MaVo said:
Don't know, I haven't made any attempt to reverse engineer any of his TH designs, but my SWAG is that due to the practical constraints of their target markets, I can't see him gambling on low volume custom drivers beyond 'bullet proofing' proven designs like he did with the LABhorn's.
GM
True, that doesnt seem to be rational. But if the driver isnt custom, then hornresp has to do something wrong or DSL lists peak levels from driving their drivers to xmech, which i really doubt.
I dont know if i really reverse engineered it, i simply modeled a tapped horn with about the same frequency response, driver and volume.
I dont know if i really reverse engineered it, i simply modeled a tapped horn with about the same frequency response, driver and volume.
CLS said:
CLS: Yep, I forgot that the horn length is not a function of the position of the driver, just the distance from the throat to the mouth. Even so, I wonder if it might be possible to tame some of those peaks at the upper end by judicious placement of the two drivers (e.g. phase cancellation). I would think it should be possible to model this.
Cordraconis said:@Craigwalsh: A few months ago I made a series of AkAbak scripts with a passive radiator in the "Tap" position of a TH. I've made these to try to create a horn system that could be calculated using the assumptions of Leach' Horn model, but regarding to a TH system the problem of finding out the exact parameters for the passive radiator stayed the same.As far as I know, the only person who can have a paper/model for THs similar to the Leach paper, is Tom Danley and I doubt he will release such a thing.
Anyway, I seem to have lost the series of scripts (I'll look again later), but I sent them to GM as study/discussion material and maybe he still has them somewhere lying around and can post them here.
Removing the PR references and dropping in a second driver to model the response, would be a quick "hack" for what you need.
Cordraconis: I'm not clear on what a PR would provide at the tap, and I'm not familiar with the "Leach" paper. I've designed speakers for many years, but horns and especially TH design are new to me (this is probably clear). It would be interesting to see what you came up with though, as I believe the whole area of TH design is really very cool. Having worked with so many closed box, bandpass, PR and ported designs, it's just amazing to me the level of efficiency that can be achieved by THs of similar volume.
MaVo said:
MaVo: Actually, I was thinking of two identical drivers, although your thinking of using drivers with different TS parameters could have some merit. While I agree that the horn clearly limits the bandwidth at the low end, the response anomalies at the high end appear to be a consequence of several factors, including driver position (according to Hornresp).
It would be very cool if it were possible to model the response of two drivers oriented in this manner. In any event, I'm keen to build a dual driver TH in a relatively small volume for pro-sound use. I will likely configure the drivers to enter the throat at the same place along the length of the horn. But, I do think it would be interesting to see the effect of varying the position of the two drivers along the horn.
Thanks to all for contributions here, BTW. This thread has been incredibly inspirational and very much appreciated.
MaVo said:
GM said:
We had this debate some time back and never resolved it -- some posters (including Tom) said that the excursion predicted by Hornresp and AkaBak was higher than in a real life tapped horn (without any real explanation as to why this should be the case), others said that the predictions agree with their measurements.
I tried pushing the issue to get an answer out of Tom -- because a 2x difference in excursion has a huge influence on whether it's worth me building a tapped horn for PA or not -- but didn't succeed.
Nobody (including Tom) has said anything to dispel the hypothesis that the TH-115 uses a standard 15TBX100.
Ian
Hello Ian,
http://www.diyaudio.com/forums/showthread.php?postid=1399717&highlight=#post1399717
Above a link where you could say the TH-115 is driven by a TBX100.
I built myself a couple off tapped horns, which was a hell of job. To fit the horn in the smallest dimension possible i had to make the tough flare panel "the labhorn cut". It`s a horn driven by the 15P1200 and with a cutt-off around 45hz.
A couple off years ago I build myself 2 labhorns and from that time i believed TD`s theory. It is the labhorn that is my reference and i can see this horn goes not as deep, but has a lot more output in
above the 45hz point and less excursion. A link to my previous labhorns.
http://www.speakerstore.nl/index.php?l=nl&pg=12&prjID=25
I did make a prototype off my tapped horn and measure it. I was finding out my throat and front volume was a bit to big. It was my conclusion from a impedance measurement.
Soon i will post some more photo`s and some measurements.
Cheers,
Marcel
http://www.diyaudio.com/forums/showthread.php?postid=1399717&highlight=#post1399717
Above a link where you could say the TH-115 is driven by a TBX100.
I built myself a couple off tapped horns, which was a hell of job. To fit the horn in the smallest dimension possible i had to make the tough flare panel "the labhorn cut". It`s a horn driven by the 15P1200 and with a cutt-off around 45hz.
A couple off years ago I build myself 2 labhorns and from that time i believed TD`s theory. It is the labhorn that is my reference and i can see this horn goes not as deep, but has a lot more output in
above the 45hz point and less excursion. A link to my previous labhorns.
http://www.speakerstore.nl/index.php?l=nl&pg=12&prjID=25
I did make a prototype off my tapped horn and measure it. I was finding out my throat and front volume was a bit to big. It was my conclusion from a impedance measurement.
Soon i will post some more photo`s and some measurements.
Cheers,
Marcel
Hi Guys,
Some questions have answers and others don’t.
So far as a difference between observed and predicted excursion, I am not very concerned as I have nothing to do with software and can measure / am stuck with the real thing anyway.
I am very grateful to have a ways to model the response of a horn that when built measures similar to the model. I don’t know how many of you can appreciate the dark days of horns when all one could do is copy or make prototypes with your best guess.
I guess in some ways that cut / try part hasn’t gone totally now that I think about it.
Where there are differences between prediction and result it is because of one or more things; The computer model you made wasn’t close enough to the real thing, it didn’t have enough important secondary details in it, the math didn’t account for enough important details or has simplifying assumptions which didn’t apply in this case, has a bug or all of the above.
For example, most programs overstate the Q of internal resonances, some factor changes with scale, smaller systems have more internal losses and so measured Q’s tend to be lower than predicted and often enough very sharp predicted features are non-existent in measurements.
Excursion would be tied in too, potentially where a difference in measured response verses predicted existed, there would be a difference in excursion as well.
So far as drivers, I try to use stock products if I can, it is important that if somebody breaks a speaker on the road, that they can get it going asap and that when we order drivers, that they are still in business.
Especially in Pro Sound and Permanent installations, I can’t afford to risk using a driver that hasn’t already been beaten to failure many times and had the all the production bugs worked out. So, with the exception that most of them are 4 Ohms now, we do use a standard 15TBX100.
Power and SPL rating in Pro sound.
There are several ways this is specified.
A 1000Watt rms rating using pink noise requires an amplifier than can put out 4000Watts as the test signal by definition has a 6 dB peak to average ratio and the average is 1000W.
This puts the theoretical peak level at; sensitivity (106dB) + dBW(30dB) + peak 6dB = 142dB peak mid band.
A common practice in Pro sound is not to use mid band sensitivity but to use a sensitivity taken at a higher frequency, say for a three letter name subwoofer, measure at 500Hz (a point truly irrelevant to its intended use but more impressive).
Here with a TH-115 if one uses the figure at 100Hz (still within its operational bw in some cases) the sensitivity is more like 112dB so that 142dB could be pushed up to 148dB.
Anyway, all of that is arm waving, marketing numbers, like “Watts” in the first place and with more arm twisting this is how some can rate a floor wedge at 148dB too.
Perhaps by directing a lightning strike to the terminals, the floor wedge could do it.
Not wanting to appear hobbled where this “perfect imaginary” rating is normal, we used some of those numbers but on the conservative side.
Also however there is the 133 dB (136 pk hold ) rating which is where it’s sonic character subjectively started to change, I would say more like its “usable loudness” but there is no standard rating for this and no one rates what loudness a speaker starts to poop out sonically.
Unfortunate too as this figure is the one more related to listening to music.
Iand, so far as telling you guys everything about everything, as much as I like building stuff and DIY, I just can’t do that.
I am trying to grow a company in the land of well funded giants by coming up with new things, I am trying to make every box a home run or as close as I can because customers jobs and casual side by side comparisons are key to cutting a path through the jungle.
Iand, I understand your skepticism, If your in Pro sound, why not have your company buy a couple or four 115’s and see what the real thing can do and do compared to what you have / what your used to.
Best,
Tom Danley
Some questions have answers and others don’t.
So far as a difference between observed and predicted excursion, I am not very concerned as I have nothing to do with software and can measure / am stuck with the real thing anyway.
I am very grateful to have a ways to model the response of a horn that when built measures similar to the model. I don’t know how many of you can appreciate the dark days of horns when all one could do is copy or make prototypes with your best guess.
I guess in some ways that cut / try part hasn’t gone totally now that I think about it.
Where there are differences between prediction and result it is because of one or more things; The computer model you made wasn’t close enough to the real thing, it didn’t have enough important secondary details in it, the math didn’t account for enough important details or has simplifying assumptions which didn’t apply in this case, has a bug or all of the above.
For example, most programs overstate the Q of internal resonances, some factor changes with scale, smaller systems have more internal losses and so measured Q’s tend to be lower than predicted and often enough very sharp predicted features are non-existent in measurements.
Excursion would be tied in too, potentially where a difference in measured response verses predicted existed, there would be a difference in excursion as well.
So far as drivers, I try to use stock products if I can, it is important that if somebody breaks a speaker on the road, that they can get it going asap and that when we order drivers, that they are still in business.
Especially in Pro Sound and Permanent installations, I can’t afford to risk using a driver that hasn’t already been beaten to failure many times and had the all the production bugs worked out. So, with the exception that most of them are 4 Ohms now, we do use a standard 15TBX100.
Power and SPL rating in Pro sound.
There are several ways this is specified.
A 1000Watt rms rating using pink noise requires an amplifier than can put out 4000Watts as the test signal by definition has a 6 dB peak to average ratio and the average is 1000W.
This puts the theoretical peak level at; sensitivity (106dB) + dBW(30dB) + peak 6dB = 142dB peak mid band.
A common practice in Pro sound is not to use mid band sensitivity but to use a sensitivity taken at a higher frequency, say for a three letter name subwoofer, measure at 500Hz (a point truly irrelevant to its intended use but more impressive).
Here with a TH-115 if one uses the figure at 100Hz (still within its operational bw in some cases) the sensitivity is more like 112dB so that 142dB could be pushed up to 148dB.
Anyway, all of that is arm waving, marketing numbers, like “Watts” in the first place and with more arm twisting this is how some can rate a floor wedge at 148dB too.
Perhaps by directing a lightning strike to the terminals, the floor wedge could do it.
Not wanting to appear hobbled where this “perfect imaginary” rating is normal, we used some of those numbers but on the conservative side.
Also however there is the 133 dB (136 pk hold ) rating which is where it’s sonic character subjectively started to change, I would say more like its “usable loudness” but there is no standard rating for this and no one rates what loudness a speaker starts to poop out sonically.
Unfortunate too as this figure is the one more related to listening to music.
Iand, so far as telling you guys everything about everything, as much as I like building stuff and DIY, I just can’t do that.
I am trying to grow a company in the land of well funded giants by coming up with new things, I am trying to make every box a home run or as close as I can because customers jobs and casual side by side comparisons are key to cutting a path through the jungle.
Iand, I understand your skepticism, If your in Pro sound, why not have your company buy a couple or four 115’s and see what the real thing can do and do compared to what you have / what your used to.
Best,
Tom Danley
Tom Danley said:
Hi Tom
I understand completely about the deficiencies of simulation, I run into these kind of problems all the time at work (chip design) -- the tools are only as good as the information you put in.
I also understand that you have a business to run, I'm not trying to find out what your "secret sauce" is in tapped horn design -- given the extended bandwidths you're achieving you're obviously doing something (or several somethings) clever that nobody else has worked out.
What I'm trying to find out is *what* (if anything) is wrong with the simulation tools -- as far as I can see they're solving fairly well-known basic mechanical/acoustic equations without making any dodgy assumptions that could obviously be in error. A tapped horn is just another way (but a rather clever one!) of building a box with a speaker in it, and ought to be as easy (or difficult) to simulate accurately as any other type of speaker.
If the tools can be relied on then other people (like me) can try and come up with their own tapped horn "secret sauce" -- not commercially without a licence of course, I'm talking about hobbyists doing this for their own use who would never buy a ready-made speaker on principle (never mind the cost
I'm certainly not sceptical about the performance of your products, it's clear that they're way ahead of anything else in the market and are far more honestly specified than most (all?) competing products.
But speaking as someone who plays in a band that certainly can't afford to buy multiple TH-115s (even if they were available in the UK) I'd like to try and build something similar, but I don't have the time to build and measure multiple iterations of a complex design on a "cut-and-try" or "maybe it'll work" basis -- you can easily justify doing this because it's your business.
Also given the very different driver parameters needed for a tapped horn compared to (for example) a reflex or bandpass box, if I spend lots of time and money building a tapped horn (with expensive pro drivers like the BMS 15N850) and it doesn't deliver what I expect, the drivers aren't really suitable for re-using in a different enclosure.
Hence my interest in trying to predict the performance first...
Cheers
Ian
Hi guys....
I'm jumping in here a little late and I'm trying to read through the thread , but I've got a couple of newbie questions here. Also just for the record, I'm a mac user so hornresp is not an option ATM, but if I decide to pursue the TH as my next project I will do the virtual PC thing and D/L hornresp and learn it.
So my question pertains to horn size (e.g. overall box size) vs. bandwidth vs. power handling.
Question 1
What factor determines band width the most, (a) length of the horn, (b) internal volume of the horn, or (c) horn mouth.
I am mostly interested in getting up to roughly 100 hz. where I can cross to my mains. The bottom end needs to go to about 40 hz roughly, but if it goes lower, that's even better.
Question 2
Does a bigger horn/box size adversely affect power handling, such as a traditional vented box does?
This would be for live sound use. I am trying to balance output, efficiency, and box size. I don't need extreme low end output, but anything under 40 hz is welcome.
Basically I am looking at a design with the eminence definimax 12 or 15. 2 or 4 cabinets depending on how much output can be had. These would be for a band PA that gets rented a couple times per year, and has to fit in a van or small trailer.
So what are the general guidelines for tuning vs. efficiency. Is it bigger box = higher tuning and more efficient, or does changing the apparent size of the horn not change efficiency that much....
Thanks in advance,
-=Tim=-
Back to reading the over 2000 posts!
I'm jumping in here a little late and I'm trying to read through the thread
, but I've got a couple of newbie questions here. Also just for the record, I'm a mac user so hornresp is not an option ATM, but if I decide to pursue the TH as my next project I will do the virtual PC thing and D/L hornresp and learn it.So my question pertains to horn size (e.g. overall box size) vs. bandwidth vs. power handling.
Question 1
What factor determines band width the most, (a) length of the horn, (b) internal volume of the horn, or (c) horn mouth.
I am mostly interested in getting up to roughly 100 hz. where I can cross to my mains. The bottom end needs to go to about 40 hz roughly, but if it goes lower, that's even better.
Question 2
Does a bigger horn/box size adversely affect power handling, such as a traditional vented box does?
This would be for live sound use. I am trying to balance output, efficiency, and box size. I don't need extreme low end output, but anything under 40 hz is welcome.
Basically I am looking at a design with the eminence definimax 12 or 15. 2 or 4 cabinets depending on how much output can be had. These would be for a band PA that gets rented a couple times per year, and has to fit in a van or small trailer.
So what are the general guidelines for tuning vs. efficiency. Is it bigger box = higher tuning and more efficient, or does changing the apparent size of the horn not change efficiency that much....
Thanks in advance,
-=Tim=-
Back to reading the over 2000 posts!
"What factor determines band width the most, "
The limiting factor seems to be the upper dip in the frequency response of the TH. This is related to the 1/4W length of the TH, and where the tap is in the mouth. The closer the tap is to the mouth, the wider the bandwidth seems to be. Lower Qts drivers seem to like being closer to the ends (throatand mouth).
"Does a bigger horn/box size adversely affect power handling"
Excursion is the limiting factor here, and that is driven by the length.
Here is one possible TH with the Definimax 4015LF (two horns shown with 2.83V in 2Pi), ±2dB 31hz~150hz, 108dB at 35hz.
The limiting factor seems to be the upper dip in the frequency response of the TH. This is related to the 1/4W length of the TH, and where the tap is in the mouth. The closer the tap is to the mouth, the wider the bandwidth seems to be. Lower Qts drivers seem to like being closer to the ends (throatand mouth).
"Does a bigger horn/box size adversely affect power handling"
Excursion is the limiting factor here, and that is driven by the length.
Here is one possible TH with the Definimax 4015LF (two horns shown with 2.83V in 2Pi), ±2dB 31hz~150hz, 108dB at 35hz.
may be a info,
a horn has only 3 oktaven band width,
the flare rate is most important,
ln (AM/AH)/l = k
AH = Horn throat
AM = Horn mouth
k = flare rate
l = Length
http://www.lenardaudio.com/education/07_horns.html
TH has an extrem low k
K WERTE:
0,37 ~ 10 Hz
0,55 ~ 15 Hz
0,7 ~ 19 Hz
0,9 ~ 25 Hz
1,1 ~ 30 Hz
1,3 ~ 35 Hz
1,48 ~ 40 Hz
a horn has only 3 oktaven band width,
the flare rate is most important,
ln (AM/AH)/l = k
AH = Horn throat
AM = Horn mouth
k = flare rate
l = Length
http://www.lenardaudio.com/education/07_horns.html
TH has an extrem low k
K WERTE:
0,37 ~ 10 Hz
0,55 ~ 15 Hz
0,7 ~ 19 Hz
0,9 ~ 25 Hz
1,1 ~ 30 Hz
1,3 ~ 35 Hz
1,48 ~ 40 Hz
A suggested design for two Difinimax 4012HO.
As we can see, it is minimal sized. Only a 49cm (19-1/4") x 49cm (19-1/4") mouth for both drivers in one cabinet, or about 35cm (13-5/8") x 35cm (13-5/8") if two singles are constructed.
Response is about 105dB at 42hz, ±1.5dB from 38.5hz to 130hz, 2.83V in 2Pi.
As we can see, it is minimal sized. Only a 49cm (19-1/4") x 49cm (19-1/4") mouth for both drivers in one cabinet, or about 35cm (13-5/8") x 35cm (13-5/8") if two singles are constructed.
Response is about 105dB at 42hz, ±1.5dB from 38.5hz to 130hz, 2.83V in 2Pi.
Dang!
That's incredible for 2 twelves!
Somewhere in this thread someone posted a pdf containing a guess at a folding scheme for the th115. Where the throat split into 2 paths. Is this necessary, or couldn't it be made simpler like a traditional bass horn? Just move the location of the driver like this....
That's incredible for 2 twelves!
Somewhere in this thread someone posted a pdf containing a guess at a folding scheme for the th115. Where the throat split into 2 paths. Is this necessary, or couldn't it be made simpler like a traditional bass horn? Just move the location of the driver like this....
Attachments
The section in the middle where a woofer would be located in a traditional bass horn is now empty space, where you *could* mount a resonator or 2. Or tune the size of that enclosure to BE the resonator.
You could also put a hatch door on it like is normally done to mount the woofer inside, except you would now use it to access your resonator(s) for tuning purposes.
Thirdly that space could house an amp.
Fourthly you could cut it out and have a hole through the middle of your speaker cabinet!
Feel free to let me know if this is not a possible folding scheme for whatever reason. I am still deep in the learning process after all....
You could also put a hatch door on it like is normally done to mount the woofer inside, except you would now use it to access your resonator(s) for tuning purposes.
Thirdly that space could house an amp.
Fourthly you could cut it out and have a hole through the middle of your speaker cabinet!
Feel free to let me know if this is not a possible folding scheme for whatever reason. I am still deep in the learning process after all....
Oh, yeah....
I thought I would offer up a bit of advice for anyone building a speaker cab that needs to be really well sealed (such as these).
I've seen some people mention liquid nails, or silicone, or even blu-tak to seal the joints. That's the hard way, here is what I do. Grab a quart of discounted latex paint from the hardware store. They always have discounts on paint that was mixed up into the wrong color. Take that and pour it straight (don't thin it) into your cabinet. About a cup (.5 liter) at a time. Then tilt your cabinet in various positions so that the thick latex runs over and into all corners and seams. It will stick pretty well, and it builds up fast. Usually one coat is all thats needed. But if you want to you can put several coats on to ease your mind.
It's amazing how well sealed you can make your cabs using this method.
Anyway, there's my little contribution to this great thread! {/thread jack}
I thought I would offer up a bit of advice for anyone building a speaker cab that needs to be really well sealed (such as these).
I've seen some people mention liquid nails, or silicone, or even blu-tak to seal the joints. That's the hard way, here is what I do. Grab a quart of discounted latex paint from the hardware store. They always have discounts on paint that was mixed up into the wrong color. Take that and pour it straight (don't thin it) into your cabinet. About a cup (.5 liter) at a time. Then tilt your cabinet in various positions so that the thick latex runs over and into all corners and seams. It will stick pretty well, and it builds up fast. Usually one coat is all thats needed. But if you want to you can put several coats on to ease your mind.
It's amazing how well sealed you can make your cabs using this method.
Anyway, there's my little contribution to this great thread! {/thread jack}