this system is 30" wide and 6 feet tall. for scale the orange woofers are 12 inches. the depth would be about 2 feet.
some parts are omitted as they wouldn’t help understand what is going on. so i am not showing any rear chamber enclosing the drivers or any kind of structural supports holding the horns in place. i am only showing the horns and the drivers - that’s it.
it will become clear what this is as you read along … and not gonna lie, the main difficulty with this system and why i probably won’t build it is i like right angles and this system doesn’t have a lot of them. every angle is different and this will require designing everything in solidworks and then leveling up my woodwork game to get all the angles to line up.
but basically the upper part is an “upper midrange” horn. the center part is three cell HF horn ( using 3 separate drivers ). the bottom part is 3 x 8" lower midranges ( purple ) and four 12" woofers ( orange ). i will go into all of this in more detail later.
i been designing and redesigning this particular system for like 5 years maybe more but now is the right time to post it
what happened is after i shelved this idea a few years ago i forgot all about it completely when i switched my focus to arrays. then recently i started designing again only to realize after the design was finished that i had the same exact design years ago. and now my mind is starting to shift to arrays again so before i forget this design again i must document it here.
OK so the part of the system you see in the sketch is strictly for operation above 80 hz. it will cross over to subwoofers at 80 hz. you may think a system this large should reach lower but that’s just because you’re stupid. this is the scientifically correct way to do it if you ask any genius with 20+ years experience in the field. i didn’t invent this - just applying best practices.
for example take JBL VTX F35:
is a large, expensive, high output system, and it has this response:
this system with the required amps and processing is over $10,000 and puts out 130 decibels yet doesn’t have the low end reach of a $100 bookshelf speaker. why ? because if you can afford to spend $10,000 on the top you can also afford to get a subwoofer - that’s why.
it appears the VTX was going for a 60 hz crossover point to subwoofer. the common crossover points in pro audio are 60, 80, 100 and 120 hz.
the common crossover point in home audio is 80 hz, because in small rooms bass standing waves dominate and even with the higher 80 hz crossover you can only localize the subwoofer by its harmonic and spurious noises ( such as port chuffing ) and not actual output.
thus our large main system will have a high 80 hz crossover - not because i can’t design it to go lower, but because i want it fully optimized and that means not trying to do things that the subwoofers will be doing anyway ( reproducing frequencies below 80 hz ).
by not trying to reach lower we can make the midbass more effortless and dynamic.
low frequencies are dominating by standing waves in a normal room and are best handled by a distributed array of subwoofers, NOT by the mains !
pushing frequencies below 80 hz to the mains is a straight up SIGN OF LOW IQ. i don’t care if this is offensive. you’re fucking stupid if you want the mains to go lower than 80 hz. those frequencies should be handled by subs and SUBS ONLY.
the crossover points will be 80 hz, 200 hz, 700 hz and 2.5 khz which means each frequency band is just under 2 octaves
for example 2.5 khz / 700 hz = 3.6 to find out how many octaves that is 2^X = 3.6 and just type that into google and AI solves for X = 1.85 so this frequency band is 1.85 octaves.
there are a total of 10 octaves from 20 hz to 20 khz and with 4-way + subs that’s 5 frequency bands that must on average be 2 octaves because 2 * 5 = 10. however because frequencies below 40 hz and above 10 khz are less important than frequencies between 40 hz and 10 khz the outer bands ( subwoofer and tweeter ) can be stretched a bit so the inner 3 bands are slightly narrower than 2 octaves and the outer 3 bands are slightly wider.
this is the orange driver in the sketch. they are arranged in a 2 X 2 array. i might separately discuss the intricacies of this layout later.
this driver is currently $400 from parts express.
this driver is sold as replacement part for JBL speakers and JBL doesn’t normally sell their drivers on open market, but this one is an exception … for the time being.
we can’t know how long it will be available, but the good news is in this design it can be replaced with any other good 12" prosound midbass, such as ones from B&C etc.
when checking TS parameters look for Xmax of about 5 to 10 mm for this application and EBP of around 150 hz, FS of around 50 hz and QTS of around 0.3 … those parameters would make it roughly suitable for 80 hz crossover point in a vented alignment. of course you then drop the driver into your box modeling software ( you should learn to use HornResp anyway, so make it HornResp ) and tune the port just below 80 hz to say 65 hz and adjust the driver choice and box volume until you get a nice roll off without any anomalies.
the 12" JBL 2262 HPL is really more of a starting point here than anything. it’s a proven driver used in many JBL speakers to cover the frequency band we’re covering here ( 80 hz to 200 hz ) but you can use any other suitable 12" driver …
it MUST be 12" though or it won’t fit the dimensions of the design … the idea is that two 12" woofers vertically are 24" and three 8" midranges vertically are also 24" so they fit into the same horn. this portion of a speaker is a bit like a Danley Unity Horn:
in the sense that drivers covering different frequency bands are sharing the same horn. but in this case instead of having a tweeter at the center there is an array of 3 midranges at the center. which has also been done before. take for example PD764i-215
it uses two 8" midranges and two 15" woofers on the same horn. my design is basically the same but it’s three 8" midranges and four 12" woofers.
danley puts the tweeter on the same horn. JBL puts tweeter on separate horn but mounts it inside the bigger horn. i put it on a separate horn and mount it outside of the main horn. the reason being that prosound speakers like Danley and JBL are meant to be stacked on top of each other, but home speakers are not. so i have basically reworked several designs from Danley and JBL into something actually optimized for home use.
in case of these 8 ohm drivers you would wire them in parallel for a total of 2 ohm, but if you are going with B&C for example choose the 4 ohm versions and wire in series-parallel for a total of 4 ohm which is the preferred amp ( amplifier ) load.
in my case because i’m using external subs with 80 hz crossover a 12" midbass was more appropriate than 15" woofers so instead of matching 2 X 8" to a single 15" vertically i am matching three 8" to dual 12" vertically. 3 X 8 = 2 X 12 = 24 inches.
also 3 drivers in parallel 8 / 3 = 2.7 ohms, which is actually ideal amp load. anywhere from 2 to 4 ohms is good for the amp, but 2.7 is ideal. for more sustained loads like subs you want to go a bit higher like 4 ohms to avoid tripping the amp. most amps are rated for 4 ohm power but will produce more power at 2 ohm and the most power at 2.7 … but for sustained loads you may not want to use that extra power and stick with rated 4 ohms to protect the amp.
you can use any other quality 8" prosound midrange here as well, such as ones from B&C. the Eitheen Sound is slightly nicer than some of the other ones but it’s not that important.
A GOOD DESIGN SHOULD TOLERATE SOME VARIABLITY IN COMPONENT QUALITY.
you may someday be forced to replace some of the drivers and it should not ruin the system overall.
this is the first controversial choice here. compression midranges have fallen out of favor because they are a bit more colored than paper midranges. however if you look at how a compression midrange was used in discontinued JBL cinema systems:
we are actually covering that range with two drivers - a lower midrange ( 8" cone ) and an upper midrange ( BMS MID ). so we aren’t quite doing the same thing as what JBL has discontinued …
in other words we aren’t quite repeating something that has proven to be a bad idea …
rather we’re operating this BMS mid in the same frequency range as RCF still operates their 4" compression drivers ( they run them down to 650 hz ).
the reason compression mids are no longer used is because to get low end you need a deep horn which not only makes the speaker deep but has a slow expansion rate which results in higher distortion. by not trying to take the driver to 300 hz ( which it is rated for ) we can use a shallower, faster expanding horn which is not only more practical but better sounding.
then we simply cover the 200 hz to 700 hz with the paper midrange.
the 4582ND is simply the popular BMS coaxial without the super tweeter section. it shares the same part number replacement diaphragm with the coaxials.
the 4599HE is a completely custom midrange with two opposing diaphragms ( different from the ones used in the coaxials ). that means that if you ever need to replace the diaphragm you better hope 4599 is still in production because it is an exotic made-to-order driver, which may very well be discontinued and then good luck finding diaphragms for it.
by contrast because 4582ND shares the replacement diaphragm part with something like 5 other BMS drivers, some of which are used in mainstream speakers such as QSC cinema speakers - you are almost guaranteed you will be able to find replacements.
furthermore to the diaphragm on 4582ND is basically a hat on the back of the driver. you don’t even need to remove the driver from the horn to replace it. on the 4599HE you have to remove the driver from the horn then disassemble it and then somehow put it back without damaging the diaphragm.
the procedure for replacing diaphragm is kind of a big deal. some drivers make it fool proof. with BMS 4599 it seems like it would be very difficult because the dual diaphragm assembly has voice coils on both sides and must be bolted in between two heavy motors without damaging the fragile coils.
whatever marginal sound quality improvement the 4599 may have over 4582 is not worth the risk of putting yourself into a difficult situation should you need to replace the diaphragm.
at this point i have to reiterate that it is the design that makes or breaks a speaker - not the components.
it is beginners and bad designers who try to compensate for lack of knowledge with using premium drivers.
top designers use fairly pedestrian drivers and then coax the desired response out of them.
if you were to swap the engines between Formula 1 and a Minivan and race them on the track the Formula 1 with a minivan engine would beat a minivan with Formula 1 engine.
this isn’t the best sounding tweeter on earth. nor even of its type. but it’s good enough and it has he practical benefits BMS is known for such as small size, light weight and easy diaphragm replacement ( same method as with the MID - diaphragm is replaceable without removing driver from the horn ).
again, it is the beginner / fool who thinks using the best drivers is what is going to make a great speaker. NO. you need drivers that are GOOD ENOUGH and then you need the BEST INTEGRATION into overall system.
if you want a slightly more Audiophile tweeter you can use Faital Pro HF106 instead:
there is a saying that all that is new is actually something that is old but well forgotten.
my design only seems impossible to idiots who don’t understand that it actually isn’t as innovative as it seems.
what my design really does is look at the entire landscape of all the exceptional loudspeakers ever made, study the principles of their operation and then combines various aspects of those designs into a single design using the latest drivers and latest understanding of physics of horns and arrays and does so in the context of creating a studio monitor quality speaker for home use that is able to effortlessly reach rock concert SPL levels, which, again are based on latest research into hearing damage ( there will be separate article on that ).
the reason this is the “baseline” system is because this is the minimum cost / size / weight system that i consider adequate in the sense that you will not wish you had more when you listen to it. instead your ears will give out before the system does all the while having the accuracy of a studio monitor ( which mostly comes from just DSP tuning ).
there will be a separate ARRAY system that be much heavier and more expensive than this but in some ways less compromised. it won’t quite be “cost no object” but rather it will be “performance over cost” whereas this baseline system is not quite “build to a price point” but rather “uncompromising performance at a competitive cost, size and weight”
then you turn it’s on the SIDE, so rotate 90 degrees so its vertical instead of horizontal. put 3 of them side by side and just cut off the parts of the flange that get in the way to construct your multi-cell horn ! it’s that simple !
just make sure the horn you’re buying has correct throat and bolt pattern for your HF driver and also that it is rated for the frequency range you will use. so in our case 2.5 khz to 20 khz and almost all horns are rated for this range. the more expensive horns can reach below 1 khz but we don’t need that here. a good practice is to use a horn rated 1 octave lower than what you will actually use so get a horn rated to go down to 1 khz or better, even though we will only use 2.5 khz.
as for the midrange horn these are not so widely available. BMS horn must be ordered from Europe and being fiberglass it could easily crack in shipping.
also the B&C is for 1.4" driver not 2" driver … although BMS mid also comes in 1.4" but i think maybe 2" is the better option for a midrange horn for both acoustical reasons and reasons of compatibility. you can adapt a 1.4" driver to 2" horn but not vice versa.
there is also SEOS 30 horn, which is made to order and ships from Poland
this is the horn Celestion used to demo their Axiperiodic driver. i have a distinct feeling this horn will cost a lot.
and of course Joseph Crowe will make a custom horn for you for about $5,000.
but i suggest you build the horn yourself.
since this is for a BMS driver i suggest looking at BMS horn as a starting point ( though you should also look at B&C and SEOS horns as well ) and then realize that the horn has two parts - the narrow throat part and the wide mouth part. what you want to do is 3D print the throat and make the mouth out of plywood.
3D printers are limited in SIZE of what they can print and also you can’t print something hanging in space - the material must be supported while it’s printed. basically you will have a hard time printing the mouth but you can print the throat.
my actual design is to print the throat in 2 parts and glue them together. specifically you would print the the mounting plate ( for the driver ) separately and glue it to the rest of the horn mouth. the mounting plate should be thicker in the middle to give plenty of area to glue to the throat. also the throat should be thicker there as well for strength.
you don’t need to buy a 3D printer - there are shops that will print your model. you just need to make sure the model is small and is a printable shape - that is something that starts out from a flat surface and builds up.
for example here Joseph Crowe is selling a file that can be used to 3D print a throat adapter:
but it can certainly be 3D printed, at least in two parts. each starting from a flat surface ( the opposite sides of the throat adapter ) and printing up and then they get glued somewhere in the middle. you will need to design the part where they glue to have enough surface and a shape that securely locks in - but that’s not hard.
then you just bolt the plywood mouth to this 3D printed adapter.
yes it’s work but it’s eminently doable and with a midrange horn performance is all about SIZE.
you will want the throat of the horn to roughly match that of BMS and B&C horns ( as those are specifically designed for the type of driver you will be using and will provide optimal loading for that driver ) but when it comes to the mouth - bigger is better - and SEOS is the absurdly large horn ( which is why it was used to demo the Axiperiodic ) and that’s the one you want to match on size of the horn mouth. it’s called “30” because it’s 30 inches in size.
at the beginning i mentioned this speaker is 30" wide - that’s why. but also 12" + 8" + 12" = 32" and when you account for the ANGLE that the woofers are installed at that gets the overall width of the woofer section just under 30 inches as well.
that’s what DESIGN means. it doesn’t mean buying the most expensive drivers. it means getting everything to work perfectly together. the speaker makes optimal use of space. of course i have also matched various acoustical characteristics but if you want to know that you have to join the site and ask because it’s too much to write up unless somebody is specifically looking for that information.
but yes i have considered driver frequency response, power handling, distortion characteristics and directivity patterns of drivers and horns - and everything is optimized ( of course to really optimize it you need to order samples of drivers and measure them and then choose the ones that perform best, rather my calculations are only to get in the ball park from where you can start that process ) you can join the site and we can talk about it in depth. here i used an example of DIMENSIONS matching because it’s something everybody can understand ( hopefully ).
and as for why the HF section is in the middle and not on top of midrange section it’s because it’s much less tall. the midrange horn is about 24 X 32 inches or so in mouth dimensions. the HF horn is probably about 10 x 20 inches. it’s a lot shorter vertically so it’s easy to slot it in there without pushing the midrange horn up much. on other hand if midrange horn came first it would push the tweeters way up to above ear level.
again, i didn’t invent this. many ultra high end speakers are designed this way:
you try to keep drivers close to ear level but they can’t all be there so you prioritize the smaller drivers because they don’t push other drivers as far out from ear level.
it’s all genius trust me. there is reason for everything in this design. none of it is random. everything is calculated. if you want to know all that stuff join the forum and ask - i will be happy to explain.
it’s just that there is hundreds of hours of considering all sorts of factors that went into this design and it only makes sense to discuss those details after somebody asks about them specifically. without anybody asking it only makes sense to lay out the general design as i have above.
but yeah i’m probably the best designer in DIY audio. possibly better than Tom Danely. they always ban me when i go after Danley so we never will find out who is best. there are also certainly better designers than me at JBL and maybe L-Acoustics, and that’s why i have studied pretty much all JBL designs ever made - not so much any Danley designs, which for the most part are just inferior versions of older JBL designs anyway.
outside of arrays, this “baseline, minimum driver” system is probably the best sound system for home use you can build or buy.
YES ARRAYS CAN SURPASS THIS SYSTEM, but they will be larger, heavier and more expensive … and in any case i will design one soon.
i was always a HUGE fan of arrays ! but the $$$ factor with arrays is real. otherwise i would have had one 20 years ago already !
this “baseline, minimum driver” system is going to get you close to Array performance level at a significantly lower cost. this is why i designed it and why i am sharing it.
but personally to be honest i would go the extra mile and build the array …
but some people ( like Romy the Cat ) simply like horns !
also arrays are so perfect they can be a bit boring. kind of like Electric Cars are so perfect they’re boring. no noise, no vibration, no gear shifts, no tire slip - nothing. you just press the pedal and it goes. some people want more personality from their car !
studio monitors are also close to perfect and many people don’t like them because of that as well.
this “baseline minimum driver” system will have more personality than an array system, both visually and acoustically. it won’t look or sound boring. if you want to build something unique yet designed around sound principles of physics and speaker design and has incredible performance - build this speaker.
if you don’t understand some of the philosophy behind it - just ask.
replace the array of three tweeters with a single JBL D2 driver:
i didn’t originally want to do this for number of reasons but it’s just an absurd value. for $420 you get a driver that is double the performance of $700 drivers.
when i originally saw it i thought it was some kind of mistake. i couldn’t believe they were selling it at this price. i thought it has to be some other D2 not the one that’s used in all the best speakers in the world. but it seems it’s the same one.
with a Single D2 replacing three Faital or BMS drivers a $420 driver replaces three $260 drivers and now you can use a regular off the shelf horn for the D2 instead of having to design and 3D print custom ones. you also avoid any kind of summation issues from 3 separate horns.
you know what - before unveiling the array system i will do an intermediate system that will be an improved version of this one ( to be presented in a separate thread )
it will consist of:
8+ khz: 4 X Beyma CP12/N SuperTweters ( vertical array nested in midbass array, horizontal diamond directivity )
