I hesitate a bit to turn this thread into something that the OP probably does not want: a debate (i.e., win/lose zero-sum game arguments) over MEH design. I'd like to think that there are reasons why you would do each type of MEH build. I wish that the opinions-stated-as-facts on any design thread (especially on this subject) could be turned into system trade-off discussions (not debates) that include things like explicitly stating "what is the objective?" in the discussion before the winners and losers are picked.
The reason why I interjected above was to illuminate a bit more on what I read to be a rather sweeping generalization that I feel didn't point out the pros/cons of using a particular type of MEH design. [The real reason why we have trained engineers in this world (and not just opinionated "physicists"--in the strict meaning of the phase) is because engineers formally are taught to do trade studies and have existing decision criteria for each alternative approach to design.]
My objectives in MEH design are to provide something that beats the sound quality of anything else on the market--including extremely expensive "boutique loudspeakers" that sometimes are being sold for $100Ks each. (Yes, I really did just say that.)
If you simply want "MEH using cheap drivers", then you can certainly go that route, i.e., using a 1" compression driver with two/four/six midrange cones, then smaller woofers--like 12" woofers. That was never my intent, and having so many midrange drivers is really a design choice based on a) using passive crossovers that can't do level balancing and EQ very well, and b) have the ability to run continuously at 120+ dB/1m was never of any use to my needs.
I simply found that the "50 or 60 degree coverage horizontally, three-way, four cone midranges, passive crossover MEH design that produces 120 dB/1m" wasn't my preferred solution, since I was actually looking for something having higher subjective sound quality for home hi-fi use (i.e., not specifically 120+dB output PA use that the Danley SH-50 and SH-60 models were designed to do). I really wasn't interested in using passive crossovers, but rather DSP crossovers which have superior PEQ capabilities, which turns out to be a real discriminator for MEH design, in my experience (although simple first order passive crossover filters can be used, but the EQ that comes with a DSP crossover is also necessary for home hi-fi use). I could actually hear the tradeoff in the midrange of the SH-50 by using four cone midranges and passive crossover/balancing network vs. the K-402-MEH. It wasn't really that subtle, but rather was quite obvious within the first few seconds of listening to the two design approaches in A-B fashion in-room.
Like Mark said above, a natural set of horn coverage angles seems to have a sweet spot at 90 x 60 degrees, with the 90 degree coverage direction horizontally oriented in-room. (The 60 degree coverage direction I find works much better in the vertical direction to control floor and ceiling bounce.) The in-room coverage of the horn I find needs to be able to seamlessly cover the entire front of the room, without gaps or sensitivity to head-in-a-vice use.
If you are really obsessively worried about power response into the room instead of prioritizing subjective listening trial horn coverage requirements, then you can go to a square horn design (i.e., 90 x 90), but note that this has consequences for the compression driver up high--having to cover more polar area than what is actually needed in-room (IME), and it puts a lot more acoustic energy on the floor and ceiling than is really needed.
I see no compelling reason to use circular or elliptical horn designs. They are difficult to do well, and I hear no difference in their sound quality (even at high in-room SPL).
I'll stop there, for now. I think even the most inexperienced consumer can see that there are real tradeoffs to MEH design based on in-room sound quality needs, costs, and other "tenuously applicable" design constraints.
Chris
One more comment that I don't presently intend to press any further (unless I'm pressed here on this subject): to submit yourself to a DIY MEH build but then make the horn mouth size so small that it doesn't control its polars below 500-1500 Hz...I find is a real crime. You're giving up the biggest single advantage of the MEH design in order to have something that "looks smaller", and you're basically giving away the whole enchilada by this one "ill advised" design choice.
Note that I have no plans to apologize for what I just said--just above--because I've heard the difference (i.e., using much smaller horns that "look nicer") and it's not a subtle difference in terms of sound quality...
Chris
Note that I have no plans to apologize for what I just said--just above--because I've heard the difference (i.e., using much smaller horns that "look nicer") and it's not a subtle difference in terms of sound quality...
Chris
This is a tricky one, and I'm beginning to come around to the mindset that "there's no such thing as too much headroom."
For about 5-10 years worth of Unity horn experiments, I began to have a preference for planars, dome tweeters, and ribbons on Unity horns, instead of compression drivers. This was mostly because the "average" one inch compression drivers lacked some "air" in the top octiave, and the compression drivers that could play past 20khz couldn't reach down to 1khz. So it was a catch-22: a compression driver like the B&C DE250 didn't sound as "airy" as a dome tweeter, and a compression driver like a BMS 4540ND DID sound about as good as any hifi dome, but COULDN'T play low enough.
I think a couple of things started to sway me back to MODERN compression drivers:
1) I made a LOT of waveguides with the SB Acoustics SB26, and every time I made one, there was just something "off." It's strange because they measure quite decent, and the tweeter sounds marvelous on a small waveguide or a flat baffle... but put it on a decently large waveguide and it starts sounding a lot like a B&C DE250. There's just something screwy about the top octave.
2) The thing that REALLY cemented my thoughts, was having the Yamaha DXR12 and the Waslo Cosyne running in my house at the same time. I had the Yamahas in my office, and the Cosynes in my living room. Both use the exact same compression driver, an inexpensive 1" unit from Celestion. The treble on these two speakers sounds absolutely nothing alike, despite using the same compression driver. I've measured both and published results from both, so folks are welcome to look up the data. (I'd do it myself but I'm on vacation and the fam will kill me if I spend all day posting on diyaudio.) My hunch is that the first 2-3 inches in the waveguide of the Yamaha is a really great match to the Celestion compression driver, and this is why the treble from the Yamaha DXR12 sounds about as good as any hifi dome. Particularly remarkable when you consider the compression driver is about $45! DXR12 doesn't sound great, out of the box, because the treble is too hot and the bass is boomy. But it has headroom to burn, and once you EQ it carefully, I'd say it sounds about as good as the average $1000 audiophile speaker... it just happens to be self powered with 1000 watts on tap and can play 20X as loud as an "audiophile" speaker.
Long story short: I'm coming back around to using compression drivers on Unity horns again, but I was off the reservation for a few years there.
I really wish I had time to run and post some hornresp sims, you can simulate what I am about to describe:
Horns are obviously "louder" than traditional tweeters. Something that wasn't so obvious to me, for a while, is that a lot of the additional output is simply because the energy of the tweeter is being focused into a narrower beam.
The first time this really hit me in the face, was with my experiments with the Danley Paraline. Those things are freaken LOUD LOUD LOUD. I had paraline horns that were about 5-10dB louder than "conventional" horns and I was trying to figure out where all the extra output was coming from. The horns had similar dimensions, the horns used the same compression driver... the Paralines were just dramatically LOUDER.
Then it hit me: this thing is like a laser. By preventing the vertical beam from expanding more than a few degrees, even meters away, you get crazy amounts of output.
This is one of those things that we don't see much of in audio, because even with ten foot tall line arrays, we're basically "throwing away" a lot of that output. In a conventional line array, we are taking a sound source that is spherical, putting it next to another sound source that is spherical, and then trying to approximate something that resembles a line source.
Paraline don't do that - it's an actual line source. The wavefront isn't perfectly flat, but it is much much much narrower (vertically) than anything you'd ever see from a radiator that's dome shaped.
Once I realized that, I began to appreciate the potential of true line sources, like planars and ribbons. But also appreciated that once they're playing a frequency that's larger than their dimensions, they blow up in a hurry!!!
For instance: 4500Hz is three inches long. So if you're playing any frequency above 4500Hz on a ribbon or planar that's three inches tall, it will be surprisingly loud, at surprisingly long distances. At the Hoover Dam in Nevada, they use BG ribbons on their PA speakers, and even while standing in one of the largest rooms in America, the sound is impressively clean and loud. Truly remarkable for a $50 tweeter, I don't think there's a dome that can do this. BUT - those speakers are pointed like a laser at where the folks on the Hoover Dam tour are standing. Walk ten feet off the path and the sound goes away. (Narrow beam.)
Where people get into trouble with ribbons and planars is that they try to push them lower than the dimension of the ribbon. For instance, if you play 2khz on a planar that's three inches tall, the wavefronts that measure seven inches in diameter (two khz) are no longer flat Depending on the shape of the waveguide or baffle, the wavefronts at 2khz can be expanding dramatically.
Again, I'm kinda slow to figure things out, but it was one of the things that baffled me about AMTs, and I blew up a planar trying to get it to play low. The diaphragm itself exploded into bits of mylar. If you look up the response graphs of the Airborne AMTs, you'll see what I describe. There is a 'knee' in the response. The 'knee' is caused when the wavefront begins to expand spherically, because the wavelengths are now longer than the diaphragm itself.
Circling back around to "the impact on Unity horns", the issue with Unity horns with wide beamwidth is that as the beamwidth gets wider and wider, the output potential on the low end of the tweeter's passband is less and less and less. This manifests itself as a dip in the xover region especially if you try and push the tweeter down to 1200 or 1000Hz. All of this is exacerbated by the fact that as the angle of the walls gets wider and wider, the midrange taps get further and further apart.
In a 90 degree triangle where the midrange taps are 3.5 inches from the throat, the midrange taps are 4.95" away from each other, left to right. In order for the midranges to behave as a virtual unit, they need to be within about one third of a wavelength. 909Hz is 14.85 inches long.
This means that in a Unity horn with a beamwidth of 90 degrees, where the midranges are 3.5 inches from the throat, the midranges are behaving as a virtual unit up to a frequency of about 909Hz.
So you wind up with a conundrum:
1) you can make the beamwidth narrower
2) you can use a compression driver that can play down to 900Hz without breaking a sweat
3) you can move the midranges closer to the throat, but that requires using small midranges.
It's tricky right? This is one of the reasons I've made so many Unity horns. It's like a puzzle with endless possibilities and all of the compromises are interesting. If you want to make a Unity horn with "airy" treble, you can do that, but then you hit a brick wall when you try to use a waveguide with a wide beamwidth. If you want to make a Unity horn with "airy" treble, and you still want to have wide beamwidth, you can TRY and make a waveguide with very narrow vertical beamwidth... but that is extremely difficult to do. I can't even count the number of waveguides I've made where I tried to get a horizontal beamwidth of about a hundred degrees and a vertical beamwidth of about 40. To me that would be The Ultimate Solution but it seems nearly impossible. I invite people to try and do it in ATH, in my own experiences, it seems likely impossible. (I haven't lost all faith yet, I think it may be possible by using a VERTICAL array to achieve the 40 degree vertical beamwidth, combined with a horizontal waveguide to achieve 100 degree horizontal beamwidth. Some other folks have posted projects like this, I don't think anyone has ever finished one. The closest thing to what I describe is the project from Follgott.)
I know this post sounds unusually meandering, even for me, but lately I've been focused on the idea of taking the music itself, converting it from two channel to multichannel, and THEN reproducing the music over narrow beamwidth speakers. The fact that I can see the Vegas sphere (with it's laser focused beamsteered array) from my backyard may have inspired me lol
It's both. The geometry rules of a Unity horn dictate that you'll get a dip in the midrange if the angle of the walls is wide. Beefy compression drivers and coaxial compression drivers "fix" the dip because they can be crossed over lower. This allows the overall output of the speaker to be maximized.
Alternatively, one could use a "conventional" compression driver, add boost to fill in the dip, and achieve flat response. But every decibel of boost you use to fill in the dip is a decibel of output that you're throwing away overall.
Come to think of it, I guess one solution that would be Truly Epic would be to set up EQ curves that are dependent on output level. I believe JBL does this with a lot of their powered speakers. Basically you would have one curve for an output level of 90dB, another curve for an output level of 100dB, another curve for an output level of 110dB, etc. This would allow you to use just about any level of boost you'd like (within reason) at low to moderate levels of output. It would mean that when the speakers are blasting, there would be a dip in the midrange, but if you're "only" listening at 95-105dB, the response curve would be flat.
This would allow for the use of relatively small/cheap compression drivers on a Unity horn AND wide horizontal beamwidth.
I have no use for wide vertical beamwidth, but it's extremely difficult to get wide horizontal beamwidth and narrow vertical beamwidth from a tweeter that produces a spherical wavefront. This is a large part of the reason that my "Square Pegs" thread on Paralines has been ongoing for over ten years now. If there was a tweeter that had a diaphragm that measured about seven inches in height and about 0.75" in diameter, that would solve nearly all of this. AMTs are about as close as it gets, but their width is too wide, and wide width makes them beam above 5khz or so.
If the Paraline or V-DOSC worked exactly as they're supposed to, all of this would be moot, it would arguably be the ultimate tweeter for a home Unity horn.
I'd be very surprised if any A For Ara speakers were ever for sale on Etsy.
There is a range of the monolith speakers now available that use a nominal 90° waveguide.
An interesting one will be ready this coming summer. The waveguides main axis is aimed 30° from straight ahead, so I've been able to narrow the pattern and the speakers don't have to be towed in.
That's been my mindset in all my builds.. to the point of making sure every driver section has sufficient excursion capability at the low end of its range, to handle +15 to 20dB peaks above average SPL, without exceeding xmax. Amp power/voltage per driver section, has to be sufficient as well to achieve that headroom, ime.
(I'm convinced clean headroom is one of the greatest deficiencies in typical home audio....particularly for passive speakers.)
Understandable. 🙂
For over 3 decades, electrostats, planars, tall ribbons, etc...were my home audio passion. Still have 2 pairs of full-range stats I can't bear to part with.
That's one of the first things I learned when dabbling with live sound....how the horn/waveguide pattern increases SPL by confining its radiation. Sound guys often describe the phenom as 'throw', with narrower H-V patterns having longer throw. Probably not the best term, but folks know what it means.
I think at its simplest, a horn does for the high end, the same thing boundaries and walls do for the low end. Confine radiated power.
Yep, true lines do focus vertical pattern like crazy, given their minimal c2c spacing of VHF producing areas.
Now, how the heck to get large planar membranes, like in electrotats to be able to make enough bass....with as you say going kaboom...(aka arcing or such)
As you say, a wider horn forces them closer to throat.
I've found it's no problem for 4" mids and 90 degree horns to get closer than 3.5" if ports are located a little towards cones' surrounds . And I think good relatively inexpensive CDs are around than can reach 1000Hz, so I see a 90 degree syn as very doable economically.
Hey, enough of this audio stuff...enjoy a fine family vacation ! And Thx for the detailed reply.
I second that experience and opinion.
If pattern control can't be achieved to at least 500Hz, I can't see any reason to build a unity/synergy. Which means for 90 degree horiz, a width of 22". And I'd call that marginal.
400Hz, and 28" starts to make sense to me, and it just gets better from there if large size permits.
For a narrow horn, say a 60 degree H .....500Hz needs a 33" width. And 400Hz needs a whopping 42" width.
Forget narrow, lol.
I've been really happy with 90 x 60, 48" wide, and pattern to 230Hz. But would go bigger if I could 😀
Hi Andrew
I have started to build a pair of MEH´s this summer. I call them "Le Grande MEH Danois". I am far from finished, but I will start a thread soon. Maybe my story can inspire you?
I have been collecting and preparing for years actually. Reading through a lot of threads here on diyAudio and also on the Klipsch forum, following the writings of members that have already contributed in this thread. I am particularly inspired by Mark100´s builds and Cask05´s K402MEH.
I have kind of made some basic decisions and started from there:
I wanted some really big MEH´s to control directivity as far down in frequency as possible/reasonable in my room
132cm/52"Hx98cm/38,5"V
80Hx60V, ratio 1,35:1
corner-placement up against the sidewalls (as Tom Danley did with his SH50´s in his living-room back in Illinois)
secondary flares with rounded edges (approximating some tractrix flares)
1,4" compression-driver (Faital Pro HF1440), that can be upgradet to a BMS 4594HE if I want that in the future. 1,4" is a sweet-spot/good compromise I think
midranges (4x4NDF34 16 ohm as Mark uses)
I start with two old 15" woofers I have (Visaton BGS38), but the MEH can be upgraded with four 10" og 12" woofers if I want that in the future
midranges and woofers mounted on top/bottom panels that can be changed, modular design!
15mm plywood for the horn and 12mm plywood for the enclosure. I found some WISA TWIN that had an OK price, as my MEH is modular I can upgrade to birch-ply.
DSP crossovers, I have a bunch of MiniDSP plate-amps with DSP.
I have decided to start with at square throat, to get started! In the future I can upgrade with a 3D printed round to square throat adapter du to my modular design.
After many many hours of contemplating (and obsessing over details), I have come up with a somewhat untraditional approach of building the MEH upside-down, starting with the secondary flares, to be able to make a modular design and round edges. More on that in my own thread. It has turned out to be a rather complicated design, but that always happens to me!!
I would probably not go that route again! I think the way Mark100 does his SYN-builds is easier.
Then I used Bill Waslo´s Synergy Calc to get some plans to start with. I actually had to recalculate the initial horn, as my prototypes had shown, that synery-calc computes som minor errors (39,78 degrees instead of 40 and 29,78 instead of 30). I choose to recalculate on paper and made my own plans. The secondary plans are build according to synergy-calc. That´s just me and my perfectionism!
BUT FIRST I made some smaller "prototypes" 73cmx54cm out of scrap material I had at hand to learn and make my mistakes (most of them!) before I started on the big ones!
It is a bit like solving a SODUKO actually, find somewhere to start an go on from there. There are many choices to make! You have to find your own way I guess.
I have an idea of where the midrange- and woofer-tabs have to go. But if I do not nail it in the first try I can just make a new panel and try again. Modular design!
The MEH´s will be placed in my room-corners on some double 18" subs. They will actually occupy less florespace then the speakers I have now!
A year ago I made a muck-up wooden frame out of some old boards, 150cm/59"x110cm/43", to find out how big a MEH I could build. Then my wife and I looked at it for a while and decided to shrink the MEH 12% to the size in the picture above.
This is what I have done/am doing. Hope that helps. I guess I will need some help from the diyAudio-community when I have my "hardware" finished.
Happy new year to everyone
Regards from Denmark
Steffen
I have started to build a pair of MEH´s this summer. I call them "Le Grande MEH Danois". I am far from finished, but I will start a thread soon. Maybe my story can inspire you?
I have been collecting and preparing for years actually. Reading through a lot of threads here on diyAudio and also on the Klipsch forum, following the writings of members that have already contributed in this thread. I am particularly inspired by Mark100´s builds and Cask05´s K402MEH.
I have kind of made some basic decisions and started from there:
I wanted some really big MEH´s to control directivity as far down in frequency as possible/reasonable in my room
132cm/52"Hx98cm/38,5"V
80Hx60V, ratio 1,35:1
corner-placement up against the sidewalls (as Tom Danley did with his SH50´s in his living-room back in Illinois)
secondary flares with rounded edges (approximating some tractrix flares)
1,4" compression-driver (Faital Pro HF1440), that can be upgradet to a BMS 4594HE if I want that in the future. 1,4" is a sweet-spot/good compromise I think
midranges (4x4NDF34 16 ohm as Mark uses)
I start with two old 15" woofers I have (Visaton BGS38), but the MEH can be upgraded with four 10" og 12" woofers if I want that in the future
midranges and woofers mounted on top/bottom panels that can be changed, modular design!
15mm plywood for the horn and 12mm plywood for the enclosure. I found some WISA TWIN that had an OK price, as my MEH is modular I can upgrade to birch-ply.
DSP crossovers, I have a bunch of MiniDSP plate-amps with DSP.
I have decided to start with at square throat, to get started! In the future I can upgrade with a 3D printed round to square throat adapter du to my modular design.
After many many hours of contemplating (and obsessing over details), I have come up with a somewhat untraditional approach of building the MEH upside-down, starting with the secondary flares, to be able to make a modular design and round edges. More on that in my own thread. It has turned out to be a rather complicated design, but that always happens to me!!
I would probably not go that route again! I think the way Mark100 does his SYN-builds is easier.
Then I used Bill Waslo´s Synergy Calc to get some plans to start with. I actually had to recalculate the initial horn, as my prototypes had shown, that synery-calc computes som minor errors (39,78 degrees instead of 40 and 29,78 instead of 30). I choose to recalculate on paper and made my own plans. The secondary plans are build according to synergy-calc. That´s just me and my perfectionism!
BUT FIRST I made some smaller "prototypes" 73cmx54cm out of scrap material I had at hand to learn and make my mistakes (most of them!) before I started on the big ones!
It is a bit like solving a SODUKO actually, find somewhere to start an go on from there. There are many choices to make! You have to find your own way I guess.
I have an idea of where the midrange- and woofer-tabs have to go. But if I do not nail it in the first try I can just make a new panel and try again. Modular design!
The MEH´s will be placed in my room-corners on some double 18" subs. They will actually occupy less florespace then the speakers I have now!
A year ago I made a muck-up wooden frame out of some old boards, 150cm/59"x110cm/43", to find out how big a MEH I could build. Then my wife and I looked at it for a while and decided to shrink the MEH 12% to the size in the picture above.
This is what I have done/am doing. Hope that helps. I guess I will need some help from the diyAudio-community when I have my "hardware" finished.
Happy new year to everyone
Regards from Denmark
Steffen
Fantastic timing of this topic!
I will probably make the first cuts of my K-402 horn this week. As inspiration @Cask05 and Stabme's design, with some modifications here and there..
Only in terms of woofers, I still haven't decided. I definitely believe in Chris his approach with dual 15" Crites, as I really like the effortless sound of these big units! But I read more often that 15" woofers sound somewhat colored between 300 and 600hz, Troy said the same in one of his recent videos as well.
Perhaps 4pcs of 10inch per loadspeaker is an option? Would this sound as effortless and relaxed as dual 15", but with the refinement of smaller woofers?
I am also wondering what to look for in terms of woofer and CD impendance value and how to add this sums up, keeping in mind that I am using a tube amp of about 10watt?
I will probably make the first cuts of my K-402 horn this week. As inspiration @Cask05 and Stabme's design, with some modifications here and there..
Only in terms of woofers, I still haven't decided. I definitely believe in Chris his approach with dual 15" Crites, as I really like the effortless sound of these big units! But I read more often that 15" woofers sound somewhat colored between 300 and 600hz, Troy said the same in one of his recent videos as well.
Perhaps 4pcs of 10inch per loadspeaker is an option? Would this sound as effortless and relaxed as dual 15", but with the refinement of smaller woofers?
I am also wondering what to look for in terms of woofer and CD impendance value and how to add this sums up, keeping in mind that I am using a tube amp of about 10watt?
Hi Horntube90
Basically direct-radiation-logic does not work for synergies! New set of rules!
One of the advantages of the synergy-design in respect to woofers (and midranges) is, that the air trapped between the woofer-cone and the horn-wall in combination with the porthole creates a bandpass-filter. Actually the bandpass-filter reduces distortion from break-up and higher harmonics that were not in the original signal! Cleaner sound!
So the response of the woofer is attenuated below and above the pass-band. You can´t play as low and as high as with the same woofer as direct radiating woofer. That means, that a woofer in a synergy sounds completely different then the same woofer as direct radiating. Some people say that you do not need expensive woofers on synergies, due to the bandpass-filter. That is what Mark100 says in post #18.
Are you saying you want to make a MEH/Synergy-horn with passive filters, so that you are driving it with one amplifier only? The K402MEH by Chriss is a two-way active speaker with DSP!
Steffen
Basically direct-radiation-logic does not work for synergies! New set of rules!
One of the advantages of the synergy-design in respect to woofers (and midranges) is, that the air trapped between the woofer-cone and the horn-wall in combination with the porthole creates a bandpass-filter. Actually the bandpass-filter reduces distortion from break-up and higher harmonics that were not in the original signal! Cleaner sound!
So the response of the woofer is attenuated below and above the pass-band. You can´t play as low and as high as with the same woofer as direct radiating woofer. That means, that a woofer in a synergy sounds completely different then the same woofer as direct radiating. Some people say that you do not need expensive woofers on synergies, due to the bandpass-filter. That is what Mark100 says in post #18.
Are you saying you want to make a MEH/Synergy-horn with passive filters, so that you are driving it with one amplifier only? The K402MEH by Chriss is a two-way active speaker with DSP!
Steffen
From my experience with Cask05's K402 based design - simplicity is best for DIY. Therefore good drivers make life simpler and few drivers make a successful outcome easier.
If you have unlimited time to experiment with cheaper drivers and find that entertaining that is your path. I have found that considerable time is spent getting the crossover correct, even with DSP, and the AXIperiodic and Crites woofers. So to get to something you can begin to enjoy in a reasonable amount of time - go with Chris A's formula.
With REW the two CRITES woofers distortion is lower than the CELESTION's in the same part of the midrange. I can assure you will will not hear distortion in the upper reaches of those woofers. The distortion will be in the lower reaches where those woofers must be boosted since the K402 loses its transformative qualities below 100 Hz. Nonetheless, I found that even with this higher measured distortion the speaker sounds infinitely better being used below 100 Hz than trying to get a subwoofer to reach up there. I am using a sub (four RYHTMIK 15" sealed) for 20 to 50 Hz . Using PEQs I get an almost brick wall low pass filter that stops attenuating around 500 Hz. For the MEH a much more gradual slope gives a smooth response between the two - again assembled with PEQs - no filters that continue to fall into infinity.
The speakers do make demands of the room that must be cured with treatments and judicious DSP.
When it comes together it is good like full range electrostatics but better since there are no dynamic limitations in comparison to any electrostatic I have ever heard. Even from my listening chair a step response that looks like one driver (of course, this is without the subwoofer). It sounds as clear as that measurement would suggest it is.
Good thing about DIY is you can spend money on good drivers. The CELESTIONs are worth every penny. The CRITES were not expensive but no longer available. I know I would have liked to use two 16 Ohms drivers in parallel instead of the two CRITES in series. I would defer to Chris A for what to use instead.
If you have unlimited time to experiment with cheaper drivers and find that entertaining that is your path. I have found that considerable time is spent getting the crossover correct, even with DSP, and the AXIperiodic and Crites woofers. So to get to something you can begin to enjoy in a reasonable amount of time - go with Chris A's formula.
With REW the two CRITES woofers distortion is lower than the CELESTION's in the same part of the midrange. I can assure you will will not hear distortion in the upper reaches of those woofers. The distortion will be in the lower reaches where those woofers must be boosted since the K402 loses its transformative qualities below 100 Hz. Nonetheless, I found that even with this higher measured distortion the speaker sounds infinitely better being used below 100 Hz than trying to get a subwoofer to reach up there. I am using a sub (four RYHTMIK 15" sealed) for 20 to 50 Hz . Using PEQs I get an almost brick wall low pass filter that stops attenuating around 500 Hz. For the MEH a much more gradual slope gives a smooth response between the two - again assembled with PEQs - no filters that continue to fall into infinity.
The speakers do make demands of the room that must be cured with treatments and judicious DSP.
When it comes together it is good like full range electrostatics but better since there are no dynamic limitations in comparison to any electrostatic I have ever heard. Even from my listening chair a step response that looks like one driver (of course, this is without the subwoofer). It sounds as clear as that measurement would suggest it is.
Good thing about DIY is you can spend money on good drivers. The CELESTIONs are worth every penny. The CRITES were not expensive but no longer available. I know I would have liked to use two 16 Ohms drivers in parallel instead of the two CRITES in series. I would defer to Chris A for what to use instead.
Hi Steffen,
I have read up on the synergy horn theory again and I think you are right. The reason I am possibly looking at woofers other than the Crites is because there is quite a bit of shipping, tax and import duties involved considering I live in Europe. But the woofers with proper TS parameters are limited, having searched the database: 15" database with correct specs
I was hoping to buy similars 15" woofer at a local store, that saves over 500$ extra cost. I could then invest this amount in perhaps "even better" woofers at a local store..
But considering it will be hard to find a comparable one with the right specs, it will most likely just be the Crites. 🙂
Considering I am very happy with my current single ended tube amp and dac with tube rectifier and tube output stage, I would prefer to stay with 1st order passive x-over for now. I am aware that Chris uses DSP and have discussed with him whether the design can be reaslized with passive x-overs.
He did mention that it will be necessary to use some EQ in the digital domain, which I am totally okay with.
In the future I love to want to experiment with active x-overs and bi-amping! I see some great benefits here.. But for now, I still find it hard to let go of the magic of a SET amp.
Thank you @rickmcinnis for you're suggestions!
I would think a horn with constant directivity and pattern controll up to about 170hz is less dependent on room interactions. I would imagine if you place them in the corners or against a wall the first reflections points are crucial to deal with or are you referring to this?
Could you minimize this by choosing a horizontal coverage of 80degrees instead of 90degrees?
Can you give some hints on where to pay extra attention regarding DSP / EQ?
I would think a horn with constant directivity and pattern controll up to about 170hz is less dependent on room interactions. I would imagine if you place them in the corners or against a wall the first reflections points are crucial to deal with or are you referring to this?
Could you minimize this by choosing a horizontal coverage of 80degrees instead of 90degrees?
Can you give some hints on where to pay extra attention regarding DSP / EQ?
Hi Horntube90
First of all, I am not sure you can make a K402 MEH without DSP! It will not be easy!
I believe Chris uses a lot of PEQ´s. And if you choose to use DSP, then you can use any woofers that are somewhat comparable, you just use a different set of PEQ´s. THAT IS WHAT I THINK without being an expert. I just use some old 15" woofers that I have, and see where it gets me.
I think, that if you use a MiniDSP 2x4 HD you could use the SET amp for the compression-driver and a cheap class D or AB amplifier for the woofers, with all the freedom of the DSP and the most magic of the SET-amps!
I hope you make your own thread to let us follow your project 🙂
Steffen
First of all, I am not sure you can make a K402 MEH without DSP! It will not be easy!
I believe Chris uses a lot of PEQ´s. And if you choose to use DSP, then you can use any woofers that are somewhat comparable, you just use a different set of PEQ´s. THAT IS WHAT I THINK without being an expert. I just use some old 15" woofers that I have, and see where it gets me.
I think, that if you use a MiniDSP 2x4 HD you could use the SET amp for the compression-driver and a cheap class D or AB amplifier for the woofers, with all the freedom of the DSP and the most magic of the SET-amps!
I hope you make your own thread to let us follow your project 🙂
Steffen
There is another option as well - and I am very happy with it - to use a hybrid approach - passive crossover to protect the drivers (I used a simple 1st order series crossover for my "Kallax" MEH) and equalize via DSP - and a SET amplifier cam be used - I use one, too.
If you use tube (valve) electronics, output tube failure modes with residual DC current directed through the drivers is a real and finite concern. (With almost any SS amplifier, not so much.) Blocking caps work well in these cases, but making sure that they have no other unintended effects is something to pay close attention to. Since I do not use tube electronics, I do not use blocking caps. I've never lost a driver thus far (and I've been building my own since 1977).
The fast limiters found in DSP crossovers beats any passive network in terms of transient (non-DC-mode) overload protection. I strongly recommend using these limiters if you are prone to overdriving your tweeters, woofers, etc. They work well, and they come along for the ride with any DSP crossover I've seen.
For instance, the old Klipsch K-77 tweeters in their Heritage loudspeakers (an old University Sound T35 tweeter--no longer available) can only take ~3 watts long term, and is extremely easy to fry the voice coil/diaphragm in setups sporting "hundreds of watts" amplifier power capability. One of the first guys I helped years ago to dial-in his surround Belles in a 5.1 setup that he tri-amped using a DSP crossover was popping K-77 tweeter diaphragms. I helped him set the fast limiter in his DSP crossover (IIRC, a Xilica XP series) and his tweeter diaphragm problems immediately ceased.
I tell you this story because I found quite a few naysayers on the old K-forum that thought the fast limiters in DSP crossovers didn't work very well. They work excellently--better than any passive protection circuit I could otherwise think of.
This is true. On MEHs, I'm mostly using the PEQs and bandpass outputs that come with them...not so much the crossover filters themselves.
And I wouldn't choose to use a DSP crossover with less than 10 biquads per channel if I could avoid it (in addition to the same number of input channel PEQs). That's the level of PEQ use that I find occurs most often in MEH and other horn loudspeaker work. What you get back in return for selecting DSP-enabled/multi-amped MEH designs and other horn-loaded loudspeaker designs...far more than makes up for the bother.
Trying to retain passive crossover filters in order to mono-amp your loudspeakers with your "favorite SET amplifier" is not a good trade for home hi-fi use, I find. You lose in terms of sound quality, and you lose again in terms of using the power available from the amplifier--much of which is thrown away in passive crossovers and all the reactances generated. (And as Steffen pointed out: SETs actually sound better when directly coupled to your high frequency drivers.) Direct coupling amplifier outputs to acoustic drivers has a quality all its own in terms of subjective sound quality, in my experience.
The MEH crossover filters I use are either first order or zeroth order (using PEQs to attenuate instead of low-pass/high-pass crossover filters). Using a first order filter in an MEH that crosses at the first notch frequency will ensure time alignment. Any use of higher-order IIR filters in an MEH will ensure that you've got phase response problems.
I find that the only time that someone might need to do something esoteric with crossovers (including passive networks) in an MEH is when driving them at PA levels: 120 db/1m and higher. Power handling and "arrayability" drive the design space for commercial PA MEHs (like Danley's SH and SM series)--not so with home hi-fi, however.
Chris
Last edited:
I should have been more detailed. If one wants to use (as a compromise) a single amp for a two way MEH, then passive crossover is necessary to connect the two bands to one amplifier. The point was that the passive crossover can be pretty simple and FR can be fixed by DSP. For my use, even a few watt amp has enough headroom for DSP compensation. It is a compromise, though.
That's called mono-amping. Some of the effects of that compromise are listed in my last post.
Chris
For my relatively large synergy horn build I will probably start with a two-way system, consisting of 2pcs 15inch woofer and a compression driver..
In the future I am considering making it a three-way system by adding midrange drivers. So I can drive 15" woofers with a Hypex plate amp and applying DSP to the Schroeder freq. The midrange and highs powered by my SET and thus with passive x-over.
Hypex FA501 specs:
Latency Analogue 350 µs (If I am correct 0.35mS)
Digital 1.8 ms (at 96 kHz sampling frequency)
X-over point for the 15'' woofers will be around the Schroeder frequency. Would the latency of 0.35mS cause problems with the summation in the synergy horn?
In the future I am considering making it a three-way system by adding midrange drivers. So I can drive 15" woofers with a Hypex plate amp and applying DSP to the Schroeder freq. The midrange and highs powered by my SET and thus with passive x-over.
Hypex FA501 specs:
Latency Analogue 350 µs (If I am correct 0.35mS)
Digital 1.8 ms (at 96 kHz sampling frequency)
X-over point for the 15'' woofers will be around the Schroeder frequency. Would the latency of 0.35mS cause problems with the summation in the synergy horn?