Is there a formula for determining the ideal depth of a waveguide or horn on the basis of mouth and throat dimensions?
I am considering building a horn/waveguide optimized for constant directivity not loading.
I am considering building a horn/waveguide optimized for constant directivity not loading.
I think most would suggest a O.S. waveguide. then the waveguide dimensions and depth are set by the formula.
Thanks!
What’s the formula?
A spreadsheet where I can type mouth and throat dimensions and then get the depth would be great.
Hmmm..
I'm not sure how to use that sheet.
I have uploaded it to google sheets now. What does the different things mean in "real words"?
The beamwidth of a waveguide is the angle of the walls.
That's it.
When I first saw that, I thought it had to be more complex than that, but it isn't.
For instance, a conical waveguide measuring 90 x 90 degrees will have a beamwidth of 90 x 90 degrees.
May be easiest would be to just use Hornresponse for calculation of your WG. Hornresp can do OS-WGs and will export dimensions in xls (or csv) format (do not remember exactly).
If you want to optimize for constant directivity, you either have to start over with simulations, or build and measure.... Both may be fun.
If you want to optimize for constant directivity, you either have to start over with simulations, or build and measure.... Both may be fun.
Beware that the critical aspect of any waveguide design is its termination into a free space. No simple formula will give you that. If you are interested I can make a TXT/CSV export function for a selected profiles in my tool, it calculates OS terminated by a clothoid: ATH - Advanced Transition Horns
Nice to wake up to a bunch of answers!
I haven’t told you the whole story about what I was considering building. For a long time I have read and re-read and re-re-read a lot of threads regarding waveguides, synergy horns and directivity. When contemplating what the ideal speaker would be, I keep coming back to point source and constant directivity.
Any way, my idea was to build a frame shaped as a traditional floor standing speaker. (I have to fit it into a living room so very wide guides as the JBL 2384 is unfortunately too big) In the frame I would stretch some 4-way stretch fabric and place a dowel with the diameter of the compression driver exit. Push it to the back of the frame where a board is placed. Fix it and then start spray painting the fabric with 2 comp clear coat. Keep doing that until I have a horn. The frame would be the horns permanent speaker stand. It is my understanding, that doing this will give me the optimal transition from square frame to circular exit. Is that right? About the termination of the horn. That was actually my next question. I need to do some round over. But how big and what shape?
Back to the depth thing. Bateman mentions that coverage angle determines the rest. But are there some coverage angles that minimize HOMs more than others. Seems to me that a very narrow coverage would increase distortions? And what about the second flare on danleys horns. Is that necessary when using the stretch fabric method?
I haven’t told you the whole story about what I was considering building. For a long time I have read and re-read and re-re-read a lot of threads regarding waveguides, synergy horns and directivity. When contemplating what the ideal speaker would be, I keep coming back to point source and constant directivity.
Any way, my idea was to build a frame shaped as a traditional floor standing speaker. (I have to fit it into a living room so very wide guides as the JBL 2384 is unfortunately too big) In the frame I would stretch some 4-way stretch fabric and place a dowel with the diameter of the compression driver exit. Push it to the back of the frame where a board is placed. Fix it and then start spray painting the fabric with 2 comp clear coat. Keep doing that until I have a horn. The frame would be the horns permanent speaker stand. It is my understanding, that doing this will give me the optimal transition from square frame to circular exit. Is that right? About the termination of the horn. That was actually my next question. I need to do some round over. But how big and what shape?
Back to the depth thing. Bateman mentions that coverage angle determines the rest. But are there some coverage angles that minimize HOMs more than others. Seems to me that a very narrow coverage would increase distortions? And what about the second flare on danleys horns. Is that necessary when using the stretch fabric method?
The frame, dowel and stretch method is what I want to try also one fine day. But I would use it to create a mold instead of the direct horn to be used later on.
Edit: Forgot to mention that I would use the exiit angle of the driver to determine how far to push the dowel.
Regards
Charles
Edit: Forgot to mention that I would use the exiit angle of the driver to determine how far to push the dowel.
Regards
Charles
Last edited:
For a driver exit angle to match, I would just shape the dowel to conical. The depth of the pushing will more affect the coverage angle. It is an interesting idea, I have seen it realized several times, also here at diyAudio (don't remember where). I guess it could work quite well after all.
Yes, this is a possibility. My idea was to make a simple carboard template that is used to check whether the initial curve shape is at the desired angle or not. If the angle is too wide one just pushes a little more until it fits.
Regards
Charles
Interesting idea!
Could you provide me/us with an example of how it would look? Exit angles are typically from 10 to 20 degrees or something, right?
I am speaking of a simple cut out piece of cardboard or the like with a shape like this:
___ ___
| \ / |
| \____/ |
|____________|
You just use it to check the angle of the waveguide at its base when you push the dowel in.
I was just thinking About the other metod with the angle milled into the dowel: Ist Advantage would be that one does not have to take care of the throat angle during the stretching process but the mouth angle instead.
regards
Charles
Edit: Tried different fonts but it doesn't want to show up the same way as in the Editor. Will make a proper drawing.
___ ___
| \ / |
| \____/ |
|____________|
You just use it to check the angle of the waveguide at its base when you push the dowel in.
I was just thinking About the other metod with the angle milled into the dowel: Ist Advantage would be that one does not have to take care of the throat angle during the stretching process but the mouth angle instead.
regards
Charles
Edit: Tried different fonts but it doesn't want to show up the same way as in the Editor. Will make a proper drawing.
Last edited:
Code:
[FONT=Courier New]___ ___
| \ / |
| \____/ |
|____________|[/FONT]Years ago, Tom Danley posted a similar method to make these:
I tried to find the post, with no luck. IIRC, he said that he used a fabric that stretched. This was used as a mold I assume.
An externally hosted image should be here but it was not working when we last tested it.
In the car audio world, if you've ever seen fiberglass boxes and dashes that are all swoopy and smooth, they basically do that by the method that you describe. They use the same type of material that's used for sweatshirts. It's a mix of polyester and spandex IIRC. Basically it's thick and stretchy. You want that, so that you don't have to lay down ten coats of epoxy.
A perfectly conical horn has a really gnarly transition from the horn to the room. If you've heard someone speaking through a megaphone, you know what I'm talking about.
If you have a conical horn with a mouth that measures 17cm in diameter, it will control directivity down to about 2khz. (speed of sound / length of wave.)
As the waveguide loses control, it does two things:
* first it narrows
* and then it broadens
If you check out measurements I've posted, you'll see this happen on the low end of the waveguide's polar response.
So when you "flare" the waveguide at it's termination, it fights these effects. Basically it widens the beamwidth at the point where the beamwidth is starting to narrow.
There's a Keele AES paper about this, you can find it on xlrtechs dot com. I'm too lazy to look it up 🙂
In regards to HOMs, I personally believe that wider angles tend to have lower HOMs, but there's quite a debate about that :
The Preference for Direct Radiators
- Status
- Not open for further replies.