I've never built or heard a Multiple Entry Horn, I probably will try it later this year as the idea is starting to appeal to me. But I'm already interested in how they typically use mid/bass cone drivers firing through smallish bandpass ports.
Particularly that it permits small (1/4-wavelength) centre-to-centre spacing yet with reasonably large and dynamic drivers, also offers reduced cut-outs for the nearby High-Frequency driver to diffract off, and (if I understand correctly) maybe some distortion benefits in constraining the cones. As always, there are down-sides, such as reduced bandwidth for the cone drivers.
Other concerns might include their output travelling down the horn and being reflected out again, HOMs, and the more complex, tricky design/testing/tweaking involved. But.. if you didn't need horn loading or the directivity control of a waveguide (e.g such as in a near-field scenareo), might using the same configuration but in a flat baffle work well? A baffle is in some ways a 180-degree waveguide, so I wonder if there is mileage in doing this. But I'm concerned that some of the MEH magic might indeed come from the horn.
All new territory for me, so before I get too involved I wondered what more experienced members though of the idea?
Thanks,
Kev
Particularly that it permits small (1/4-wavelength) centre-to-centre spacing yet with reasonably large and dynamic drivers, also offers reduced cut-outs for the nearby High-Frequency driver to diffract off, and (if I understand correctly) maybe some distortion benefits in constraining the cones. As always, there are down-sides, such as reduced bandwidth for the cone drivers.
Other concerns might include their output travelling down the horn and being reflected out again, HOMs, and the more complex, tricky design/testing/tweaking involved. But.. if you didn't need horn loading or the directivity control of a waveguide (e.g such as in a near-field scenareo), might using the same configuration but in a flat baffle work well? A baffle is in some ways a 180-degree waveguide, so I wonder if there is mileage in doing this. But I'm concerned that some of the MEH magic might indeed come from the horn.
All new territory for me, so before I get too involved I wondered what more experienced members though of the idea?
Thanks,
Kev
A flat baffle can be thought of as a 180 degree ( non-optimized ) horn. If the loudspeaker is remote from walls, you also have to consider the effects of the cabinet edges, a large roundover or chamfer here is worthwhile.
I have recently entered the MEH discovery process on my own, modifying an off the shelf 18Sound XT1464 horn. Having made a few measurements here and there after my normal job and life commitments, I can say what I see is quite enlightening.
There are many paths to take, you could chose to use a compression driver, or if you don't need the spl capabilities of a compression driver, you could try a full range driver at the throat. ( modified throat with a roundover )
This is what I have mostly found out:
Where to cross? That depends on your HF device and it's limitations. This will set your tap point for the midrange driver(s). You want to tap in at a point in the horn where the HF device is already very close to or slightly larger than 1 wl at the expected crosover, in terms of area. Note that differing horn profiles will have different areas at the same z axis length from the throat flange. The tap to tap measurement and the throat need to be within 1/4wl of each other. You could use any number of midrange drivers, provided you can physically fit them on the horn and they don't get in the way of each other or the compression driver etc. Think of it as a 3D puzzle that you don't have all the pieces until you do some mockups.
Other considerations, to avoid or reduce pattern flip you want to keep the horn either axisymmetric or at least similar horizontal and vertical angles. Axisymmetric profiles will tend to have a slight measureable dip on axis, so keep that in mind.
I have recently entered the MEH discovery process on my own, modifying an off the shelf 18Sound XT1464 horn. Having made a few measurements here and there after my normal job and life commitments, I can say what I see is quite enlightening.
There are many paths to take, you could chose to use a compression driver, or if you don't need the spl capabilities of a compression driver, you could try a full range driver at the throat. ( modified throat with a roundover )
This is what I have mostly found out:
Where to cross? That depends on your HF device and it's limitations. This will set your tap point for the midrange driver(s). You want to tap in at a point in the horn where the HF device is already very close to or slightly larger than 1 wl at the expected crosover, in terms of area. Note that differing horn profiles will have different areas at the same z axis length from the throat flange. The tap to tap measurement and the throat need to be within 1/4wl of each other. You could use any number of midrange drivers, provided you can physically fit them on the horn and they don't get in the way of each other or the compression driver etc. Think of it as a 3D puzzle that you don't have all the pieces until you do some mockups.
Other considerations, to avoid or reduce pattern flip you want to keep the horn either axisymmetric or at least similar horizontal and vertical angles. Axisymmetric profiles will tend to have a slight measureable dip on axis, so keep that in mind.
I experimented a bit with covered drivers, i.e. bandpass. I tried various shapes in front of a BC 8PE21 and it did as expected - much better directivity and rolled off high end. I am not sure if you can get the mids close enough to the tweeter - you can move the bandpass hole to the side of the driver, but it further reduces the high end of it due to cancellations from the uneven distance from the cone. I measured a Visaton FRS5X in a very small 3D printed bandpass box on a baffle as a test for a MEH - and it was easily usable up to 1.5kHz with the hole offset from the center and volume under the cone filled to allow 2 mm cone travel.
And I guess you have seen this thread https://www.diyaudio.com/community/...nging-loudspeakers.408415/page-2#post-7590579 - very inspirational, using 3 bandpass sections.
And I guess you have seen this thread https://www.diyaudio.com/community/...nging-loudspeakers.408415/page-2#post-7590579 - very inspirational, using 3 bandpass sections.
Thank you everyone, this may not be such a daft idea then; in fact it might well be worth trying a prototype. I've a bit more to learn first, and need to install some simulation software to see how such a thing might best be configured.
GM you may be reading my mind! I had indeed thought about a driver over/surrounding the tweeter, but need to work out if things like cone size, cavity size and port length would be acceptable. The flat baffle might even be taken further, into a convex shape such as a sphere, which may help such a configuration and also help with baffle-edge diffraction.
My tweeter is actually a 1" compression driver (HF108) which I've been trialing without a waveguide and am quite excited about the results. Part of the impetus for this thread is next wanting to pair it with cone driver(s) also of high sensitivity and dynamics. I can come close to 1/4 wavelength separation with a conventional speaker layout but only with pretty small cones. Firing through ports should allow the kind of driver(s) I would like yet at a better centre-to-centre distance and may even offer less edges near to the HF driver. At least that is my hope, we'll have to see how it goes in practice, and in my slightly inexpert hands.
GM you may be reading my mind! I had indeed thought about a driver over/surrounding the tweeter, but need to work out if things like cone size, cavity size and port length would be acceptable. The flat baffle might even be taken further, into a convex shape such as a sphere, which may help such a configuration and also help with baffle-edge diffraction.
My tweeter is actually a 1" compression driver (HF108) which I've been trialing without a waveguide and am quite excited about the results. Part of the impetus for this thread is next wanting to pair it with cone driver(s) also of high sensitivity and dynamics. I can come close to 1/4 wavelength separation with a conventional speaker layout but only with pretty small cones. Firing through ports should allow the kind of driver(s) I would like yet at a better centre-to-centre distance and may even offer less edges near to the HF driver. At least that is my hope, we'll have to see how it goes in practice, and in my slightly inexpert hands.
Hi,
1kHz is 34cm long, so 1/4wl is about 8cm at 1kHz. So you could crossover the ~10cm radius HF108 to a ~6cm diameter woofer, or aperture, to get within 1/4 at 1kHz. At 2kHz it wouldn't be posible anymore, because the tweeter body is already bigger than 1/2 wavelength, so you'd must use a waveguide of sorts to make room for woofer flange to get it physically close enough, right beside the throat. And soon you'd be on the MEH with less than 180deg coverage to have smoother directivity 🙂
Why not use coaxial driver without any holes? Or fullrange driver with woofer to get crossover << 1kHz?
1kHz is 34cm long, so 1/4wl is about 8cm at 1kHz. So you could crossover the ~10cm radius HF108 to a ~6cm diameter woofer, or aperture, to get within 1/4 at 1kHz. At 2kHz it wouldn't be posible anymore, because the tweeter body is already bigger than 1/2 wavelength, so you'd must use a waveguide of sorts to make room for woofer flange to get it physically close enough, right beside the throat. And soon you'd be on the MEH with less than 180deg coverage to have smoother directivity 🙂
Why not use coaxial driver without any holes? Or fullrange driver with woofer to get crossover << 1kHz?
This does not have the same meaning as you might consider with a conventional design.
Thanks, yes those are the distances I'd arrived at, too, which are quite an attractive proposition in a few ways.
My journey so far has also involved testing (a few sizes of) full-range markaudio alpair drivers, though I haven't tried any modern coaxials yet. The 5.3 and 7MS sounded nice, and I agree they'd offer a solution that I'd be pretty happy with. Going down to a few hundred hertz or less offers great design flexibility and convenience without much complication; in a sense this is a fall-back that I have banked.
But... having tested compression drivers I'm more keen to use these if I can; they sound extremely articulate and genuine to me. Not quite as easy to implement, but happily not too bad either; at the SPLs I want, even quite modest ones can easily go below 1khz without horn loading. So that is what has brought me to the idea in this thread.
Oh, that is interesting, Allen, thank you; I may have misunderstood something important as I hadn't made a distinction. Could you possibly point me towards understanding this?
I was reading some threads around the Danley designs before making the OP. As I understand it the distance between each tap/port should be less than 1/4 wavelength of the frequencies they're used up to, as well as being placed less than 1/4 wavelength from the high-frequency driver. Some people seem to feel this should apply to the diagonals rather than just adjacent taps, which might make things slightly less generous.
Though it isn't simple; another suggestion is that the taps/ports should be placed at a position where the horn internal circumference equates to one wavelength. That clearly can't apply to a flat baffle or even very wide angle waveguides. So maybe that is some magic I'd lose by not using a horn?
Thanks again,
Kev
In the Danley horns the taps are expected to travel back to the compression driver source back again, and cause a cancellation when you reach half a wavelength. This is used to advantage with the crossover.
Clearly this is a different issue altogether than the commonly discussed 1/4 wavelength spacing.
The thing about regular 1/4 wavelength spacing is the wide lobe at the cross. In this case you're narrowing the directivity with the waveguide, and this is going to change the parameters. You won't need the same lobe width.
Thirdly, the waveguide expands the source and the woofers are going to be right next to it in any case. This needs to be looked at through different eyes. If unsure, take polar measurements when you cross and you should be able to make it all come together.
Clearly this is a different issue altogether than the commonly discussed 1/4 wavelength spacing.
The thing about regular 1/4 wavelength spacing is the wide lobe at the cross. In this case you're narrowing the directivity with the waveguide, and this is going to change the parameters. You won't need the same lobe width.
Thirdly, the waveguide expands the source and the woofers are going to be right next to it in any case. This needs to be looked at through different eyes. If unsure, take polar measurements when you cross and you should be able to make it all come together.
Ah, many thanks, Allen. I see now that things are indeed more different than I had appreciated.
The reflected/cancelling waves in particular; avoiding them with a flat baffle seemed to have some advantage, but it actually defeats a key aspect of the design. Probably the waveguide expanding the source is also more important than I'd given credit to.
More to learn then, as always. I will make the attempt but accept that I might well end up reverting to a more conventional design once I understand the trade-offs more clearly.
Thanks again!
Kev
The reflected/cancelling waves in particular; avoiding them with a flat baffle seemed to have some advantage, but it actually defeats a key aspect of the design. Probably the waveguide expanding the source is also more important than I'd given credit to.
More to learn then, as always. I will make the attempt but accept that I might well end up reverting to a more conventional design once I understand the trade-offs more clearly.
Thanks again!
Kev
I don't think the cancellation is an essential aspect of the design although it's certainly difficult to ignore.
In any case I expect you'll find some parallels with an MTM, but with the additional variable of the waveguide to do with as you wish.
Good luck, you may make a breakthrough 🙂
In any case I expect you'll find some parallels with an MTM, but with the additional variable of the waveguide to do with as you wish.
Good luck, you may make a breakthrough 🙂
The flat baffle would certainly work well with a small high efficiency mid(s).
There's a big difference between a hornless compression driver and a horn/MEH in a typical room...
The main question to answer is what kind of presentation do you want to live with on a daily basis?
This thread was interesting....open baffle eventually won him over in a later thread.
https://www.diyaudio.com/community/...synergy-horn-versus-balls-of-prestige.329636/
There's a big difference between a hornless compression driver and a horn/MEH in a typical room...
The main question to answer is what kind of presentation do you want to live with on a daily basis?
This thread was interesting....open baffle eventually won him over in a later thread.
https://www.diyaudio.com/community/...synergy-horn-versus-balls-of-prestige.329636/
Or this thread is also interesting https://www.diyaudio.com/community/...directivity-and-horbach-keele-filters.284117/
Thank you, everyone. Those links are indeed interesting.
After reading those, it seems likely that I'm over-complicating (again). I'm interested in close driver (or port/tap) spacing because it intuitively seems that as one gets nearer to a speaker the effects of separate drivers should get more noticeable. But if lobing is suitable for the listening position and time/distance differences are small enough to seem spacious rather than appear as distinct confounding sources.. well then things get easier.
Perhaps I'm on an unneccessary track that could bring more difficulties than benefits; more testing of driver separation needed, and easy enough to do!
Thanks,
Kev
After reading those, it seems likely that I'm over-complicating (again). I'm interested in close driver (or port/tap) spacing because it intuitively seems that as one gets nearer to a speaker the effects of separate drivers should get more noticeable. But if lobing is suitable for the listening position and time/distance differences are small enough to seem spacious rather than appear as distinct confounding sources.. well then things get easier.
Perhaps I'm on an unneccessary track that could bring more difficulties than benefits; more testing of driver separation needed, and easy enough to do!
Thanks,
Kev
I might not bother with a baffle or enclosure in your case.
Pair the HF108 with a high quality midrange like this BMS.
https://en.toutlehautparleur.com/br...s/5-pouce/speaker-bms-5n160-8-ohm-5-inch.html
BMS coaxials are also very nice.
https://en.toutlehautparleur.com/br...axial-speaker-bms-5cn162-8-16-ohm-5-inch.html
Find a way to hang them with a minimum or no baffle. Sounds like you can cross around 800hz with your spl requirements. Spacing issues won't be a big problem here.
Add a 15" or two with healthy xmax and some voltage behind it. Enjoy.
Lots of options of course. There's no right answer.
Pair the HF108 with a high quality midrange like this BMS.
https://en.toutlehautparleur.com/br...s/5-pouce/speaker-bms-5n160-8-ohm-5-inch.html
BMS coaxials are also very nice.
https://en.toutlehautparleur.com/br...axial-speaker-bms-5cn162-8-16-ohm-5-inch.html
Find a way to hang them with a minimum or no baffle. Sounds like you can cross around 800hz with your spl requirements. Spacing issues won't be a big problem here.
Add a 15" or two with healthy xmax and some voltage behind it. Enjoy.
Lots of options of course. There's no right answer.
Yes, in some ways the number of options is exciting, but taking me some time to explore.
By this point, it seems likely that I'll end up building two different pairs of speakers (instead of trying to satisfy my uses in one design). That would be one compact pair for close listening and a bigger pair for more typical room distances. The latter may well use big drivers for directivity control, unless I go the MEH route.
For the smaller pair then yes the HF108 doesn't need an enclosure or baffle at all, though it could be put in one if wished, just for reasons of diffraction or aesthetics. So it is usefully compact for the range above several hundred hertz or so. The driver below it (in frequency) could be open-backed or in a small enclosure; the latter might work best for compact baffle dimensions. But Ive yet to decide if this will be more than a 2-way; 2-ways would suffice but it would also be appealing to use a smaller driver, which would mean crossing either to side-firing woofers or subwoofers.
Thanks,
Kev
By this point, it seems likely that I'll end up building two different pairs of speakers (instead of trying to satisfy my uses in one design). That would be one compact pair for close listening and a bigger pair for more typical room distances. The latter may well use big drivers for directivity control, unless I go the MEH route.
For the smaller pair then yes the HF108 doesn't need an enclosure or baffle at all, though it could be put in one if wished, just for reasons of diffraction or aesthetics. So it is usefully compact for the range above several hundred hertz or so. The driver below it (in frequency) could be open-backed or in a small enclosure; the latter might work best for compact baffle dimensions. But Ive yet to decide if this will be more than a 2-way; 2-ways would suffice but it would also be appealing to use a smaller driver, which would mean crossing either to side-firing woofers or subwoofers.
Thanks,
Kev
In retrospect, my original reasons for looking at this were probably flawed. It appears that the lobes resulting from 1/3rd or maybe even 1/2 wavelength driver-separation would actually be fine for my near-field situation. For normal/room distances they'd be less ideal, but I'd be more interested in other designs for room speakers anyway (e.g. for reasons of directivity).
But along the way, I have become quite keen on having a woofer behind the 'tweeter' for practical reasons. It keeps the height of a 2-way speaker smaller (or permits a bigger woofer for the same height). It also seems an elegant way to combine my preference for a sphere with two separate drivers (obviously, there are already coaxial drivers and full/wide range drivers available, but this might be a little different).
So I need to start some simulations of such a bandpass arrangement, and see where that takes me.
But along the way, I have become quite keen on having a woofer behind the 'tweeter' for practical reasons. It keeps the height of a 2-way speaker smaller (or permits a bigger woofer for the same height). It also seems an elegant way to combine my preference for a sphere with two separate drivers (obviously, there are already coaxial drivers and full/wide range drivers available, but this might be a little different).
So I need to start some simulations of such a bandpass arrangement, and see where that takes me.