@ Bill Brown, ... "like trying to hold 10 ping pong balls under water"
...i like that !!
Or like trying to herd 10 cats
The more i think about it, i figure countless folks have pondered this one...how to tight pack drivers.
It may be what led to synergy's existence, conical "horn-rings," etc...
in that what can't be geometrically spaced tight enough on a flat plane, can be done on a conical plane...
Just scratching my head here....
@nc535,
Good to see you you hear Jack
I just can't keep up with the sims and such...
but i would honestly try if i wasn't so dang suspicious as to how they equate to reality.
I mean, the SG26 tweeter measurements i've made vary so incredibly much just due to baffling and diffraction.
Tests being:
flat textbook baffle SG26 alone (beautiful),
vs a flush mount baffle SG26 with surrounding 3 fe25s also flush mounted (half-***acceptable),
vs none of them flush mounted (ugly).
Can sims handle these cases?
I guess maybe so? if i make polars of all the cases to put into Vcad?
But if i do that much work why bother with sims to begin with??
Pls don't think I'm knocking sims, and i certainly don't want to shift the focus in this thread away from the great discussion on driver spacing and ring array combining....
Simply want to know if sims can really do it... ???
...i like that !!
Or like trying to herd 10 cats
The more i think about it, i figure countless folks have pondered this one...how to tight pack drivers.
It may be what led to synergy's existence, conical "horn-rings," etc...
in that what can't be geometrically spaced tight enough on a flat plane, can be done on a conical plane...
Just scratching my head here....
@nc535,
Good to see you you hear Jack
I just can't keep up with the sims and such...
but i would honestly try if i wasn't so dang suspicious as to how they equate to reality.
I mean, the SG26 tweeter measurements i've made vary so incredibly much just due to baffling and diffraction.
Tests being:
flat textbook baffle SG26 alone (beautiful),
vs a flush mount baffle SG26 with surrounding 3 fe25s also flush mounted (half-***acceptable),
vs none of them flush mounted (ugly).
Can sims handle these cases?
I guess maybe so? if i make polars of all the cases to put into Vcad?
But if i do that much work why bother with sims to begin with??
Pls don't think I'm knocking sims, and i certainly don't want to shift the focus in this thread away from the great discussion on driver spacing and ring array combining....
Simply want to know if sims can really do it... ???
Hi Mark:
Its good although somewhat surprising to see you at it again after your great results with your last project. (And I agree - a large horn like that is possibly the best way to go if you have room for it.) That was always my problem. As soon as I had a design done, I would lose interest, set it aside and go one to the next. The bar (and the cost) for that is pretty low for that when I'm just simulating.
Vituix would be very good at answering questions like:
-do I have the drivers close enough to each other?
-how loud will it be able to play?
-where are the boundary peaks and nulls?
-what is the overall directivity?
-are there any lobes in the pattern?
you would/could make prototype XO and EQ filters based on the sim but of course the final ones would be based on measurements
Accuracy of the sims depends primarily on the accuracy of the measurements. If you have 5 or 10 degree polars of each driver type in a test box it will be very good. What Vituix does is add up the contributions of the individual drivers, taking their directivity and relative distances into account, but not necessarily the baffle as each driver in an array may have a different distance to baffle edge than it did in the test box. If you trace data sheet graphs and synthesize directivity based on piston diameter using Vituix SPL trace and diffraction tools, it will give a good general idea of what the response will look like - one much better than one based on Edge or Xdir.
Without simulation, there is a lot of cut and try and sometimes even a re-start. My goal is to do all the what ifs and cut and try in simulation and to try many more designs than I could if I had to build each one in order to evaluate it.
Its good although somewhat surprising to see you at it again after your great results with your last project. (And I agree - a large horn like that is possibly the best way to go if you have room for it.) That was always my problem. As soon as I had a design done, I would lose interest, set it aside and go one to the next. The bar (and the cost) for that is pretty low for that when I'm just simulating.
Vituix would be very good at answering questions like:
-do I have the drivers close enough to each other?
-how loud will it be able to play?
-where are the boundary peaks and nulls?
-what is the overall directivity?
-are there any lobes in the pattern?
you would/could make prototype XO and EQ filters based on the sim but of course the final ones would be based on measurements
Accuracy of the sims depends primarily on the accuracy of the measurements. If you have 5 or 10 degree polars of each driver type in a test box it will be very good. What Vituix does is add up the contributions of the individual drivers, taking their directivity and relative distances into account, but not necessarily the baffle as each driver in an array may have a different distance to baffle edge than it did in the test box. If you trace data sheet graphs and synthesize directivity based on piston diameter using Vituix SPL trace and diffraction tools, it will give a good general idea of what the response will look like - one much better than one based on Edge or Xdir.
Without simulation, there is a lot of cut and try and sometimes even a re-start. My goal is to do all the what ifs and cut and try in simulation and to try many more designs than I could if I had to build each one in order to evaluate it.
Last edited:
The simulation can only be as good as the data you feed into it and the assumptions being made in the simulation.
Vituix is not ideal to deal with the sort of issue you are describing there with diffraction from surrounding driver cones and non flush mounting.
Those are functions of geometry and can be simulated quite successfully with BEM and FEM but those are not quite as quick as the sort of things you can do in Vituix.
Given how fast you can make a prototype and measure it in reality would impact on the benefit you would get from simulating. Where it can be useful as Jack says is in finding issues that prevent you from wanting to make a prototype or trying out different ideas quickly to see which has the best chance of giving the response you want.
Thanks nc535 and fluid ,
Your comments make full sense, on how simulations can fit, as well as not fit this project.
The diffraction thing going on kinda has my head spinning....I really don't know how I could make the baffle any smoother, the way it fits the tweeter and surrounding 3fe25's.
Maybe a "tight-pack" ring is a non-starter simple due to inevitable diffraction from adjacent drivers too closely spaced?
Ken, looking at your latest design, as well as FoLLgoTT's, driver spacing looks moved out a fair amount from the minimum possible. The vistaton waveguide does the same thing i think, moving spacing out.
Maybe the separation is necessary for reducing diffraction????
Just spitballing, wondering out loud....
PS, Jack it is kinda crazy i'm working on another project given how happy i am with the large horn synergies. They continue to thrill me. I've had great clarity, dynamics, imaging, and SPL for quite a while in previous designs, but now i have to add a newfound appreciation for beautiful timbre nuances.
But i have to admit, the prototype i deemed too heavy to continue with that led to the current synergies, had a little extra bit of magic that the current synergies don't quite have.
The prototype for current synergies was an equally large horn, very heavy, narrower pattern, that had a little less center-to-center distance between the mid entry ports, and CD to ports.
I'm thinking the smaller c-to-c port distances was maybe the major source of the magic. Maybe plain ole closer acoustic coupling....
It's why I'm fascinated with the ring arrays, their potentially tight c-to-c's, and pattern control that might match the big horns.
(But the magic could have been from the heavy proto's tighter pattern, larger mid drivers with way stronger BL, or even different secondary flare materials..... simply dunno...
Your comments make full sense, on how simulations can fit, as well as not fit this project.
The diffraction thing going on kinda has my head spinning....I really don't know how I could make the baffle any smoother, the way it fits the tweeter and surrounding 3fe25's.
Maybe a "tight-pack" ring is a non-starter simple due to inevitable diffraction from adjacent drivers too closely spaced?
Ken, looking at your latest design, as well as FoLLgoTT's, driver spacing looks moved out a fair amount from the minimum possible. The vistaton waveguide does the same thing i think, moving spacing out.
Maybe the separation is necessary for reducing diffraction????
Just spitballing, wondering out loud....
PS, Jack it is kinda crazy i'm working on another project given how happy i am with the large horn synergies. They continue to thrill me. I've had great clarity, dynamics, imaging, and SPL for quite a while in previous designs, but now i have to add a newfound appreciation for beautiful timbre nuances.
But i have to admit, the prototype i deemed too heavy to continue with that led to the current synergies, had a little extra bit of magic that the current synergies don't quite have.
The prototype for current synergies was an equally large horn, very heavy, narrower pattern, that had a little less center-to-center distance between the mid entry ports, and CD to ports.
I'm thinking the smaller c-to-c port distances was maybe the major source of the magic. Maybe plain ole closer acoustic coupling....
It's why I'm fascinated with the ring arrays, their potentially tight c-to-c's, and pattern control that might match the big horns.
(But the magic could have been from the heavy proto's tighter pattern, larger mid drivers with way stronger BL, or even different secondary flare materials..... simply dunno...
Hi Mark:
I think these tight packed rings do have potential but you do need to do what you can to minimize diffraction. A small, shallow waveguide instead of a flat mounting plate for the central tweeter is probably what is called for. Diffraction for drivers on the outer rings shouldn't matter given the dimensions and the frequencies at which they play.
I think these tight packed rings do have potential but you do need to do what you can to minimize diffraction. A small, shallow waveguide instead of a flat mounting plate for the central tweeter is probably what is called for. Diffraction for drivers on the outer rings shouldn't matter given the dimensions and the frequencies at which they play.
What if at least the first ring was behind baffle firing through band pass ports (like in a MEH) to get the c-2-c a bit closer and a bit less diffraction? Maybe try to attach them ring drivers a bit rotated instead of flat against the baffle, like if a MEH was in a car crash and went flat?
I remember somewhere someone wrote that MEHs are increasingly difficult to implement for wider than about 60 degrees (straight sided horn) since the mid range taps get further and further apart as one widens the horn so need to get ever closer the throat, or something like that. If you fellows succeed here with this 180 degree "horn" that myth is busted Jolly weekend!
I remember somewhere someone wrote that MEHs are increasingly difficult to implement for wider than about 60 degrees (straight sided horn) since the mid range taps get further and further apart as one widens the horn so need to get ever closer the throat, or something like that. If you fellows succeed here with this 180 degree "horn" that myth is busted
Jolly weekend!
Last edited:
Choice of WG
Good points.
That's part of the reasoning that led me to choose the particular WG I use, its edge is shaded from the cone. I keep a table of thresholds for audibility of delayed reflections at hand while evaluating impulse responses. The edge of the WG can be of concern.
Ken
Good points.
That's part of the reasoning that led me to choose the particular WG I use, its edge is shaded from the cone. I keep a table of thresholds for audibility of delayed reflections at hand while evaluating impulse responses. The edge of the WG can be of concern.
Ken
See above
I refer to that idea above. A significant difference between that and a true MEH, is that it has less tolerance to minor differences in the Helmholtz resonances at the ports. On the other hand, the coupling down the WG is weaker, and there's much less of a cancellation dip.
A solution with CNC or 3-D printing in mind would likely lead to having "tilted" mid drives just as you suggest, to give the shortest ports. Such a thing is way beyond my abilities to make.
The first array I made had the inner mids behind ports - and the differences among the port resonances was significant, and annoying (at least for measuring).
Ken
I refer to that idea above. A significant difference between that and a true MEH, is that it has less tolerance to minor differences in the Helmholtz resonances at the ports. On the other hand, the coupling down the WG is weaker, and there's much less of a cancellation dip.
A solution with CNC or 3-D printing in mind would likely lead to having "tilted" mid drives just as you suggest, to give the shortest ports. Such a thing is way beyond my abilities to make.
The first array I made had the inner mids behind ports - and the differences among the port resonances was significant, and annoying (at least for measuring).
Ken
Hence 10 years!
On the other hand, even highly imperfect attempts like those I've made, produce an attractive sound, at least in some sense. Until I tried such wide and shallow speakers, I hadn't appreciated the significance of the rear-wall reflection.
OTOH my recent experience suggests to me that ceiling reflections are not as disturbing as I had previously thought (comparing the sound of the arrays described up thread).
The main lesson for me is that I won't ever consider speakers with narrow baffles. Strange that nearly all speakers are narrower than they are deep, that now seems misguided.
Ken
Yeah it looks like every trick known to diyaudio folks must be exploited in order to make this work
On the other hand, even highly imperfect attempts like those I've made, produce an attractive sound, at least in some sense. Until I tried such wide and shallow speakers, I hadn't appreciated the significance of the rear-wall reflection.
OTOH my recent experience suggests to me that ceiling reflections are not as disturbing as I had previously thought (comparing the sound of the arrays described up thread).
The main lesson for me is that I won't ever consider speakers with narrow baffles. Strange that nearly all speakers are narrower than they are deep, that now seems misguided.
Ken
Should have replied here
Yes, but if you are EQing a whole ring, they'd better be reasonably well matched.
It's not like inside an MEH where differences in the response of individual mids would average out as the sources combine.
Ken
Yes, but if you are EQing a whole ring, they'd better be reasonably well matched.
It's not like inside an MEH where differences in the response of individual mids would average out as the sources combine.
Ken
For an on wall speaker wide and flat makes a lot of sense, for other placements it might not be quite as clear cut. I haven't read the whole thread so forgive me if it is covered, but what makes you say the above quote?
perhaps a longer answer than you wanted
Fluid,
good question, the reasoning could be specific to my situation where on-wall/near wall works well.
I managed OK with speakers far from the rear wall firing along the long dimension of my room (which has a 2:1 aspect ratio, approximately. That worked well enough with strongly toed-in Geddes Harpers (supported by added woofers & subwoofers). The speakers were ~2m apart, and I could listen in a zone about 2m in front. I could never achieve as good a sound when I tried to arrange the Harpers facing across the width of the room – at that time it did not occur to me that most of the problem was from the rear wall (which I now think it might have been).
Many years ago ('90s) I had a chance to listen to Quad 63s in a large room and liked the sound. Since then I've been wondering about large-panel speakers, but without having a large enough room for "big dipoles" that was going nowhere. Since then I've explored many ideas even vestigial baffles (including ambiphonic arrays of tweeters and mids on sticks in an arc around the room - which worked well with some recordings and when I could hold my head still in the right place).
Having read what BButterfield wrote in another forum, under a different pseudonym, I was intrigued to revisit the ideas mentioned in my OP. The resulting arrays now fire across the room, are farther apart, but still far from the side walls, and don't require to be toed-in (the first ones were wedged but the last ones are not).
The first two arrays were designed to reduce floor and ceiling reflections too, but I can't find evidence correlating what I measure and what I hear that this aspect makes much difference to me. The effect seems to be very different from side-wall reflections, or perhaps it’s because the times delay between floor, ceiling and side-wall reflections are now such that the brain can "sort them out" - it just came to mind that this aspect may also have been important when the Harpers were toed in to strongly reduce the reflection from the nearer side-wall.
I'm reluctant to enter into descriptions of the “imaging” impressions but the main reason for my statement is that I perceive a natural quality to the sound of the arrays, perhaps the same as referred to by people who had heard BButterfield's arrays (again on the other forum, the link to which I have misplaced).
For years I’ve looked for a solution better than I could achieve with Harpers when facing across the room. It’s hard to find speakers to replace Geddes ones (except perhaps larger Geddes ones were they obtainable).
It’s during that phase that I started to doubt the sense of narrow baffles, for various reasons (waveguide size, step frequency range, round-overs). Safe to say that I didn’t find a commercial solution that I could afford (nor even many that I can’t afford). A couple of ideas that looked attractive along the way are Grimm LS1 (which I have been tempted to imitate) and Taipuu 2 way (which given the cost would mean putting a large BMS co-ax in a larger box) – but I think they are still too narrow (17.1”/434mm so 1.7x as wide as Harpers and WG holding down no more than an extra octave).
If I ever move to a new house, I'll be looking at whether I can use arrays or fit in-wall speakers, not placing speakers away from walls.
Ken
Fluid,
good question, the reasoning could be specific to my situation where on-wall/near wall works well.
I managed OK with speakers far from the rear wall firing along the long dimension of my room (which has a 2:1 aspect ratio, approximately. That worked well enough with strongly toed-in Geddes Harpers (supported by added woofers & subwoofers). The speakers were ~2m apart, and I could listen in a zone about 2m in front. I could never achieve as good a sound when I tried to arrange the Harpers facing across the width of the room – at that time it did not occur to me that most of the problem was from the rear wall (which I now think it might have been).
Many years ago ('90s) I had a chance to listen to Quad 63s in a large room and liked the sound. Since then I've been wondering about large-panel speakers, but without having a large enough room for "big dipoles" that was going nowhere. Since then I've explored many ideas even vestigial baffles (including ambiphonic arrays of tweeters and mids on sticks in an arc around the room - which worked well with some recordings and when I could hold my head still in the right place).
Having read what BButterfield wrote in another forum, under a different pseudonym, I was intrigued to revisit the ideas mentioned in my OP. The resulting arrays now fire across the room, are farther apart, but still far from the side walls, and don't require to be toed-in (the first ones were wedged but the last ones are not).
The first two arrays were designed to reduce floor and ceiling reflections too, but I can't find evidence correlating what I measure and what I hear that this aspect makes much difference to me. The effect seems to be very different from side-wall reflections, or perhaps it’s because the times delay between floor, ceiling and side-wall reflections are now such that the brain can "sort them out" - it just came to mind that this aspect may also have been important when the Harpers were toed in to strongly reduce the reflection from the nearer side-wall.
I'm reluctant to enter into descriptions of the “imaging” impressions but the main reason for my statement is that I perceive a natural quality to the sound of the arrays, perhaps the same as referred to by people who had heard BButterfield's arrays (again on the other forum, the link to which I have misplaced).
For years I’ve looked for a solution better than I could achieve with Harpers when facing across the room. It’s hard to find speakers to replace Geddes ones (except perhaps larger Geddes ones were they obtainable).
It’s during that phase that I started to doubt the sense of narrow baffles, for various reasons (waveguide size, step frequency range, round-overs). Safe to say that I didn’t find a commercial solution that I could afford (nor even many that I can’t afford). A couple of ideas that looked attractive along the way are Grimm LS1 (which I have been tempted to imitate) and Taipuu 2 way (which given the cost would mean putting a large BMS co-ax in a larger box) – but I think they are still too narrow (17.1”/434mm so 1.7x as wide as Harpers and WG holding down no more than an extra octave).
If I ever move to a new house, I'll be looking at whether I can use arrays or fit in-wall speakers, not placing speakers away from walls.
Ken
Thanks Jack!
I'll try a Wavecore TW030WA14 i have laying around next, which i think fits what you're recommending.
One thing i can't picture intuitively though, is why there would be a diffraction difference between a flat mounting plate surrounding the central tweeter, vs a flat baffle surrounding the tweeter. Seems in either case, a flat surround is a flat surround.
I guess you're saying the shape and depth of the waveguide are the important parts of holding diffraction down....??
I've been thinking I need some kind of waveguide around the tweeter to get the pattern control handshake started right with the first ring .
What if at least the first ring was behind baffle firing through band pass ports (like in a MEH) to get the c-2-c a bit closer and a bit less diffraction? Maybe try to attach them ring drivers a bit rotated instead of flat against the baffle, like if a MEH was in a car crash and went flat?
I remember somewhere someone wrote that MEHs are increasingly difficult to implement for wider than about 60 degrees (straight sided horn) since the mid range taps get further and further apart as one widens the horn so need to get ever closer the throat, or something like that. If you fellows succeed here with this 180 degree "horn" that myth is busted
Hi tmuikki,
Yep, i think Ken has mentioned ports behind the baffle, and talked about the increased precision needed for such (as compared to a synergy where the horn does some healing blending)
And i think you're correct about mid-range port spacing issues in synergy's wider than 60 degrees.
My latest 90x60 syn has the midrange ports on the horn's top and bottom walls to take advantage of the closer 60 deg spacing, compared to the side walls 90 deg spacing.
Hi Ken, that brings up the question I've been meaning to ask you for some time...had you considered in-wall speakers.
I really think this is probably the way to ultimately go, and what I'll do if i ever build a room again.
What little i know of high $$$$ studio installations, it appears such is the strongly favored install technique.
A fwiw re planars.... my experiences with full range electrostats have been that they almost always work better in small rooms than large.
Especially small rooms with stiff wall and ceiling construction.
No need for a sub then, and gives beautiful coherency top to bottom.
Interesting to see you two thinking similarly to me. I can't understand the narrow baffle approach.
Especially in DIY, even though there are some good simulation tools to help with diffraction/radiation pattern. My thinking has always been an "infinite baffle" measuring large enough to go down to perhaps the Schroeder frequency of the room, where it seems to begin to matter less or rebuilding my front wall (sheetrock over studs) and going in-wall.
I read somewhere else, after someone reported lesser imaging with these approaches, that it was an optical illusion, the soundstaging fantastic if the eye-ear-brain wasn't being misled.
To eliminate the cancellation based rear wall reflections and the baffle-edge diffraction issue seems very worthwhile to me.
Bill
Especially in DIY, even though there are some good simulation tools to help with diffraction/radiation pattern. My thinking has always been an "infinite baffle" measuring large enough to go down to perhaps the Schroeder frequency of the room, where it seems to begin to matter less or rebuilding my front wall (sheetrock over studs) and going in-wall.
I read somewhere else, after someone reported lesser imaging with these approaches, that it was an optical illusion, the soundstaging fantastic if the eye-ear-brain wasn't being misled.
To eliminate the cancellation based rear wall reflections and the baffle-edge diffraction issue seems very worthwhile to me.
Bill
Thanks for the reasoning, your placement situation no doubt factors into your preferences. Picking a poison for everyone is not easy
My next speaker will be somewhat similar to a Summa or JBL M2 in size. I've posted some directivity simulations of it this thread 2 way waveguide speaker build ABEC modelling
Woofer simulations are currently in the works as they have turned out to be trickier than imagined.
BTW bbutterfield has posted info on his speaker to diyaudio recently also interesting that is considering a WWMTMWW version also wide and flat.
First build: Slim wall speakers. Some startup questions
The waveguide helps to guide the sound where you want it and avoid it spreading towards the midranges. Whatever is on the front of the wavecor tweeter seems to be doing almost nothing as the directivity looks just like you would imagine a 1.3" tweeter to look.
If you are able to measure the dome and surround I could see what sort of guide I could come up with for you.
Might also be a candidate for a Tekton style array.