|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
71.39.40.4
In Reply to: How to get the most efficiency in horns from the throat design? posted by Artemius on January 30, 2007 at 12:34:42:
Take a look at Keele's "Low Frequency Horn Design using Thiele/Small Driver Parameters" available from his website.There is an interesting quote from Neville Thiele towards the back which is particularily enlightening concerning throat size. I'll quote it here:
"It must be pointed out that the transformed desing formulas used in this paper are based on traditional low frequency horn design methods. These traditional methods under some situations may not yield a design which has the optimum combination of response, efficiency and maximum acoustic output. This is primarily due to the fact thattraditional horn design dictates a specific value of throat area which maximizes the nominal efficiency. Beacause a number of the horn's performance charateristics depend heavily on throat area, constraint of this parameter to a specific value removes one valuable degree of design freedom."
Says much.
DM
Follow Ups:
The horn MOUTH size is determines the smoothness of the frequency response, NOT the efficiency.The horn mouth size is related to an antenna suffering reflections due to an impedance mismatch if the mouth is undersized. The idea is to match the mouth to the impedance of the radiation angle it exhausts (or transmits) into.
A mismatch (typically from too small of a mouth) will cause reflections back down the horn to the throat seen by the driver as changes in acoustic impedance (variations in reactance).
A general rule of thumb also described in Keele's paper is the mouth size (Sm) is commonly acepted as mouth CIRCUMFERENCE = WAVELENGTH (Fc). This is for a 1/2 space radiation angle (floor-standing, etc). The mouth can be half of that value for 1/4 space, and one-quarter the 1/2 space mouth size value for corner placement (1/8 space).
Some go even further - the 1980's Gillum/Klipsch MCM patent indicates that the mouthsize can be as small as 1/12th wavelength (Fc) when properly annulled.This would shorten the horn quite a bit. Should the throat size also be made smaller? Won't work - because other considerations of performance are at play in the throat size.
So the bottom line is that 1/4 wavelength HORN PATHWAYS are not required when other considerations are at play in the horn design. Hopefully, this does not confuse anyone.
DM
That's right. If a horn's mouth is too small, it becomes more resonant. As taper is reduced, the cross section becomes increasingly more like a pipe.Reactance annulling can be used to smooth response, but it can only cancel one node. The idea is to use the resonance of the driver in the rear chamber to counteract one standing wave node, usually the primary one. Reactance annulling cannot be used to counter all standing wave nodes.
Standing wave nodes can be countered with another method. Using careful driver placement, some antinodes can be cancelled or reduced. This is usually accomplished by offsetting the driver some distance from the end of the pipe. For several years, Martin King has been championing the idea of using careful placement to modify what standing wave nodes are energized, thereby cancelling or smoothing the antinodes.
A poster (ELENA) from Italy recently posted a design and a pdf document on the Klipsch forum MODS section that features tuned cavities in the horn pathway to reduce the reactance peaks of a reflex-ported front horn, and various tuned ports along the horn pathway (in addition to a front-port) to "fill in" the troughs.His example was a single driver short-pathway reflex-ported folded corner horn.
Same issues we are talking about here. Quite interesting.
DM
More of ELENA's design.
I should mention, ELENA's approach is necessary due to the horn's Fc being well above the Fr (back chamber resonant frequency) required to be much lower in frequency to capacitate the front reflex port which is tuned well below the horns Fc. The issue is exascerbated by the porting, in other words, as it means the driver cannot be annulled.This is not quite the same as a front-loaded of short pathway length which is capable of being annulled. Through the combination of multiple flare-rates AND reactance annulling, the reactance problem cited previously caused by the short pathway is reduced sufficiently.
Very interesting. Incorporating several resonators, each to individually annul a specific standing wave node from the horn. That's an interesting way to smooth the impedance peaks. Thanks for the link!
I don't believe it is all that unique. A number of the lowther cabinets have ported cavities adjunct to the horn path, though off the top of my head the only one i can see in my mind is ported to the outside.
How many of those Lowther-types are front-loaded reflex-ported horns?
My pleasure, sir.I thought it was a unique approach too.
Another methodology to control reactance peaks is disclosed in the 1956 Glenn (General Electric) patent # 2,765,864.A rear-loaded, short-length, unitary-pathway folded horn, this time with a purposely undersized mouth...
To optimize a horn for efficiency, bandwidth, or whatever criteria, one should look at the whole system . The driver, rear chamber, front chamber, compression ratio, flare length, area and profile are all variables that can be modified for a specific goal.See the Marshall Leach paper below:
"A procedure is presented for the design from specifications of moving-coil drivers for low-frequency horn-loaded loudspeakers. The method permits specification of the upper and lower system cutoff frequencies, the volume of the cavity behind the driver, the driver area, the horn throat area, and the desired system electrical impedance. From these specifications, the required Thiele-Small small-signal parameters and the electromechanical parameters of the driver are determined under the condition of a maximum-sensitivity constraint on the system. The procedure can be easily modified for a maximum-efficiency constraint. "
- "On the Specification of Moving-Coil Drivers for Low-Frequency Horn-Loaded Loudspeakers" (Open in New Window)
HiI have been designing horns with THAT paper reduced to mathcad for either case, for about 25 years, for a living.
I have found it to be very good at proportioning driver and horn parameters IF the horn is reasonably large (working close to theory).For situations any distance from an near ideal horn, one must have a way to computer model exactly what ever you have chosen to try.
Do that and DMoore will see what the expansion rate and quarter wavelength mode is and why they matter. You can’t just shrink a bass horn for your convenience and have it still be a bass horn.
Best,
There are many examples of bass horns out there that work quite well and do not exhibit reactance problems AND they are quite short in pathlength AND they have a large throat. Everything they are not supposed to, in other words.The Klipsch Jubilee, for instance. The Klipsch MCM.
I've cited references, what more do you want?
Everything I said is in the available literature for checking. So don't blame me if you disagree. I didn't make it up, nor am I wrong about what it says.
HiI suppose it could happen, but I have never seen a horn / driver combination that didn’t measure very much like a careful computer model of it.
Also, are we talking bass horns such as 60-70Hz LF corner like in the old days or bass as in 30-40Hz or even 20-30Hz?
A quick look at the data sheet for the current Klipsch mcm grand shows its LF section as being a modest 67 inches wide, 32 inches tall and 45 inches deep, with a cutoff at + - 3 dB ,40Hz up (they note that is with with eq applied) and –10db at 35Hz.
The lf section is rated at 800 Watts, combined with the stated sensitivity of 108dB and perfect linearity would produce a maximum SPL of 137dB @ 1M in 55.8 cubic feet of cabinet volume.Alternately, a more modern horn design I am intimately familiar with, based entirely on modeling the actual item..
http://www.danleysoundlabs.com/pdf/Danley TH-115 Spec Sheet_r2.pdf
Two of these would be a fair comparison given the size difference.
These boxes are 40 inches by 22 ˝ by 28 inches and two occupy 29 cubic feet instead of 55.8 for the MCM woofer.
Two have a low frequency sensitivity of 107dB 1W 1m and are – 3 at about 36Hz (no eq). Sensitivity graph is conservative, measured in half space at 10 meters, -10dB from 1 meter and at 100Watts, +10 dB from 1 Watt to remove nearfield cabinet size errors.
Two have a power rating of 2000Watts continuous, 4000 Watts Program, combined with an average 107dB sensitivity and perfect linearity would produce a pressure of 140 dB continuous and 143 peak.So, you have a different approach which is nearly half the physical volume, has a lower cutoff and can go 2 to 4 times louder if needed.
What direction leads to a smaller box for a given cutoff?Tom
The Klipsch examples are not to show that they can't be bettered - but that they are practical examples of very short bass horns that don't suffer from reactance problems as described previously.Multiple flares - described by Olson, used by PWK in 1945. Yes, I tend to like Klipsch designs for the fact that he published, and the products are out there and have been for many years for all to experience whether what he said actually worked or not in a practical manner. It is not intended to be anything other than an easily recognised reference.
Do you find this program to be at all accurate in predicting horns? I have run all sorts of desings thru it and have yet to come up with one that I would put any time into building. I have 515B's, 2205's and 2245's. I figure on putting in a mono sub horn in my attic ( and I have 25 feet of usable length) and venting it into the front of the room, but nothing looked worth while.
HiThis math (Marshal Leach’s) is the best I know of for defining the missing bits in a horn problem. That being said, one MUST keep in mind that it deals with an acoustically ideal (full size) horn and cannot predict what is best when a horn is too small.
For that (real world bass horns), there are better programs that allow one to deal with the real dimensions.
The problem always with bass horns is that they ARE too small to act like the “ideal” so one must know what they actually do to get the most out of them.
It’s sort of like with passive crossovers, you can only get part of the way to the goal if your limited to formula’s and a calculator. If you want to be much closer to perfection, then you must include both the magnitude and phase for both the acoustic and electrical legs of each driver into your computer model.Mc Beans horn simulator is a freeware program which is very popular and seems to do a decent job predicting some things. A program called AKABAK is a more detailed acoustic modeler which is both a real pain in the rear to use and very powerful.
If your fortunate enough to have worked on a NASA project that warranted the writing of software to do this (oddly by a graduate student of Marshal’s Leach’s assigned to help me back then), then of course there is the proprietary type software.
That is what I used to design the Lab subwoofer 4 or 5 years ago, a DIY horn project on Pro-sound web.The key is being able to model what you have to deal with and then play with things.
Try that Mc Beans simulator, read up on what folks are doing with it.
Happy Horning,
Here are a few links to Hornresp models:1. Brad Litz LABhorn model , both for his version and for the original LABhorn.
2. 12Pi and LABhorn models , 20 liter front chamber version of the 12Pi
3. 20, 30 and 40 liter front chamber versions of the 12PiYou can modify the front and rear chambers in the Hornresp model and see a correlation in measured results from physical models. This is shown in the 12Pi measurements, where different versions were made with different front chambers. Hornresp program simulations correspond well with measured results.
Do you credit the prior work of Martin King anywhere in regards to your tapped horn products? For many years, he has proposed offsetting the driver in a tapered pipe away from the closed end so to smooth response. Your technical description of the tapped horn is almost identical to the King method, in many cases using the same terminology to describe the system. Did you license your system from him, or base your designs on his prior work?
Hi V“Do you credit the prior work of Martin King anywhere in regards to your tapped horn products?”
No.
“Did you license your system from him, or base your designs on his prior work?”
Nope, not familiar with his work, the Tapped horn evolved out of an observation in the Unity and later Synergy horns which also have drivers forward of the apex.
You mention Martin King has a Tapped horn, can you direct me to it?
Thanks
I see. Well, it wouldn't be the first time people arrived at the same conclusion independently. You should contact Martin King then, he's a nice guy and you would find you've treaded the same terra.
HiHey, thanks for the link.
Actually I think had run across Martin about 10 years ago on the old Bass List forum.
It is cool to see he has been very busy, I will look through his stuff as I have time.
Thanks again,
Hi Tom - - I'm going to guess that Martin did not try to model such an arrangement until requests arose from curiosity in your tapped horns & white paper plus re-interest in Jensen's transflex. Cowan's attempts at TD-style T-H look interesting - two plyed ok - one had a pretty big hole.Willian's Pages
Hi FreddyiHuh, another fellow who’s been busy making sawdust.
I remember him from when Nick was selling the horns, his boxes turned out nice.
One may notice on his curves that moving the driver position AND / OR changing its parameters can have a similar effect on the response curve.
What he cannot see is the effect this also has on excursion for a give output.
It’s hard to make one by trial and error but he has gotten some good results.
Best,
For several years, Martin has been proposing driver offset in the line specifically to cancel standing wave nodes. That's part of the design process for him, and is outlined in his papers.
Back in the late 80's I was building gently tapered lines with not one but two offset woofers, with the midpoint between the woofers approximately 1/3 of the effective distance from the closed end of the line. I was trying to smooth out the sound coming from the end of a relatively unstuffed line so that the peaks & dips weren't so huge. I wrote an article about it to submit to SpeakerBuilder, but then decided to keep the idea to myself and use it in a commercial loudspeaker venture (which subsequently got derailed by my divorce).I've since gone over to vented enclosures because they enable me to use woofers whose characteristics are in my opinion better suited for good midrange reproduction.
The system described in my unsubmitted article had another trick or two up its sleeve, which I'm still hoping to one day revisit and possibly use in a commercial speaker system.
thanks V - its obvious I've not used nor studied MJK's sheets - for some reason I want a karlson-coupler to "be something" (other than bizarre toy of past) - hahaha - at regular aspect k have coupled-cavity feature with possible peak before going into a dip (3rd Z peak = 235 on K12/155 on K15), tone chamber effect (kind of slap-echo effect in upper half of coupler when fingers are snapped in the coupler) - k might only have limted use at best.those horn traps are cool
Best,
Freddy
See this one, for example:"If the driver is offset to a location at approximately one third the length of a straight transmission line, the 3/4 wavelength standing wave can be totally supressed. In fact, the impact from every other quarter wavelength mode (3/4, 7/4, 11/4, ...) will be essentially removed from the system SPL response."
(3rd try to reply) - thank you very much. Hopefully AJ-HOrn will be able to do (for simulation fun) similar work. If not then still will be fun for 'half-horn'
Good chatting with you.
Hi VBack at Servodrive / Sound Physics Labs, we began selling speakers for commercial sound with the drivers tapped forward of the apex in 1999 and I later received a patent (Unity horn) for it.
I built the first Tapped bass horn about 5 years ago and have applied for a patent on it also, at Danley Sound Labs, we have been selling them as a commercial product since January 2005.
There is / was no association with Martin’s work other than the field of interest.
Do read through Martin's papers. I think you'll find yourself on terra cognita.
Yeah, that’s called a transmission line stub in Radio Antenna land.
There are a number of antenna analogues which are useful in speaker design..
Tom
You'll notice his papers were written some 5 years ago. They describe a tapered pipe - which is what a basshorn really is - and he proposes driver offset specifically to counter standing wave nodes.
While that is similar to the old Bassmaxx style horn which had a stub, it is not like the Tapped horn in that there is no stub and side B of the driver, to provide the variable source impedance the column needs, it is near (but not a fixed proportional distance from) the open end.
Tom
The tapped horn looks to be a folded tapered line made in the way Martin King proposes. Using his method, both sides of the drivers are used as radiating surfaces, just as yours is. You have folded your pathway, but the two methods look exactly the same to me. I'm confident his spreadsheets will perfectly model your tapped horn, because it uses the same principle. Study his papers closely, I think you'll feel like you're looking in the mirror.
Hi VI have not had time to look at everything there; can you point EXACTLY to what you think is the same thing and the spreadsheet which calculates it?
See this one, for example:"If the driver is offset to a location at approximately one third the length of a straight transmission line, the 3/4 wavelength standing wave can be totally supressed. In fact, the impact from every other quarter wavelength mode (3/4, 7/4, 11/4, ...) will be essentially removed from the system SPL response."
Hi VI am not clear where you think this applies, in the Tapped horn, one driver face is at the closed end (not some distance away as he proposes) and the other face is near the open end. His concern seems to be “not driving” resonances as opposed to dealing with acoustic impedances.
You said “I'm confident his spreadsheets will perfectly model your tapped horn, because it uses the same principle. Study his papers closely, I think you'll feel like you're looking in the mirror.”
This is what I wanted you to point out to me, where is "that part" of his site?
Tom
Well, you might start on the main page. The principles are fundamentally the same as what you are doing, so I think probably you'll be interested in seeing what MJK has written. I'm surprised you hadn't seen this before, since he's been pretty visible in the DIY community.
Greets!MJK said they won't in this post, has he since changed his WSs?
GM
As you can see, MJK says, "I thought about the problem yesterday and concluded that a worksheet to model this type of enclosure is possible. It would require some rearranging and extension of the math but I believe it could be done." The concept is there already, and you can see in his prior work that he was thinking along these lines. I think his work is worth mentioning in this regard, perhaps giving some credit where due.
Greets!I'm all for giving the man his due, and have on numerous occasions and nearly as many forums, but he only thought about it after being presented with the concept. Clearly, he didn't have it in mind when he was developing his WSs and after talking with him yesterday still has no plans to modify his WS to model a tapped horn or TL.
Hi VSo you can’t actually point to a Tapped horn on his site then?
MJK tends to concern himself with ideas more than specific implementations. However, the MJK analysis describes the principles used in what you are calling a tapped horn. Not to sound like a broken record, but again, I encourage you to read through his papers if you haven't already.
Hello V,It looks like you do not completely understand what a 'tapped horn' is. Offsetting the driver in a TL is a very old trick, Martin did not invent this. I think the credits for the 'roots' of the Tapped horn go to Jensen with their Transflex design.
Best regards,
Hi WaltI agree about the Transflex, the first time I saw it I thought OMG, it’s nearly a Tapped horn. In modeling the Jensen, the reason it isn’t a modern product came out though.
As you have seen, it isn’t easy to get all the relationships right as there is no written procedure to rely on.
Best,
It really is old news"Sound reproduction"
G.A. Briggs
1953
Yes, perhaps you are right. However, MJK did a very nice job of quantifying things through mathematical analysis. I think some credit is deserved by several people, Jensen, King, Danley, to name a few. By the way, I like what you've done too.
Hi VI don’t belong in “that” list.
Understand, I quit the AES 11 years ago and have not published any papers or references on the Tapped horn other than the White paper and talking about it at trade shows etc..
Other than the explanation in the white paper (which you’ll notice is not at all like Martin’s) and occasionally talking about it on forum’s, that is about it so far as public disclosure.My focus was first seeing if I could exploit “something” I saw with the Unity and Synergy horns, then on making it work well enough to be a product and then getting the most out of it.
My understanding and explanation of how it works, the path to designing them and all the design tools involved were things I derived without outside help / input.
Now, the determination of “if” this is novel is up to the Patent office to decide but the performance of the current Tapped horns compared to regular small bass horns is such that “if” someone else had made it work properly, they certainly would have used it.
Keep in mind the th-115 and th-215 are both smaller / lighter than all but one speaker model tested at a recent Pro-Sound subwoofer shoot out.
http://srforums.prosoundweb.com/index.php/mv/msg/20327/0/16/0/
http://srforums.prosoundweb.com/index.php/t/20399/0/
Arnold Klayman patent number 5, 177, 329 Arnold is 82 now and retired.
I also found another patent that came out about ten months after Klayman's, Hayakawa US # 5, 197, 103 which is similar. Then there is also patent US 4, 064, 966 worth looking at as well. Going back to 1962 you can see another similar patent US# 3. 047, 090 look at fig. # 4. Regards Moray James.
moray james
This post is made possible by the generous support of people like you and our sponsors: