This is the latest in an ongoing journey into new territory for me. Previous tests around the near-field theme have arrived at various (somewhat personal) conclusions:
The circles are based on (existing drivers) an alpair 5.3 and a scanspeak 15w/8530k01, which can work rather well together IMO. The mock up comprises Ikea blanda bowls in 28cm and 20cm diameters; the height is somewhat less than two stacked spheres would be, and I calculate will offer about 12Litres internal volume (before internal bracing). Probably I'd go with sealed enclosures, partly for simplicity under the circumstances and partly because I'm happy enough with how the drivers sound in them.
The alpair would want a bit less than 2Litres for a Qtc of 0.7, which would leave the 15W with sufficient for a qtc of perhaps 0.75 (depending on stuffing and braces and how the internal volume is apportioned and divided into two chambers). This will be an an active system so could be corrected in any case, but I'd just like to begin from 'approximately' suitable volumes for future flexibility.
So that is the new embryo of a plan. I don't get a lot of time off and have no workshop these days so it may take a while to realise. Combined with the complicated shape, I'll probably therefore make use of the bowls and go for a painted finish, which will allow much use of glue and filler underneath. I had considered making a mould and then using glass-fiber and resin for the walls, but that all adds to time and cost. If this works then I can always use a finished speaker as a mould later, should I want to make a fiber-glass version in the future.
Cheers,
Kev
- My situation is indeed well suited to speakers of modest SPL positioned quite close to the listening position (say around 50cm give or take). I like this, where room effects are reduced and the neighbours don't complain as much, and yet the apparent location of instruments and vocals (etc) seem more natural to me than with headphones.
- My first attempts with full-range drivers were very positive indeed; I'm surprised by how much I like them. However (at least with a small enclosure) it transpires that I prefer them used in a WAW configuration.
- Speakers positioned so close can feel intrusive and also be a bit annoying for practical reasons. I've found this to be greatly improved if they are a rounded shape (rather than with corners and points), and also if they're not taking up desk-space or blocking lines of sight.
- It is to be a bespoke creation for a slightly unusual setup, so I'd like something that looks a bit unusual too.
The circles are based on (existing drivers) an alpair 5.3 and a scanspeak 15w/8530k01, which can work rather well together IMO. The mock up comprises Ikea blanda bowls in 28cm and 20cm diameters; the height is somewhat less than two stacked spheres would be, and I calculate will offer about 12Litres internal volume (before internal bracing). Probably I'd go with sealed enclosures, partly for simplicity under the circumstances and partly because I'm happy enough with how the drivers sound in them.
The alpair would want a bit less than 2Litres for a Qtc of 0.7, which would leave the 15W with sufficient for a qtc of perhaps 0.75 (depending on stuffing and braces and how the internal volume is apportioned and divided into two chambers). This will be an an active system so could be corrected in any case, but I'd just like to begin from 'approximately' suitable volumes for future flexibility.
So that is the new embryo of a plan. I don't get a lot of time off and have no workshop these days so it may take a while to realise. Combined with the complicated shape, I'll probably therefore make use of the bowls and go for a painted finish, which will allow much use of glue and filler underneath. I had considered making a mould and then using glass-fiber and resin for the walls, but that all adds to time and cost. If this works then I can always use a finished speaker as a mould later, should I want to make a fiber-glass version in the future.
Cheers,
Kev
Last edited:
Got a bit of time to play with the idea some more today. Nothing very profound but thought I'd post it as a record of how the idea is starting to take shape.
I've decided to use the bowls just for the overall shape, and not to dictate how and where this is divided into mid-tweeter and woofer compartments (as conjoining spheres would try to do). It gives more flexibility on where the drivers can be placed and on how the internal space can be apportioned between them. It also helps avoid having the same internal dimension in all directions.
So really I just want an outer shell to start with. A rebate in the bowls should allow some (slightly tapered) slats to be epoxied all around. They'll create a many-sided cone as an outer shell, epoxy filler will later help make that fully round. Just four slats are shown (and of course only half of the speaker), to illustrate:
The slats are quite thin, and mostly intended to make a reasonable curve. For more structural purposes there would be additional ply slats laminated behind these, and internal bracing added around where they join the bowls.
I've decided to use the bowls just for the overall shape, and not to dictate how and where this is divided into mid-tweeter and woofer compartments (as conjoining spheres would try to do). It gives more flexibility on where the drivers can be placed and on how the internal space can be apportioned between them. It also helps avoid having the same internal dimension in all directions.
So really I just want an outer shell to start with. A rebate in the bowls should allow some (slightly tapered) slats to be epoxied all around. They'll create a many-sided cone as an outer shell, epoxy filler will later help make that fully round. Just four slats are shown (and of course only half of the speaker), to illustrate:
The slats are quite thin, and mostly intended to make a reasonable curve. For more structural purposes there would be additional ply slats laminated behind these, and internal bracing added around where they join the bowls.
Last edited:
Thinking further, these speakers will have probably the widest choice of possible crossover frequency range that I've personally ever built. The 15W is sensible up to a few khz, whilst the 5.3 has been successfully used right down to around 250hz. Though of course I'm intending to use the 5.3 as a mid-tweeter, or I'd have used an actual tweeter, so probably that gives at least some kind of general range to work to.
With the intention to have the speakers so close, a small change in listening height is actually quite a large change in listening angle. So I'd prefer the driver separation to equate to less than half the wavelength at crossover frequency, rather than the often used 1:1. At about 130mm centre-to-centre, that would still allow a XO of up to 1.3khz, so hardly a restriction in this case. In fact even Planet10's frequently suggested 1/4 wavelength would mean a 660hz XO, which is trivial for the 5.3. Interesting; the easy use of small drivers might be another benefit of close listening that I've not experienced with typical room speakers before.
So probably I shall start testing somewhere around the 660hz point, and see how it sounds. The 15w 'bass' driver in this case is small enough that there isn't really much need for the mid-tweeter to go especially low, and the 15w will certainly have more 'umph' down there. Though conversely maybe even an active crossover might be best kept below about 300hz, especially if I decide to stretch the 15w with any low-end dsp. I shall have to see; happily the XO will be active and just software so no cost implications in trying alternatives.
With the intention to have the speakers so close, a small change in listening height is actually quite a large change in listening angle. So I'd prefer the driver separation to equate to less than half the wavelength at crossover frequency, rather than the often used 1:1. At about 130mm centre-to-centre, that would still allow a XO of up to 1.3khz, so hardly a restriction in this case. In fact even Planet10's frequently suggested 1/4 wavelength would mean a 660hz XO, which is trivial for the 5.3. Interesting; the easy use of small drivers might be another benefit of close listening that I've not experienced with typical room speakers before.
So probably I shall start testing somewhere around the 660hz point, and see how it sounds. The 15w 'bass' driver in this case is small enough that there isn't really much need for the mid-tweeter to go especially low, and the 15w will certainly have more 'umph' down there. Though conversely maybe even an active crossover might be best kept below about 300hz, especially if I decide to stretch the 15w with any low-end dsp. I shall have to see; happily the XO will be active and just software so no cost implications in trying alternatives.
Last edited:
Ideally the centre-to-centre is less than a quarter wavelength, keeping the XO below that is always one of my goals (not always achievable). #50 Hz is just over 24cm.
dave
Thanks, yes I've seen your posts on the 1/4 wavelength separation and believe that I understand it to at least some extent. Ideally I'd like to not have nulls etc falling within the angular dispersion of the driver, but perhaps could compromise if necessary on just those angles covering the listening position (which in this case might be reasonably large anyway, due to proximity).
That would be easy enough to achieve by itself, so could well be the result. It just depends on how this particular ideal sits within the wider bundle of compromises that I inevitably end up making, whatever they turn out to be.
Thanks again,
Kev
That would be easy enough to achieve by itself, so could well be the result. It just depends on how this particular ideal sits within the wider bundle of compromises that I inevitably end up making, whatever they turn out to be.
Thanks again,
Kev
Well, I did a bit more listening to the alpair 5.3 (in a suitable sealed enclosure) in order to get a feel for roughly where I might want to cross it over. Having done so, I can't imagine wanting to cross over any higher than 500hz (and in all probablility it would be lower - 300-400hz perhaps).
The 5.3 doesn't have a massive amount of impact at lower frequencies (given its small size) and not a huge amount of excursion either, so I needed a bit of reassurance. But yes, it is absolutely fine in this application of close, quiet listening.
It also means I can get Dave's preferred 1/4 wavelength separation even with an inch or more of space between the drivers. So a pleasing result.
The 5.3 doesn't have a massive amount of impact at lower frequencies (given its small size) and not a huge amount of excursion either, so I needed a bit of reassurance. But yes, it is absolutely fine in this application of close, quiet listening.
It also means I can get Dave's preferred 1/4 wavelength separation even with an inch or more of space between the drivers. So a pleasing result.
Don't know if the "1/4 wavelength rule" applies to desktop distance, like <1m. What about... omni? Both drivers lie horizontal.
The Alpair 5.2/3 by itself in a compact tapered TL can play organ and piano deep bass fortissimo with easy authority; bested my Joseph Audio RM7XL (BR) in bass extension/clarity/impact.
https://www.diyaudio.com/community/...rkaudio-alpair-5-2-in-evansound-3-75l.393909/
The Alpair 5.2/3 by itself in a compact tapered TL can play organ and piano deep bass fortissimo with easy authority; bested my Joseph Audio RM7XL (BR) in bass extension/clarity/impact.
https://www.diyaudio.com/community/...rkaudio-alpair-5-2-in-evansound-3-75l.393909/
I too don't know if the 1/4 wavelength ideal still applies. Though the angular separation between drivers (as seen from the listening position) increases as one gets closer, so it seems at least possible that their separation might (in some way that I'm ignorant of) become more important for close listening. Happily though, it looks like the 1/4 wavelength distance will happen by default in this case, so I shall have satisfied it whether it matters or not.
Here the 5.3 is to be in quite a modest little enclosure, so I'm not expecting it to perform as it would in a TL; instead the small 15w woofer will handle lower frequencies. The 5.3 will still easily cover down to 500hz (and lower) though, which is really just midrange territory, so at the meoment I'm not expecting to cross over any higher.
Here the 5.3 is to be in quite a modest little enclosure, so I'm not expecting it to perform as it would in a TL; instead the small 15w woofer will handle lower frequencies. The 5.3 will still easily cover down to 500hz (and lower) though, which is really just midrange territory, so at the meoment I'm not expecting to cross over any higher.
Nice cut on the bowl sides, how did you cut them?
I don't have a feel for bamboo (that's the material, yes?), so the fiber/grain might not be amenable to this option: You might think about making the slats thicker than the rebate, angled on the side (or gluing the even # slats first, then fitting the odd # in with matched angles works too). Then you could carve/sand it down to the profile without a lot of filler needed. Just need to ensure you use fully dried/shrunk material, but again, I don't know bamboo.
Thanks. I just cut the bowls by hand: a tenon saw to cut a groove of the right depth for the end of the rebate, then a rasp to remove the material up to the groove. In a few places chissels worked, but mostly the grain (if it is called that for bamboo) was in the wrong direction for clean chisselling. If I'd still got workshop facilities there would have been better and easier options so I'm not really advocating this method, but currently I've just got a chair in the living room with a dust sheet under it.
Yes it is bamboo, and I too am unfamiliar with it. It seems quite light and strong and yet reasonably hard, so I'd probably not want to sand it too much by hand. It is said to be acoustically quite good though, and there have been quite a lot of speaker builds using these bowls (mostly to make spheres), so I hope that bodes well for the finished product.
Yes it is bamboo, and I too am unfamiliar with it. It seems quite light and strong and yet reasonably hard, so I'd probably not want to sand it too much by hand. It is said to be acoustically quite good though, and there have been quite a lot of speaker builds using these bowls (mostly to make spheres), so I hope that bodes well for the finished product.
My first thought is to use 2 separate bowls, one on top of the other but then I thought just one bowl for the woof and a planer tweet on top in free air. What is WAW configuration?
Here is an interesting study.
Here is an interesting study.
Bravo! Even better work done by hand!
I modified some bamboo cutlery boxes and ran into the fiber/grain issue. I was able to use a table saw with fine-tooth blade, band saw, and to some extent the scroll saw (although that courted disaster), otherwise anything I tried to do with a knife or chisel was completely controlled by the fibers. But from what I remember sanding wasn't really an issue.
For the sanding down, I might think about making a temporary [lathe]. When I built a speaker from Acaia bowls (Target brand) they were too big for my lathe, but I wanted to get the store finish off - and like you did not relish sanding by hand (and trying to keep it's shape). So I got a lazy-susan/cake-spinner for the open end and then pin-nailed a block with a bolt on the back. That way I could use a hand drill to spin the bowl and more evenly sand the finish off. Thinking about it now, why didn't I use the drill press?
If you are making another side for this, you might be able to tack them together and tack a block on top and bottom and similarly spin it round to do the shaping/sanding. It would work to smooth out the wood or the filler/epoxy either way. Not sure how you are planning to mount it, hanging, pedestal, or something sticking into the side. But if it fits your mounting plan, a threaded rod running through the center-line of the whole thing might let you spin it on that and provide a mount to help keep any stuffing in place.
I modified some bamboo cutlery boxes and ran into the fiber/grain issue. I was able to use a table saw with fine-tooth blade, band saw, and to some extent the scroll saw (although that courted disaster), otherwise anything I tried to do with a knife or chisel was completely controlled by the fibers. But from what I remember sanding wasn't really an issue.
For the sanding down, I might think about making a temporary [lathe]. When I built a speaker from Acaia bowls (Target brand) they were too big for my lathe, but I wanted to get the store finish off - and like you did not relish sanding by hand (and trying to keep it's shape). So I got a lazy-susan/cake-spinner for the open end and then pin-nailed a block with a bolt on the back. That way I could use a hand drill to spin the bowl and more evenly sand the finish off. Thinking about it now, why didn't I use the drill press?
If you are making another side for this, you might be able to tack them together and tack a block on top and bottom and similarly spin it round to do the shaping/sanding. It would work to smooth out the wood or the filler/epoxy either way. Not sure how you are planning to mount it, hanging, pedestal, or something sticking into the side. But if it fits your mounting plan, a threaded rod running through the center-line of the whole thing might let you spin it on that and provide a mount to help keep any stuffing in place.
If the drivers are this close then they are essentially coincident. And time aligned (ignoring XO). Means they appear as a single source with no lobing.
dave
Like the Frugel-Horn Lite it is more likely just doing a really good job of the first harmonic.
dave
Thanks for the suggestions. I certainly do miss having a lathe for such projects these days, but a temporary one may indeed be better than nothing. Perhaps I've hampered that option slightly, both with the rather crude manual alignment of things and with the orientation of the large bowl, since they aren't entirely round in that direction. But we will see, it might still be made to work if needed.
Whatever the means used, I don't expect this to be especially perfect even at best. So I won't be going for any kind of smooth shiny finish, which would show up all the imperfections. I'm thinking more like a matt textured finish to it; in the past I used some flecked stone-effect textured spray paint, so probably something like that.
It's why I thought about a threaded rod going through the entire center. If you can get the center nailed down, the large bowl direction shouldn't hurt getting the slat section evened out, it's just like the square section on a table leg. Especially if you are just using sandpaper strips.
If you are thinking about it, those Fidget spinners that were a fad a while back have surprisingly decent bearings for things like that (most are not sealed, so require some oil/care).
If you are thinking about it, those Fidget spinners that were a fad a while back have surprisingly decent bearings for things like that (most are not sealed, so require some oil/care).
I think TLonken ought to sound very different from a FH. For clarification, the 3.75L partially-constructed 0.88m-line TLonken bested the Joseph (~5 times cab volume), but when completed 1.05m-line played much lower still with gusto. The difference isn't in the "first harmonic".
Acoustic-instrument concert/recital music don't usually go 30hz fundamental for pragmatic reasons (nobody could hear unless played terribly loud). I experienced Engleskyts for soprano and organ at the Grace Cathedral in SF, and routinely play a CD recording (not of the live performance) for basic sound evaluation as well as enjoyment.
Thanks again for the suggestions. As usual, the finished speakers will also be the first prototype, so I'll be making a lot of things up as we go! For now, I got some time last night to glue the slats to the outside of the bowls with epoxy adhesive. Then some bits of ply were glued across the gaps on the inside, using a high quality grip-fill type adhesive:
The result seems pretty stiff and strong, whilst fairly dead. I was going to add some reinforcing/bridging rings on the inside to join the ends of the inner slats to the inside of the bowls. However, on reflection I'll probably just add a third but longer layer of slats on the inside that will sandwich the bowl rims, like they do on the outside; that'll then be very strong.
The profile looks quite pleasing (to me, anyway). It isn't the most accurate construction in the world, but as a first-stab at this I'm hoping to simply work to the profile of the machine-made bowls. So, sand/file down any high spots of the outer slats, cover with a thin layer of filler and use a straight edge to smooth it, resting on the top and bottom bowls to get the line. Later when the filler is dry some sand-paper wrapped around a straight bit of wood should work similarly. If not then we'll give the temporary lathe idea a go!
It is interesting to see that the outside of the shape (on the left) appears smaller than the true width/height (on the right), due to the curves.
The result seems pretty stiff and strong, whilst fairly dead. I was going to add some reinforcing/bridging rings on the inside to join the ends of the inner slats to the inside of the bowls. However, on reflection I'll probably just add a third but longer layer of slats on the inside that will sandwich the bowl rims, like they do on the outside; that'll then be very strong.
The profile looks quite pleasing (to me, anyway). It isn't the most accurate construction in the world, but as a first-stab at this I'm hoping to simply work to the profile of the machine-made bowls. So, sand/file down any high spots of the outer slats, cover with a thin layer of filler and use a straight edge to smooth it, resting on the top and bottom bowls to get the line. Later when the filler is dry some sand-paper wrapped around a straight bit of wood should work similarly. If not then we'll give the temporary lathe idea a go!
It is interesting to see that the outside of the shape (on the left) appears smaller than the true width/height (on the right), due to the curves.
Last edited:
Looks great!
The slats could be a design accent! Almost like a corset on a nesting doll.
I used a rescued marble table insert with sheet sandpaper as a convenient way to true the [gasket/sealing] surfaces.
The tolerances on the Target bowls wasn't that great, so I spent some time working out which bowl to mate to which and how to rotate them as I was clamping/testing the sound, then forgetting to mark them in a way I wouldn't sand or cut off, so I had to do it all over again, and again. But I often serve as a bad example. 🙂
I'll try to shut up now.
The slats could be a design accent! Almost like a corset on a nesting doll.
I used a rescued marble table insert with sheet sandpaper as a convenient way to true the [gasket/sealing] surfaces.
The tolerances on the Target bowls wasn't that great, so I spent some time working out which bowl to mate to which and how to rotate them as I was clamping/testing the sound, then forgetting to mark them in a way I wouldn't sand or cut off, so I had to do it all over again, and again. But I often serve as a bad example. 🙂
I'll try to shut up now.
Ha, I'd not spotted the corset or nesting doll similarities 🙂
Yes, I suppose if I'd made things 'much' more accurately then a bamboo finish might have been possible. But it is all done too roughly for that, and IMO the slats are in the wrong orientation compared to the rings of the bowls, which wouldn't work too well visually. To do that, I'd really want to glue up a hollow blank with everything aligned harmoniously and turn it on a proper lathe, but I don't have those facilities any more. So all in all, an opaque (painted) finish hiding lots of filler still seems the most practical answer.
Anyway, there will now be a long and repetitive interlude whilst I tediously make two more of these by hand in spare moments. Because they aren't very accurately made, the current two halves will be for left and right speakers. Their corresponding/mating halves will be made by taping the new bowls to the existing halves before gluing everything up, so they will at least match well in spite of the dodgy workmanship!
Yes, I suppose if I'd made things 'much' more accurately then a bamboo finish might have been possible. But it is all done too roughly for that, and IMO the slats are in the wrong orientation compared to the rings of the bowls, which wouldn't work too well visually. To do that, I'd really want to glue up a hollow blank with everything aligned harmoniously and turn it on a proper lathe, but I don't have those facilities any more. So all in all, an opaque (painted) finish hiding lots of filler still seems the most practical answer.
Anyway, there will now be a long and repetitive interlude whilst I tediously make two more of these by hand in spare moments. Because they aren't very accurately made, the current two halves will be for left and right speakers. Their corresponding/mating halves will be made by taping the new bowls to the existing halves before gluing everything up, so they will at least match well in spite of the dodgy workmanship!