I've been awarded a small grant through my college to design some high frequency horns, and it has been one hell of a ride. Before this project, I didn't really have any knowledge at all of formal acoustics, aside from some basic stuff I picked up here and there from DIY sources.
I'm currently following parts of Leo Beranek's Acoustics: Sound Fields and Transducers as well as Quarter Wave's Horn Theory texts for reference.
The driver I'm using for the tweeters are Selenium D220Ti. I'm planning on crossing them at 2000 Hz, and as such I'm setting the cutoff frequency of the horn to be 1500 Hz.
Using the calculations in part 5 of the Quarter Wave papers, I've come up with an exponential horn who's geometry is defined by:
But that seems incredibly small, especially considering that the horn used for the spec sheet measurements, Selenium HL14-25 is 6 inches in diameter at the mouth and 10 inches long (although it's designed for a lower cutoff).
For mid-frequency, the results seem much more reasonable. Using the same calculations, and choosing a Selenium d250-x as the driver, I get a horn with:
If I'm doing something wrong, I can't seem to figure out what it is; the only difference with the Quarter Wave/Beranek texts and my scenario is that I'm using compression drivers while the they seem to be focused more on building horns for larger speakers in a baffle. However, I haven't been able to find on literature on how to design horns specifically for compression drivers so I've been assuming there's little, if any, difference.
Am I crazy, or should I just go with what the calculations are telling me?
I'm currently following parts of Leo Beranek's Acoustics: Sound Fields and Transducers as well as Quarter Wave's Horn Theory texts for reference.
The driver I'm using for the tweeters are Selenium D220Ti. I'm planning on crossing them at 2000 Hz, and as such I'm setting the cutoff frequency of the horn to be 1500 Hz.
Using the calculations in part 5 of the Quarter Wave papers, I've come up with an exponential horn who's geometry is defined by:
- A 1" throat diameter
- A 2.875" mouth diameter
- And 1.51" in length
- For a cutoff of 1500 Hz.
But that seems incredibly small, especially considering that the horn used for the spec sheet measurements, Selenium HL14-25 is 6 inches in diameter at the mouth and 10 inches long (although it's designed for a lower cutoff).
For mid-frequency, the results seem much more reasonable. Using the same calculations, and choosing a Selenium d250-x as the driver, I get a horn with:
- a 1" throat diameter
- a 8.614" mouth diameter
- and a 9.28307" length
- and a cutoff of 500 Hz.
If I'm doing something wrong, I can't seem to figure out what it is; the only difference with the Quarter Wave/Beranek texts and my scenario is that I'm using compression drivers while the they seem to be focused more on building horns for larger speakers in a baffle. However, I haven't been able to find on literature on how to design horns specifically for compression drivers so I've been assuming there's little, if any, difference.
Am I crazy, or should I just go with what the calculations are telling me?
Just pick up a bair of these:
Denovo Audio DW-62S round waveguide DIY Sound Group
They sound great for $5.50 each.
Denovo Audio DW-62S round waveguide DIY Sound Group
They sound great for $5.50 each.
An externally hosted image should be here but it was not working when we last tested it.
Thanks for the replies!
The project was given funding so I could *design* horns; they're for personal use, and the whole process is meant to be more educational than constructing something that sounds good on the first try. To that end, buying a premade horn is out.
Unfortunately, hornresp is also out as I don't have steady access to a windows computer. Does anyone know if there is a Linux equivalent?
The project was given funding so I could *design* horns; they're for personal use, and the whole process is meant to be more educational than constructing something that sounds good on the first try. To that end, buying a premade horn is out.
Unfortunately, hornresp is also out as I don't have steady access to a windows computer. Does anyone know if there is a Linux equivalent?
That's a great deal - do you have measurements?
The price is great but $15 minimum for shipping makes it $26/pair. Something wrong about that.
These PRV's go down to 400Hz and are $15ea and shipping is $6, for $36/pair. They are 45 deg though, but being able to XO at 1kHz is important... Hmm...
http://www.parts-express.com/prv-audio-wgp14-25-chrome-1-45-x-45-waveguide-1-3-8-18-tpi--294-2904
These go down to 1200Hz and are $11.
http://www.parts-express.com/prv-audio-wg16-25-b-1-45-x-45-abs-elliptical-tractrix-waveguide-2-3-bolt--294-2890
Question: is there a good STL file for horns similar to this for those with access to a 3d printer? I would like to have it go down to 1200Hz and have 45 deg to 60 deg coverage.
Last edited:
An exponential horn would work. It would be an unusual choice for a tweeter from a hi-fi perspective. The LeCleach profile is a variation of exponential but looks fairly different in practice and would be the choice of some. The other choice would be a waveguide with a more conical profile. I don't know if this is important to you.
Simply, the needs change vs. frequency and an exponential horn would be a more common choice at lower frequencies.
One other thing to know is how you'll hand off at the lower end. Eg: at 2000Hz you might match directivity off to a direct radiator, say an 8" woofer which would need to have good behaviour aty those frequencies and appropriate sensitivity.
Simply, the needs change vs. frequency and an exponential horn would be a more common choice at lower frequencies.
One other thing to know is how you'll hand off at the lower end. Eg: at 2000Hz you might match directivity off to a direct radiator, say an 8" woofer which would need to have good behaviour aty those frequencies and appropriate sensitivity.
I've finally gotten the impedance graphs to work out. This is what it would look like with a 500 Hz cutoff, and the geometry stated in the OP:
And this is what it would look like with a 300 Hz cutoff:
With the only difference being the size of the mouth and the length of the horn.
This is talking about the size of the mouth, or the length of the horn? Pardon my ignorance, but I'm not sure I've actually read an explanation of what is meant by "Quarter Wave"; it's the name of the guy's website, so I just assumed it was a moniker.
But, if it's dealing with the area of the mouth, my 300 Hz cutoff horn should suffice, no? Definitely picking up that Olson book when I get the chance.
An externally hosted image should be here but it was not working when we last tested it.
And this is what it would look like with a 300 Hz cutoff:
An externally hosted image should be here but it was not working when we last tested it.
With the only difference being the size of the mouth and the length of the horn.
This is talking about the size of the mouth, or the length of the horn? Pardon my ignorance, but I'm not sure I've actually read an explanation of what is meant by "Quarter Wave"; it's the name of the guy's website, so I just assumed it was a moniker.
But, if it's dealing with the area of the mouth, my 300 Hz cutoff horn should suffice, no? Definitely picking up that Olson book when I get the chance.
Lemmee see if I have this about right? I'm fairly incredulous, sorry.
Your college gave you a grant to rediscover horn acoustics and the related math??
Assuming your are at an engineering college (??) - perhaps not, since they'd likely find the idea of paying for this funny - then your library is literally full of texts and journal articles from the last century (funny to be able to say such a thing!) full of horn theory and horn designs. As is the internet.
Two texts have just been mentioned, Berenek and Olsen being two authors to start with. Then you want to find the JAES anthology series. After that read Solomon's paper on his tractrix expansion...
After you get your feet wet, then you can follow up with the various newer designs intended to control the higher frequency dispersions - some of that info is to be found on this site.
Bottom line, you can use a conical expansion and make a "horn" that will work as a horn down to your design frequency. Your job is to find out IF this is a good idea, or not, and WHY. Then you will have earned your grant.
Btw, getting US (as a group) here to do your work for you, to get the result fast gets you no benefit at all, as you learn next to nothing.
_-_-
Your college gave you a grant to rediscover horn acoustics and the related math??
Assuming your are at an engineering college (??) - perhaps not, since they'd likely find the idea of paying for this funny - then your library is literally full of texts and journal articles from the last century (funny to be able to say such a thing!) full of horn theory and horn designs. As is the internet.
Two texts have just been mentioned, Berenek and Olsen being two authors to start with. Then you want to find the JAES anthology series. After that read Solomon's paper on his tractrix expansion...
After you get your feet wet, then you can follow up with the various newer designs intended to control the higher frequency dispersions - some of that info is to be found on this site.
Bottom line, you can use a conical expansion and make a "horn" that will work as a horn down to your design frequency. Your job is to find out IF this is a good idea, or not, and WHY. Then you will have earned your grant.
Btw, getting US (as a group) here to do your work for you, to get the result fast gets you no benefit at all, as you learn next to nothing.
_-_-
Hi Bear,
I should first state that this grant is a grand total of $400; its nothing serious. As the description of the grant stated, it's simply "enough to cover materials and costs to undertake a project that may not otherwise be possible"; they don't want me to revolutionize horn theory... The point of mentioning that I got a grant to do it was to let y'all know that I'm meant to be coming up with my OWN horn design for personal use, not just buying one premade to sound good. The goal of the project is personal research, not necessarily fantastic results; however, having a prototype is important to show that I actually produced SOMETHING, even if it sounds like crap. As long as I can explain the mistakes I made and how to fix them, I'll be satisfied that the college won't think I'm just trying to get money out of them. This is especially important for when I'm applying for future funding.
I'm also feeling that people on this thread are getting the idea that I'm asking for them to design a horn for me, and this is certainly not the case. I'm a music and mathematics undergraduate attending an institution that doesn't have very strong STEM programs, and virtually no acoustics courses are offered, save for an "Introduction to Sound" course for non-STEM majors who's syllabus basically ends at describing that sound waves exist. As a result, I've had a grand total of about eight weeks to research this topic, and its produced about 250 pages of notes, learning circuit theory and the first three-ish chapters of Beranek and most of chapter 9. I've had to try to piece everything together myself, without anyone to ask questions to if there were gaps in my understanding; and with no formal background in physics or engineering, those gaps are bound to exist.
I'm realizing that I'm probably in over my head getting something quality finished by the deadline, but I've put in about 25 hours per week for this project, and I need to start building a prototype by Monday of next week to have something to show for it. To that end, all I was hoping for was for someone to tell me if my calculations were entirely off point for the horn, and point me in the right direction so that I might have a better understanding of what I did wrong. I'm not afraid of making mistakes, but I want to learn from them.
To touch on the other part of your post:
This is exactly what I'm hoping to find out; I'm not sure I understand the difference of a horn that "works" and a horn that is a "good idea"; speaking strictly in terms of impedance matching and not as a wave guide, isn't a horn that has a relatively stable and primarily resistive impedance throughout the desired operating range a horn that "works"? I realize the directivity, especially at higher frequency, will be less than desirable but I'm only aiming to explore the impedance matching aspect. Waveguiding can come in the future!
Thanks for the suggestions on the JAES articles and Solomans' paper; I need to find out if they'll come up in any of the databases my campus is subscribed to. Like I said, they're not big on STEM, so hopefully I can find something. For JAES, is there a particular place you'd recommend starting?
iirc, it's Webster's equations that are the basis for classical horn designs.
There are two parts (well, important parts) to horn design:
- loading/usable F3
- dispersion (or lack thereof) WRT frequency
A conic horn will load and have a usable F3, but may not have the dispersion WRT frequency that you (or one) might find desirable. A key idea is that as the frequency goes up, the energy tends to go straight ahead - aka "beam". You can see this in the polar response plots in Berenk (iirc).
Many of the newer designs are aimed at trading off some of that loading in favor of better HF dispersion characteristics.
You should look up diffraction horns too - another technique for trying to get better HF dispersion.
Then there is the oxymoronic "Constant Dispersion" vs. "Constant Directivity". Two terms that imho are idiotic to describe the polar response behavior of different horn objectives, and so designs.
Your LIBRARY can obtain the JAES Anthologies for you on loan, and likely the library has an automatic subscription to most journal sites on the web, so you can read them online. Audio Engineering Society.
There are two parts (well, important parts) to horn design:
- loading/usable F3
- dispersion (or lack thereof) WRT frequency
A conic horn will load and have a usable F3, but may not have the dispersion WRT frequency that you (or one) might find desirable. A key idea is that as the frequency goes up, the energy tends to go straight ahead - aka "beam". You can see this in the polar response plots in Berenk (iirc).
Many of the newer designs are aimed at trading off some of that loading in favor of better HF dispersion characteristics.
You should look up diffraction horns too - another technique for trying to get better HF dispersion.
Then there is the oxymoronic "Constant Dispersion" vs. "Constant Directivity". Two terms that imho are idiotic to describe the polar response behavior of different horn objectives, and so designs.
Your LIBRARY can obtain the JAES Anthologies for you on loan, and likely the library has an automatic subscription to most journal sites on the web, so you can read them online. Audio Engineering Society.
"A conic horn will load and have a usable F3, but may not have the dispersion WRT frequency that you (or one) might find desirable. "
Uhh, a conical horn will have the best dispersion at HF, better than an exponential, tractrix, etc.
You may be confusing a circular shape with a conical expansion.
Uhh, a conical horn will have the best dispersion at HF, better than an exponential, tractrix, etc.
You may be confusing a circular shape with a conical expansion.
djk, perhaps you would like to amplify your comment??
the maximum angle possible is equal to the angle of the conic section...
...geddes worked very hard to get even dispersion out to ~20deg off center axis.
I may have missed something along the way, happy to correct incorrect thinking.
_-_-
the maximum angle possible is equal to the angle of the conic section...
...geddes worked very hard to get even dispersion out to ~20deg off center axis.
I may have missed something along the way, happy to correct incorrect thinking.
_-_-
K402 does not appear to be conic...
even so, I am pretty sure that if one ran the polar response of that K402 one would not find any lack of beaming, unless there was a diffraction at the throat entry or something in the driver causing HF dispersion.
Can't seem to quickly find the AH! horns on a search, something came up but nothing definitive. I seem to recall reading about them before, maybe here?
Are these horns starting with a conic (straight side) expansion then converting to something else to limit issues at the mouth?
_-_-
even so, I am pretty sure that if one ran the polar response of that K402 one would not find any lack of beaming, unless there was a diffraction at the throat entry or something in the driver causing HF dispersion.
Can't seem to quickly find the AH! horns on a search, something came up but nothing definitive. I seem to recall reading about them before, maybe here?
Are these horns starting with a conic (straight side) expansion then converting to something else to limit issues at the mouth?
_-_-
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.