18Sound 15" and Audiohorn X-Shape 2-way build

[TulseLuper]

Components have arrived for a build I've been wanting to do for a while - a pair of large horn speakers. Mildly old-school aesthetics, but with modern drivers, modern waveguide design, and DSP crossover.

18Sound 15NTLW3500
18Sound ND1460
Audiohorn X-Shape X40 w/ FEA roundover kit

Current bass cabinet plan is ~100L net, vented, and mount the horn/compression driver on top. The goal is response to mid/low 30s in a medium/large room. No nightclub SPL levels required.

Tips/tricks for clever brackets or system for mounting a compression driver and horn on top of a box would be appreciated. Also, if anyone here has experience with Audiohorn's X-shape horns, feel free to chime in with thoughts or tips. The horns seems well built and the smoothing/putty job on the inside is nicely done. The flanges are maybe a bit thin but will work fine I'm sure.

X-shape frequency response no EQ w/ ND1460 (from Audiohorn):

Polar with a JBL compression driver:

Bass cabinet something like this, with 2" radius roundovers:

parts:

 

[ppataki]

Components have arrived for a build I've been wanting to do for a while - a pair of large horn speakers. Mildly old-school aesthetics, but with modern drivers, modern waveguide design, and DSP crossover.

18Sound 15NTLW3500
18Sound ND1460
Audiohorn X-Shape X40 w/ FEA roundover kit

Current bass cabinet plan is ~100L net, vented, and mount the horn/compression driver on top. The goal is response to mid/low 30s in a medium/large room. No nightclub SPL levels required.

Tips/tricks for clever brackets or system for mounting a compression driver and horn on top of a box would be appreciated. Also, if anyone here has experience with Audiohorn's X-shape horns, feel free to chime in with thoughts or tips. The horns seems well built and the smoothing/putty job on the inside is nicely done. The flanges are maybe a bit thin but will work fine I'm sure.

X-shape frequency response no EQ w/ ND1460 (from Audiohorn):
View attachment 483511

Polar with a JBL compression driver:
View attachment 483512

Bass cabinet something like this, with 2" radius roundovers:
View attachment 483513

parts:
View attachment 483516

Do you know already how your crossover settings will look like?
What DSP tool will you use?

 

[digitalfrost]

Looks interesting. Drivers are pricey, but if you can get good directivity this ticks a lot of boxes.

edit: https://audiohorn.net/how-to-implement-my-horn-and-my-speaker/ I just read this. Is it really that easy? If you can measure und use REW that is all there is to it?

 

[TulseLuper]

Do you know already how your crossover settings will look like?
What DSP tool will you use?

I have not decided yet but either a FusionAmp built in or a minidsp unit. Probably minidsp for it's ease of use and because I don't mind separating out the amps. In fact it's cleaner sometimes with only one cable going to the speakers instead of both power and signal like with a plate amp. I'm considering building a box for a rack mount multichannel Buckeye amp and minidsp SHD - would veneer the box to match the speakers and have a nice looking kit, I hope.

For crossover settings, I'll measure the woofer in box as best I can and just try to figure out where we have the best directivity match with the horn. Looks like there should be a lot of flexibility between ~750Hz and 1000Hz. Steep slopes could help with the lower xover points, will just have to try a few things and measure.

 

[TulseLuper]

Drivers are pricey

They are pricey. I managed to find both the woofers and compression drivers on sale at different times from different retailers. My actual price paid, shipped and tarriffed to USA, for all parts including horns is $2,860. The horns were by far the most expensive part (almost 50% of the cost).

 

[ppataki]

I have not decided yet but either a FusionAmp built in or a minidsp unit. Probably minidsp for it's ease of use and because I don't mind separating out the amps. In fact it's cleaner sometimes with only one cable going to the speakers instead of both power and signal like with a plate amp. I'm considering building a box for a rack mount multichannel Buckeye amp and minidsp SHD - would veneer the box to match the speakers and have a nice looking kit, I hope.

For crossover settings, I'll measure the woofer in box as best I can and just try to figure out where we have the best directivity match with the horn. Looks like there should be a lot of flexibility between ~750Hz and 1000Hz. Steep slopes could help with the lower xover points, will just have to try a few things and measure.

I think for such a huge woofer, the lower the crossover point, the better (would be great to see off-axis measurements about the 18Sound 15NTLW3500 though)
If I am not mistaken, the JBL M2 has it at 700Hz (also using a 15" woofer)

 

[Elgrosso]

Cool, I have the same project, my x40 is in the printing queue.
Still undecided on the drivers as I’m currently studying the final placement and expected response before picking a direction for the design.
I’m doing big changes on the room too so it might be full range, or multi subs + satellite.
it will also depend of the price, so far on my list are the wom380, the ntlw, the ftr15 and the SBA depending of the targeted box.

I already have an x28 on a 3 way (1+10+2x15) and am very happy with it.
But I’ve made some choices mostly for aesthetic at the time that came with impacts on directivity that I under-evaluated.
So this time I will take a better approach with external measurements (allowed by a much smaller speaker), vituix simulations, care on diffraction, baffle shape etc.
Just following Nicolas’ advices here, I'm sure you got plenty too.

I’m with you on the pwr-ice, much friendlier than fusion, minidsp plugins are so good (but I wonder if device console manages the pwr serie now?)
But when I had them they were too noisy on the compression drivers, just something to adjust with a passive component maybe.

So it’s great, I’l follow your build for sure! Mine might take few months.

 

[tmuikku]

I think for such a huge woofer, the lower the crossover point, the better (would be great to see off-axis measurements about the 18Sound 15NTLW3500 though)
If I am not mistaken, the JBL M2 has it at 700Hz (also using a 15" woofer)

15NTLW in quite small 46x46x40cm or so closed cab. If I remember correctly this is with 5ms windowing so take accuracy on < 1kHz with grain of salt.
edit. Added ideal radiator in ideal baffle (vituixcad diffraction tool) for comparison.

It's huge issue for home builders to get reliable measurement data for < 1kHz crossovers.

 

[ppataki]

Thank you for sharing this @tmuikku
Based on this I would definitely cross it (well) below 1kHz

 

[tmuikku]

Yeah it's good idea, for any driver in any size, to cross below diameter wavelength or so, due to all kinds of issues emerging above.

Same goes for tweeters as well, if you need to get clean to 20kHz get driver with diameter close to 20kHz wavelength. If it's bigger expect some "issues" in graphs at least. Might or might not be audible.

 

[Fledermaus]

15NTLW in quite small 46x46x40cm or so closed cab. If I remember correctly this is with 5ms windowing so take accuracy on < 1kHz with grain of salt.
edit. Added ideal radiator in ideal baffle (vituixcad diffraction tool) for comparison.

It's huge issue for home builders to get reliable measurement data for < 1kHz crossovers.

Interesting !
In order to better determine the crossover frequency range, perhaps you could bring your VCAD heatmap settings in line with AudioHorn's (scale from -36 to +6 dB with yellow color = -6 dB, angle -90 to +90°) - then one could play mental cut and paste more easily

 

[tmuikku]

Well, here you go.

This is not sufficient to make crossover though. Crossover could be almost anywhere say 500-1200Hz for quite similar results horizontal directivity in mind, assuming your waveguide is sized roughly similarly as the woofer box. After you build yours, measure it, juggle around in vituixcad and you'll find out what the suitable xo is for your application.

In general, you'd want the xo as high as possible but no too high For some reason there is trend to make xo as low as possible. I suggest look at data you have, experiement around, and make it where it sounds best, if it matters at all in your application.

 

[Fledermaus]

Nice, thanks !
Depending on where the X40 loses directivity control in its lower range, from my casual DIYer perspective I'd aim for an Fx around, say, 700 (lower CD flat-axis SPL range) or 800 Hz (avoiding the 15-inch's 1.1 kHz breakup)... but there is indeed room for experimentation - fortunately, since that's where the fun mainly lies

 

[TulseLuper]

Cool, I have the same project, my x40 is in the printing queue.

Excellent, please share your progress and plans.

Just following Nicolas’ advices here, I'm sure you got plenty too.

Nicolas has been helpful and forthcoming. Seems like he's overseen a good amount of builds with these horns. My impression is that although they're relatively flexible, there have been a lot of builds with 18Sound drive units and that seems to be a relatively optimized solution. For example, I asked about exit angles of compression drivers with the X40 and he told me the ND1460 was a match.

pwr-ice

My experience with these is that they are too noisy and almost surely very nasty with a compression driver. Some passive filtering could help but I'd rather keep that to a minimum. I wish miniDSP would update these amps. I'll use the SHD or Flex HTx (or consider the Flex8 if you want FIR filters on the inputs) and a separate multi-channel amp if I do not go with the Hypex Fusionamp.

 

[TulseLuper]

if you need to get clean to 20kHz get driver with diameter close to 20kHz wavelength

I expect to forfeit graph cleanliness in part of the top octave but we'll see how it goes. With my required SPL I might get lucky. Not too worried.

Thanks for the graphs - I haven't been able to find much data/measurements/builds with the 15NTLW.

As far as Xover point - I'm looking to keep it on the low end in part because I'm free mounting the compression driver and horn (Ctc distance from horn to woofer will be a bit longer that way). That said, directivity match will be most important.

 

[tmuikku]

Well, yeah you see that in the measurements. I got ~42cm diameter freestanding waveguide on it as well, 44cm c-c distance and it's good with 800Hz xo with only mild bump on ERDI. But, shapes of your system are different, c-c is different, so you your's might end up somewhere else.

About lobing: Lobing is issue only when nulls get near (vertical) listening axis and the listener is not at the axis but off so that frequency response dips. Luckily, dips aren't that audible so it doesn't seem much of an issue. Lobing, big c-c, smoothens DI but brings the nulls closer to 0-axis. If you always listen at listening axis then it's not an issue at all I'd say. Listening distance affects of course, further awway the importance lessens even more as even small angle covers quite a height (variation). For this reason I think vertical lobing should not be very high in priority, and crossover region chosen with other metrics if necessary, like max SPL before bad sound.

Whole system simulated response with sub. DSP with FIR capability but only limited amount of taps, barely enough for 800Hz xo LR8. There is still some narrowing ~500-1200Hz. Smaller waveguide would help some, for example. I'm also bit sceptical about the measurement data, done outside as high as I could get it and it's still very low resolution around xo. Otherwise quite similar what you have there.

 

[TulseLuper]

Larger bass bin, wider baffle, big roundover for me should all change up the 15" response a bit. Also my horn looks shorter than yours so might be able to get it closer to the woofer.

How did you mount the horn/cd to the box?

 

[tmuikku]

I used two strips of ply as stand. There is sockets in the throat section of the waveguide the strips fit into. Another socket mounted on top of the cabinet. Wood screws to hold everything in place. This was simplest I could come up with. Had 3D printed stand but it felt flimsy so used ply instead. Works as a handle as well.These get hauled around.

 

[Elgrosso]

Excellent, please share your progress and plans.

My experience with these is that they are too noisy and almost surely very nasty with a compression driver. Some passive filtering could help but I'd rather keep that to a minimum. I wish miniDSP would update these amps. I'll use the SHD or Flex HTx (or consider the Flex8 if you want FIR filters on the inputs) and a separate multi-channel amp if I do not go with the Hypex Fusionamp.

ok I misunderstood the use of the pwr, the flex/shd are sure a lot better.
Not much to share for now but basic ideas:

I will probably starts with small proto boxes to play with sizes and edge treatment impacts on measurements.
Probably minimal but still I’m curious, even a box slimmer than the woofer if possible.
I also like the more modular approach as it allows to test different low bass types below, like a dipole one for ex crossed up to 2-300hz.
And if the final placement allows good lows then why not a full ranger, but it gets much more complex quick.


Thx tmuikku for all thee measurements! Cool to finally see these with the ntlw.
Do you happen to have any distortion plots too? Or maybe any other driver you tested before?

As a side note, I’m always amazed by the use of LR8 or alike, every-time I tried I didn't like it (for ex I use LR2 right now).
I prefer a sense of « fuller » sound that the LR2 gave me over the "precision/detailed » effect of the LR8.
It might not have been related to the slope itself so much and more to everything else, like the room or just personal taste.

 

[Elgrosso]

OP, have you seen this page?
Very interesting about the interactions.

It seems the horn roundover/FEA allows the wave front to just detach itself at the right time, at the end of the round over.
Of course they’ve been calculated for that, so not much of it goes rearward (top section of the 1st animation).
And below the diffraction along the woofer cone seems minimal.

Then on the 2nd animation the woofer diffraction due to the horn seems to be a bigger problem (well all thing equal).
A free air mount with a not so short ctc should help to fix a lot of these.
At the minimum the empty space in between box and horn would "suck" the waves and avoid a bit of the diffraction (or delay and reduce it).
And no roundover on the box would help the woofer beaming a bit earlier to reduce again.

Now in opposite I also wonder how far can we go to reduce ctc, by moving the free air horn above the woofer a bit.
Probably creating a diffraction mess but also moving all these closer to a point source.
Well I just want to find time and a way to easily test all this.