All sims are in 2pi, 1.41v @ 2 ohm, measured from 10m with values scaled to represent a measurement from 1m. 

Image of how the standard BG horns are simmed.

Image of how the BC horns are simmed. 

Red=BC, Blue=standard BG

BG directivity

BCBG directivity

I think these results are interesting, solid 5hz more extension is nothing to ignore. I suspect with a larger hole, it may have slightly increased efficiency. Let me know what you think. If you want me to try anything let me know! 

It's definitely interesting, some thoughts.

I've measured quite a few horn extensions myself, some followed the sim accurately, some behaved really different somehow. I suspect that is because of the large steps in effective surface area and different behavior to what the mind thinks the air path will be like. I would measure it and then correlate the simulation to the measurement.

Long way of saying, simulate, build, measure, repeat.

The Mouth does not need to be as big as the last part of the throat, following traditional horn terminstion rules, as the real termination is the air bubble contained by the flat baffle, this comes fromt the Bc creator voice... at lower Wl the mouth is not seen and the important thing is continuity between throat and exterior. At higher and smaller Wls where the mouth is seen the size is enough from what I understand.

Interesting threads on diyaudio about Bc and its predecessor  Bdeap from some years ago.

Nice threads lately, this place seems more lively  thanks!

Sorry citizens sir I didn't get back to you regarding the labs. 

I am thinking of junking them so not a big waste to cut a hole in them. 

First problem is I have no idea how big the hole should be. How does it relate to fs? I need to put two together and measure the sq mt idge. 

Next is to do a plot outside rite now would be impossible. I could do one in the workshop but what would it mean? Two labs standard and two labs bc. But as I said inside measurement. I would get wholey flamed for that on this forum.

As always good work sir. 

Remember four labs stacked like that would be 2.4 mtrs high, about. not possible I think two at a time.

mobiele eenheid wrote: The hole looks like an eye in a pyramid, Illuminati confirmed

aahaha, I cant unsee that now. 

mobiele eenheid wrote: Would increasing the hole not simply decrease average horn path length if it's indeed the limiting factor with regards to efficiency

If you look at that hole, then divide it by 4, the cross sectional area per cab is actually really small, <25% of the mouth area of a BG horn. If I make it ~40% of the cross sectional area, it might shorten the horn by 50mm but I doubt that will have a tangible impact compared to not properly coupling with the exterior of the cabs. 

mobiele eenheid wrote: I've measured quite a few horn extensions myself, some followed the sim accurately, some behaved really different somehow. I suspect that is because of the large steps in effective surface area and different behavior to what the mind thinks the air path will be like. I would measure it and then correlate the simulation to the measurement.

What simulation software did these simulations use? I am not saying AKABAK 3 is perfect or even that I have mastered it but it is special when compared to other software used by the DIY community. It doesn't just simulate a straight horn path, that approximates a folded horn. It simulates the actual 3d model of the horn, including the exterior panels, the shape of the driver diaphragm, all the folds, any mistakes made when folding the horn in the design phase. Maybe these sims wont perfectly overlay a real life measurement but I have some confidence that they will tell you what the impact of a change is. Eg. the BC arrangement in my above post does not improve efficiency buy does improve LF extension. 

if you compare the bc218 measurements you`ll see that the bc config adds little to the low corner compared to the gain in directivity due to the `boundary`involved?!

:)

This is 4 standard BGs in a standard config compared to the largest cutout. About 1dB improvement in sensitivity and 3hz lf extension. 

Here are some top-down view maps of pressure at 63hz 

Standard BG

BC 1110cm^2 cutout

BC 1400 cutout

BC1765 cutout - Key got a little messed up on this one, the values are correct, colours are just shifted. 

I am thinking the BG horn is not a good candidate for boundary coupling, the pressure maps above show the standard BG horn to radiate from the whole mouth. The BC horns have a hot spot around the cutout but this reduces by ~9dB at the edge of the baffle. The 15BG100 is a relatively high Qts, this means to achieve reactance annulling via the method layed out in the letters to the editor section of the Leache paper, you need a high T-value. High T-values do not have fast flare rates at the mouth of the horn. If I find the motivation I might try to design a dedicated BC horn using the 15SW115 or even the 15DS115 then repeat this analysis on it. 

Let me know what you think, does my hypothesis make sense to you? 

citizensc wrote: Standard BG BC 1110cm^2 cutout BC 1400 cutout

Wow, top contribution citizensc!!!! Very encouraging results. I see hypercardioid dispersion patterns for the BC version 

Instead of trying to achieve a 2x2 array, it would be easier to just have an horizontal array, place the horns with the mouth downwards and have the aperture on ground level.

taken from:https://www.grc.com/acoustics/an-introduction-to-horn-theory.pdf

I am sorry for the pencil drawing but it was quick and I am very busy rite now. 

Maybe we are missing the point a bit here I don't think that it is about gaining spl and extension so much. Modern speakers are pretty bloody loud as they are. The point for me would be directivity I am interested in how that would sound. Because the bass would be more focused would it sound louder, cleaner? am I wrong in thinking that the Danley are less omnidirectional? More energy focused on the dance floor. Less wasted.

It will sound effortless and clean compared to a 130dB system pushed to the max.

And last longer.

Jo bg agreed 100% here. I'm quite new to this, but initially I was planning on just picking up a Behringer NX6000 for one of these designs because I always read headroom will improve the sound, but would I be fine or better off with a much more trusted brand like Crown or QSC putting out only 1000 watts total to 2 paralleled 8ohm drivers for a total of 2 4ohm loads? In other words what would be better a trusted brand with less headroom or Behringer with more headroom? Because I also thought I've read that not enough power can do damage as well, but with these designs I can't fathom needing to run them anywhere near full power rating for the types of gigs I'm planning (maybe 200-500 people), especially considering hornresp is telling me I'm approaching 130db with only 80watts through 4 drivers??. Anyways here's some hornresp charts for the 15NW100:

With 10 cm backchamber are you putting magnet in throat?

You are not powering a heater with 80w continuous, music is dynamic,  you have to allow for short peaks and transients that will require higher power , so in practice you will need way more if you don't want to clip amd compress.

While they could be called the heart of the system, until you take them to limits (and hear the sound of limiters), amplifiers are the less important part in sound quality. Once you've got nuff voltage and a decent power supply, difference will be minimal compared to difference other parts of the system could make.

A good quality driver, in a fitting enclosure built with proper ply and bracing, good processing and gain staging, will make an enourmous difference compared to difference in amplifiers, as long as those are operated within their limits.

So in your case i would go with a new Berry with warranty, you will have double the power, instead of (usually overpriced) heavyweights with uncertain history and more failure probabilty.

Unless you are going to rent or host a lot of events on this, then you may look for something more reliable and many times more expensive. But berrys are very good for beginners, warranty gives you 3 years of peace of mind.

And new ones are way less ugly than the inukes i bought. Those served me well for almost three years until i upgraded to labs, but at 8 times the price , and more for added reliability and rentability than sound quality. There is an improvement but not overwhelming. 

a very short (0,1) last segment with s4 the size of the eye and s5 the size of the front will approssimate the flat front. 

As for the amp question thanks for the info! I still would guess though that running 400watts out of an amp that's capable of 1000 could be enough headroom to not be concerned about clipping at the peaks? At that point you start to run into displacement issues anyways at about 34hz at 134db, a level I most likely also will not be pushing considering that 116db was about the strongest bass I've felt on a funktion rig measured with the DecibelX app on an iPhone (so not super accurate but maybe not horribly off) and that was approaching uncomfortable for sure.

Try put an hipass filter with filterwizard and drive it with a sensible voltage to get a feeling of real world excursion.

Impedance plot is n9t very readable, can you post 1 speaker with a readable scale (you can change it in options) ?

Seems a waste of driver if you can push only 400w into it. Seems a rating for a 70s sub...

Here's Max Spl

Now here's where I have a question. Does adding a high pass filter really reduce displacement by this much? It seems like for such a modest reduction in SPL you gain big reductions in displacement. Also note that I used the xmax of 9mm instead of the xvar of 11mm which B&C suggest may be a more accurate measurement.

Back to the amp, I know it would theoretically be a waste considering these drivers can handle 1000 watts rms but again I can't fathom needing even 130db for 200-500 people. I also used my phone to measure infront of Mungo's Hifi stacks and that was some chest thumping bone rattling bass coming in at 108db which makes me think maybe 116 might be the limit but I'd assume the Mungo's measurement could still be ballpark.

If 108 had been a reliable measurement, it would have been averaged over a long time, and you can see differences like 20 dBs between peak dB recorded and the average. The speaker needs to provide for the peak, so your 108 dB measurement could contain 128 dB peaks....

Heck we are in 12v territory, not big stacks.

About hi pass... ANY pa sub i know about needs an hipass filter.

And usually higher in frequency and higher order than the one you simmed.

20 subs stacked doing 110 dBs  will actually move LESS air than  1 sub doing 110dBs , because they need less displacement due to directivity increased by increased front size.

20hz is a low hi pass for speciality subs, especially 1st order, nothing common. And too low for a 40hz horn with a 15.

And that is a big sub tuned lower than yours with way more displacement thanks to 2 x 18 drivers.

Those 128 dbs are peaks as luckily music is dynamic, if it were continuos it would be far more dangerous.

Remember that operating any system at it's limit is not a good thing, distortion arises, drivers start to behave unlinearly near xmax, heating changes speaker parameters... so you want some headroom.

If you want to go to the low 30s you need a huuuuge long horn.

Yours is too short to load that low, so the driver has to do the work alone, and as you see in the excursion graph, excursion skyrockets from 40 hz so no sense in going lower than that. 33 hz is quite low for a straight horn, maybe you should look at ths or bandpass. Or bigger horns...

No sense in squeezing 30 hz from a 40 design, you limit your spl a lot from 40 hz up to reach that low with this design.

No sense in building a design that is loud from 40 if you need to play 30 hz, you will not use thay efficency , wasted wood and space. 

The higher the filter order the steeper ( less bass extension ) the cut, but higher order filter also reduce less the level around crossover frequency( a by product ).

While low order seem to reduce level more right around crossover frequencies, they allow more bass to go through beneath, and those frequency down low usually cause more excursion problems. So higher order are usually safer and provide more impact at xrossover frequency, but with worse time behaviour.

In real life, to make 100  people dance modern music, 116 db is hardly enough.

That is the level of one 15 inch plastic speaker on a pole at a birthday party.

And here's a slightly modded version of the NWBC horn in 2pi

Now here's the same horn in 1pi and the reason I think this is important is because in all the other forums on the BC series the size of the face of the horn is often speculated as being responsible for a lot of directivity gains. So while I know the size of this total horn's face (max maybe 5' by 7') is not large enough for the wavelengths targeted, it would be large enough to partially add some directivity gains in the mid to low 30s:

I'm going to keep at it and see what else can be done, but to me just purely based on these response graphs (I'll have another round of displacement and the rest) it seems as though this horn could be quite effective down to about 35hz. Not the 33hz I'm targeting so will keep tweaking, but just for some context that's the lowest I anticipate needing to go. This will be used for dub/reggae/dubstep/drum and bass/funky bass/space bass/house/disco you name it, just pretty much not expecting to do any hardcore or trance or hardstyle as I find that those genres have so many fast kicks that they don't leave enough space between them for proper low frequencies. Not hating because I've had some fun times with those genres but I crave the deep stuff.

That spl graphs tells NOTHING about usable extension without looking at displacement and power, and with real world hi pass.

Take some time to read about hi pass and loudspeaker behaviour, you seam well versed and learning quickly but at the same time you are missing some foundations and you are basing assumptions on wrong deductions.

Again, take some time to study the matter and avoid stupid expensive mistakes - like deducing your subwoofer needs form a measurement that does not inlcude bass.

I just want you to save money and don't build something that will not satisfy.

You can do 200 people with four cabs (well a lot of context would be needed but let's keep it simple) but your design will not do low 30s with satisfaction.

Here you Will find jbl sound system design guide part 1 and 2 for free.

If you like having a good book , bob mc carthy "sound system design and optimization" is a good one to buy.

Alright I'm going to have to get back to work soon here but I'm on a quarantine right now for international travel so the goal is to have this drafted and folded by the end of the next 2 weeks. I'll read through the suggested literature but for now based on previous comments here's some optimizations:

And here's the filter sim with 250 watts per driver in 2pi (ignore the low pass) so I'd expect in reality maybe a couple extra hz and some sensitivity on the low end (again no where near the 1200 rated but this is already hitting 140db and I'd love to keep the drivers nice and cool ideally)

Group delay is looking a bit suboptimal so maybe a 3rd order could work instead?

Maxspl is useful to see at a glance power and displacement limits together but does not provide a real excursion or frequency graph.

You should just put a proper voltage in eg and apply filters without using max spl. 

You could also look at driver power in the acoustical power window tools.

High pass way under box tuning like this wastes a lot of the potential of the box as excursion skyrockets down there, the horn is not loading and the driver is doing almost all the work, while the big back chamber  does not help in controlling the excursion. I doubt the BC part not simmed will make this a 30 hz horn.

Your hp still looks too low, maybe about 35 - 38hz should allow you to use the cab potential from 40hz.

But if you really want that 33 hz with authority a longer horn, or th or bp, maybe even reflex, would be a better choice.

Vtc looks low for 4 18 inch drivers.

So with a 31hz 4th order high-pass at 1800 watts across 4 drivers we're getting an F3 of around 32.5hz keeping within xvar but slightly exceeding xmax all around 136db. I'm up for slightly extending the length a bit but I think this will get reconfigured as I do the folding and we'll see more accurate sims then, but for now does this look far off from what to expect? The group delay still isn't fantastic so it would be great if a 3rd order high pass were possible which according to hornresp actually lowers displacement more. This is still probably way beyond the output I'd be needing but I trust your judgement on whatever SPL you think really brings the eye rattling bass. I have no idea why the BC218 was made to go all the way down to 26 hz other than just to flex or for movie fx (I saw they used them at a big Star Wars show at Disney).

But for my taste HP is still too low. You see how 30 hertz causes much more excursion than the 55 peak? 

It means that you are hindering performance over 40 hz to go that low.

Seems a big waste of space and drivers building such big subs, loaded with neo 18s, that reaches xvar at 1/3 of aes power and can do max 136 in block of 4, and without even going extra deep.

Try to up tour hipass to 35 4th order and see how mich more pressure you can put in it

By the way you wanted to demonstrate that this design could work with "normal" drivers, bur seems like a case where a bigger badder driver could make this work, as you are not loading the lows in the horn you need a strong one.

The volts you input in hornresp are those at amplifier output, nothing to do with outlet voltage, or nobody would run big subs in the us.

Questions like this make me wonder if you should consider getting a bit more knowledge before  designing your sub. Not many places to use some in the next future anyway...

Xvar us is also ok, just know what it means that woofer has allready changed his electrical parameters at this level of excursion and things start to get worse ( a little bit worse than xmax usually),  but still no short term mechanical issues. But in real life excursion will be less for thermal compression and other issues.

Read b&c definition of xmax an xvar so you know what you are using !

I might do some tweaking to improve group delay or again experiment with a 3rd order highpass.

I strongly recommend you check this thread out  https://forum.speakerplans.com/hyperbolic-bass-horns_topic105331.html" rel="nofollow - https://forum.speakerplans.com/hyperbolic-bass-horns_topic105331.html

Especially these two papers:

http://leachlegacy.ece.gatech.edu/papers/HornPaper/HornPaper.pdf" rel="nofollow - http://leachlegacy.ece.gatech.edu/papers/HornPaper/HornPaper.pdf

http://kolbrek.hornspeakersystems.info/images/misc/LFhorns2018.pdf" rel="nofollow - http://kolbrek.hornspeakersystems.info/images/misc/LFhorns2018.pdf

Read the Leach paper before Kolbrek, as it provides the foundation for the Kolbrek paper. If like me you are not an engineer, it may take a bit of work to understand the papers. Trust me though, it will be very worth it.

KaphaSound wrote: Also out of curiosity how many mid/tops are you running on top of those punishers?

We use 4 xtros, they work well for techno/house but I think in 2020 there are some better options. 

My current assumption based on the hornresp and akabak sims is that the size of the smaller mouth exit in the BC config seems to determine the boost in the low corner and resulting additional group delay, while matching the rate of expansion into the flat front face should determine how much additional efficiency is gained. In the diyaudio thread from 10 years ago I see Tom says "For instance for a 30Hz exponential horn, the area cannot expand any faster than doubling about every 2 feet", and "The object of the BC horns is to have the rate of expansion proper so that the end of the bubble is formed by the 60 by 60 inch face dimension and have the inner part of the horn tie acoustically into that external bubble. The horn inside of the cabinet has a larger mouth than the opening in the side of the cabinet and by forcing the output to pass through a restriction puts it in the right place to tie into that bubble and in effect makes the mouth larger and horn path longer." 

Art Welter also says "Hornresp seems to ignore the effect of a segment going out to a larger flat boundary, while I have measured a 3 dB increase in LF response doubling frontal area. Just another one of those things that keeps me making sawdust rather than spending too much time simulating."

A question remaining is how exactly a less than perfect expansion rate would affect the response. In other words, is it really possible to simply cut a hole in the side of most FLHs to gain 3-5hz extra extension without drastically affecting the rest of the response as shown in citizensc's Akabak sims? The sensitivity doesn't appear to change much, but Tom does point out that he uses "Akabak which is a powerful, somewhat cumbersome program but it has real limitations and in bass horn land, should be viewed as a “ball park prediction”".

Anyways, without further ado I present the "HBH15" (name is of course subject to change at any time.) Here's a quick photo of the assembled cab. Plan is to apply tung oil for a natural finish: 

This was a super quick and dirty measurement in REW, nearfield 1pi:

And here's the original simmed response in 1pi (I lost this exact one but remember the values well enough): 

A few notes. I suspect the nearfield measurement is coloring things, hopefully making the response appear more like a 2pi measurement but tbd. Even so, a close response. The main thing for me was making sure that low corner was in fact there, since in Hornresp this required some guesswork. The panel that's ratchet strapped on would start resonating like crazy at a certain point as well (I guess more reason to build the second soon enough) which more than likely colored this not quiet measurement.

Group delay measurement vs sim:

As for subjective notes: I'm very happy so far, but much more testing to be done. The room I have this in is big with a tall ceiling, but the nearfield was a pleasant and "tight" massage. I'm starting real small and just have a 450 watt amp, but I was barely seeing the signal light. Simmed response suggests this cab will hit xvar at around 900watts. 

I definitely want to shoutout everyone that's helped on this project here in the forum, especially Jo bg and citizensc for humoring my ramblings and running some great Akabak sims that convinced me to cut some wood and chance it. I also need to thank several local sound crews in Seattle for their input (Milestone Productions and another crew to emerge soon...) And of course Tom Danley for all the inspiration and for creating this type of design (maybe Lambda Labs Digital horn too, although I believe it's a TH). Oh and of course David McBean for creating Hornresp!

Let me know what y'all think and if you have any Qs. Hoping for more measurements soon.

Surely horn resp factors this into the math. As in, if your schematic is accurate to the tee, your simlulated response should be reliable.
Or not?

 

The BC215 seems like the most versatile version Danley make at the moment, and it's one of the most power dense cabinets from what I know but as a weekend warrior/DIYer I really wanted/needed something under 150-200lbs (let alone 500!) 

I know mate, good point 

The larger the stack face/baffle the closer you get to a flush mounted speaker in 1pi and the lower the frequency that gets a boost up to 6dB. Then depending on how well you're able to actually couple the internal horn to the baffle, you'll see additional sensitivity gains (not sure what the theoretical maximum is here.) This is the real art of the design where apparently even Akabak doesn't quite get it right.

With that in mind, this design still has benefits over conventional FLH arrangements even if no additional coupling to the baffle is achieved. You can achieve deeper extension with less wood in the same size cab if you're willing to put up with a slight increase in group delay, and you can balance that by playing with the size of the new mouth (smaller = more extension and more group delay).

Love your work Stu, but I do think the current box is a bit half baked. 

The All_1 BC118 thread is also well worth looking through!

https://www.diyaudio.com/community/threads/boundary-control-bc-subwoofer-bc218-2-design.413237/post-8145708" rel="nofollow - https://www.diyaudio.com/community/threads/boundary-control-bc-subwoofer-bc218-2-design.413237/post-8145708