct-radiators, and a pair of indirect-radiators, and I have come to prefer the indirects. The pair of indirects I am listening to, have not been refined yet, but the sound is cleaner, the sound stage much better, seem to get along with the room better. I cannot find glaring faults with this design yet.
As to crossovers, I would let the mid-range run full range and make whatever tweaks to it to keep all the nasties under control, the tweeter would be brought in on a single cap, while I would do my best to find sub-woofer with the natural response curve that I would only need to use a mechanical low-pass filter, either through enclosure design or dampening of the driver itself, so I can eliminate the power sucking inductor that would normally be needed. If the response of the HF is a concern we could always mount a tweeter in a box on top.
This whole concept is very much in the raw, and I am sure there are plenty of more experienced listeners out there that can tell me why this is not a good idea, so I would like to hear from them.
That is it for now,
Lynn aka Surf, Sun & Sound
One of the parameters I have laid down for these speakers is that they must be efficient enough to produce 100db+ with low-powered amps. I believe the more efficient a driver is the less inherent distortion there is for the whole system of sound reproduction. OK, before you slaughter me, I know this is not that simple, and this part of my theory is just my brainwave.
So let’s think about this abit, firstly, we can use amplifiers that are inherently more linear, such as Class A, whether tube or solid state. Also we do not have to work the voice coil as hard thus reducing heat and distortion caused by over-excursion. I remember the Mr. Nelson Pass saying he would rather be able to produce 100db with 1 watt than 100 watts. So I have this hunch that all things being equal the more efficient a speaker the greater its potential of having less distortion. Just an observation….
One of the laws of speaker design is the smaller the driver, the less efficient. So while drivers such the small Tagband’s, Jordan’s. Bandor’s and such fit the bill quite nicely in terms of response, on and off-axis, the price paid is quite low efficiency, most are under 90db/watt. So for efficiency we have to look at larger drivers, but then we are faced with very poor off-axis response, which is one thing we want as little as possible of.
A side product of the low efficiency of these smaller drivers is to get the spl out of then we want, with any amount of power, we hit a wall of serious over excursion on the lower end, just insert the numbers for any of the above mentioned drivers into any box plot, and you will see that they start to run out of linear excursion about 125Hz. We need approx. 8-10mm of linear excursion out of a small driver, 3” – 5”, to be able to hit 90Hz. The efficiency of a driver like this would be even lower. So we can see that a small driver is not the answer, not efficient enough and does not have enough linear excursion. The laws of physics are cruel……
So to satisfy the 2 criteria above, we have to move to larger drivers, and if you look at the larger Fostex’s, Jordan’s, Bandor’s, Tagband’s, and the like you can see they can be very efficient and quite comfortably cover the range we want. But….. we have to deal with very poor off-axis response which is not acceptable.
So where does the answer lie? There have been many attempts, and some extremely successful, and I will not run the list here, go to The Single Driver Site and you will see an entire site dedicated to this cause.
Well, here is what I envision, we take a larger FR/WR driver, between 6” –10”, and load it into a sealed box of appropriate size and make up-firing, against a dispersion plate mounted with our tweeter down-firing into the “mid-range” driver. We would then use a large down-firing woofer, 15” –18”, mounted in a sealed enclosure on the bottom firing on to a dispersion plate on the floor.
There it is folks, call me a complete lunatic, but I have been listening to 2 sets of speaker for the last 6 months, a pair of direct-radiators, and a pair of indirect-radiators, and I have come to prefer the indirects. The pair of indirects I am listening to, have not been refined yet, but the sound is cleaner, the sound stage much better, seem to get along with the room better. I cannot find glaring faults with this design yet.
As to crossovers, I would let the mid-range run full range and make whatever tweaks to it to keep all the nasties under control, the tweeter would be brought in on a single cap, while I would do my best to find sub-woofer with the natural response curve that I would only need to use a mechanical low-pass filter, either through enclosure design or dampening of the driver itself, so I can eliminate the power sucking inductor that would normally be needed. If the response of the HF is a concern we could always mount a tweeter in a box on top.
This whole concept is very much in the raw, and I am sure there are plenty of more experienced listeners out there that can tell me why this is not a good idea, so I would like to hear from them.
That is it for now,
Lynn aka Surf, Sun & Sound
Re: The Next Installment
The old ominidirectional, huh? There is nothing new under the sun. Or surf. 🙂
When I first became interested in audio back in the seventies, the omnidirectional concept was alredy on the way out, used only by extremely cheap speakers as a gimmick. I read a few descriptions in audio books about it not being good because of "phase problems", but that was it. Perhaps it is due for a revival.
If the highest frequencies you plan for the subwoofer is 70 Hz, won't your subwoofer already be omnidirctional? Why do you need the dispersion plate? Or even a downfiring driver?
The old ominidirectional, huh? There is nothing new under the sun. Or surf. 🙂
When I first became interested in audio back in the seventies, the omnidirectional concept was alredy on the way out, used only by extremely cheap speakers as a gimmick. I read a few descriptions in audio books about it not being good because of "phase problems", but that was it. Perhaps it is due for a revival.
surf said:
If the highest frequencies you plan for the subwoofer is 70 Hz, won't your subwoofer already be omnidirctional? Why do you need the dispersion plate? Or even a downfiring driver?
Coming in late to this thread I though I might add my own comments regarding the general design objectives spelled out originally (hope this is still a topic of interest). The first post indicated that we want to avoid crossovers in the 90 Hz - 4-6khz range.
What follows is my own opinion, and I respect the fact that different people hear and experience music differently. I normally refrain from discussing my personal preferences, but I will "risk it" this time.
My own experience with regards to phase distortion effects is remarkably consistent with the stated design objectives. Being an engineer I performed some tests (listening) to gain a better understanding of the issue. I found that a Linkwitz-Riley 4th order crossover introduces unacceptable phase distortion if it is placed within the 125hz-4khz frequency range. Better would be 100hz at the bottom end and 8khz at the high end. A 4khz crossover would be "acceptable" according to my tests, but reducing the crossover frequency below this VERY QUICKLY destroys my enjoyment of the music. I like electrostatic speakers, and I believe this is part of the reason (absence for crossovers in the desired range).
From a technical standpoint it would seem from the above observations that some people are extremely sensitive to relative time shifts in "sine waves" in the 1khz-3kz region. A performance metric of interest would be the second derivative of the phase distortion within this region ("curvature" in the phase response). A crossover at 200 hz, for example, still has significant "phase distortion curvature" at 1khz. As the crossover frequency rises above 3khz the "phase distortion curvature" in the critical region diminishes quickly. This is all theory, but consistent with my observations.
Note: some people enjoy the sound of live human voice to the extent that a state of ecstasy is achieved. Others seem to not respond in this way at all, as I observe when at a live event observing the response of others (and discussion afterward). I find myself in the first group, and also find that such an emotional response to the listening experience can be achieved from recorded music only if phase distortion is low as stated above, with the added requirement that the ultrasonic signature is also left intact via lack of ultrasonic resonance (e.g., dome tweeters) and preservation of ultrasonic content via SACD or 24/96 digital. A good LP (a big black disc) played back on a good system can also be quite good. Again, different people hear and experience such things differently.
These observations have lead me to designs that meet your original objectives. I am using electrostatics in my own ultimate design, but a good wide-band midbass (Jordan JX92s, for example) combined with a decent ribbon (Bohlender-Graebener Neo3-PDR w/o faceplate) may work if a 4th-order L-R crossover is employed at about 6-8 khz. In such a design I might consider an IB (infinite Box) implementation for the midbass (not infinite baffle). The IB concept was described in the audioXpress magazine (early 2002, I believe). Many other designs and drivers are also compatible with the requirements.
What follows is my own opinion, and I respect the fact that different people hear and experience music differently. I normally refrain from discussing my personal preferences, but I will "risk it" this time.
My own experience with regards to phase distortion effects is remarkably consistent with the stated design objectives. Being an engineer I performed some tests (listening) to gain a better understanding of the issue. I found that a Linkwitz-Riley 4th order crossover introduces unacceptable phase distortion if it is placed within the 125hz-4khz frequency range. Better would be 100hz at the bottom end and 8khz at the high end. A 4khz crossover would be "acceptable" according to my tests, but reducing the crossover frequency below this VERY QUICKLY destroys my enjoyment of the music. I like electrostatic speakers, and I believe this is part of the reason (absence for crossovers in the desired range).
From a technical standpoint it would seem from the above observations that some people are extremely sensitive to relative time shifts in "sine waves" in the 1khz-3kz region. A performance metric of interest would be the second derivative of the phase distortion within this region ("curvature" in the phase response). A crossover at 200 hz, for example, still has significant "phase distortion curvature" at 1khz. As the crossover frequency rises above 3khz the "phase distortion curvature" in the critical region diminishes quickly. This is all theory, but consistent with my observations.
Note: some people enjoy the sound of live human voice to the extent that a state of ecstasy is achieved. Others seem to not respond in this way at all, as I observe when at a live event observing the response of others (and discussion afterward). I find myself in the first group, and also find that such an emotional response to the listening experience can be achieved from recorded music only if phase distortion is low as stated above, with the added requirement that the ultrasonic signature is also left intact via lack of ultrasonic resonance (e.g., dome tweeters) and preservation of ultrasonic content via SACD or 24/96 digital. A good LP (a big black disc) played back on a good system can also be quite good. Again, different people hear and experience such things differently.
These observations have lead me to designs that meet your original objectives. I am using electrostatics in my own ultimate design, but a good wide-band midbass (Jordan JX92s, for example) combined with a decent ribbon (Bohlender-Graebener Neo3-PDR w/o faceplate) may work if a 4th-order L-R crossover is employed at about 6-8 khz. In such a design I might consider an IB (infinite Box) implementation for the midbass (not infinite baffle). The IB concept was described in the audioXpress magazine (early 2002, I believe). Many other designs and drivers are also compatible with the requirements.
goudey said:
IIRC, IB is when the air (inside box) compliance is 3x (or more) the cone suspension compliance (I was reading some material this weekend).
goudey said:
And is just a variation of Hartley's Boffle system (that is boffle not baffle) he published in the 50s.
dave
Mr. Surf, etc
There IS a high end super efficient system with horizontal mid/woof pointed up and a tweet also vertical .The woofer sound is dispersed by firing at a cone. I'm on the trial trying to find it in my links
IF the basket were very minimal , which is getting more common with lots of drivers, I would think that the larger speaker could be mounted pointing down into a box and the back of the cone radiates into the room. Would look like a Walsh driver (but not work like one) A Fertin driver could be remade with braces further away since they just bolt on. (too pricy for me though)
I like this tack you are taking. Note that it doesn't have to be necessarily omnidirectional., Part of the driver could be surrounded by absorbent stuff, like felt or could be enclosed in a highly stuffed box. I can't believe you might have come up with a scheme that might just fulfill your requirements!! Way to go!
How about this:
First Remove the magnet assembly from a fullrange driver, by cutting the legs that attach it to the rim. So you now have a magnet assembly with no basket.
Epoxy a cylinder onto the center pole piece to extend it about 6"
Now cut the legs off the rim. So now you also have a rim with a cone, suspension, and voice coil.Now remove the dustcap.
Ok, Bolt the rim to an opening in the top of the box (box stuffed) so the cone sticks up like a volcano. Take the magnetic assembly and place the extended pole piece through the cone voicecoil from above. The extended pole piece rests on a bridge that I forgot to mention that goes across the box opening under the cone.
So there you have it: An "inside out" driver with no basket in the way of the outer surface. The outer surface radiates omnidirectionall unless surrounded as mentioned above.
Don't go telling me that it will be impossible to center the voice coil. People do it all the time to recone woofers. You just need some paper shims to center it. Also to simplify the description, I didn't mention how you have to perserve the leads to the voice coil or at least re-attach them after surgery.
There IS a high end super efficient system with horizontal mid/woof pointed up and a tweet also vertical .The woofer sound is dispersed by firing at a cone. I'm on the trial trying to find it in my links
IF the basket were very minimal , which is getting more common with lots of drivers, I would think that the larger speaker could be mounted pointing down into a box and the back of the cone radiates into the room. Would look like a Walsh driver (but not work like one) A Fertin driver could be remade with braces further away since they just bolt on. (too pricy for me though)
I like this tack you are taking. Note that it doesn't have to be necessarily omnidirectional., Part of the driver could be surrounded by absorbent stuff, like felt or could be enclosed in a highly stuffed box. I can't believe you might have come up with a scheme that might just fulfill your requirements!! Way to go!
How about this:
First Remove the magnet assembly from a fullrange driver, by cutting the legs that attach it to the rim. So you now have a magnet assembly with no basket.
Epoxy a cylinder onto the center pole piece to extend it about 6"
Now cut the legs off the rim. So now you also have a rim with a cone, suspension, and voice coil.Now remove the dustcap.
Ok, Bolt the rim to an opening in the top of the box (box stuffed) so the cone sticks up like a volcano. Take the magnetic assembly and place the extended pole piece through the cone voicecoil from above. The extended pole piece rests on a bridge that I forgot to mention that goes across the box opening under the cone.
So there you have it: An "inside out" driver with no basket in the way of the outer surface. The outer surface radiates omnidirectionall unless surrounded as mentioned above.
Don't go telling me that it will be impossible to center the voice coil. People do it all the time to recone woofers. You just need some paper shims to center it. Also to simplify the description, I didn't mention how you have to perserve the leads to the voice coil or at least re-attach them after surgery.
Here is the link to the speakers that are mounted vertically.
Man this web is sure something!!!
http://www.cd-konzert.de/english/
Man this web is sure something!!!
http://www.cd-konzert.de/english/
A review:
http://www.cd-konzert.de/test/Etests.htm
Also this brings to mind my idea of having a dome tweeter facing up, with another one facing down on top of it, separated with spacers so the domes almost touch at full excursion. The sound is sqeezed out the side 36o deg-maybe more effortlessly and faster a la heil tweeeter? Maybe the faceplates need to be changed a bit to create a horn?
http://www.cd-konzert.de/test/Etests.htm
Also this brings to mind my idea of having a dome tweeter facing up, with another one facing down on top of it, separated with spacers so the domes almost touch at full excursion. The sound is sqeezed out the side 36o deg-maybe more effortlessly and faster a la heil tweeeter? Maybe the faceplates need to be changed a bit to create a horn?
Variac said:
That's not a bad idea at all. I'll have to see if i have 4 suitable tweeters to try that idea out.
dave
Goudey, Jag, Planet 10-all you guys with knowledge of the Infinite Box, Hartley Boffle, etc.
The author of this thread is looking for some advice on them.
http://www.diyaudio.com/forums/showthread.php?threadid=7860
Is that the same Hartley who came up with the oscillator configuration?
The author of this thread is looking for some advice on them.
http://www.diyaudio.com/forums/showthread.php?threadid=7860
Is that the same Hartley who came up with the oscillator configuration?