Paul W,
there is one little error in the sheet, look a few post earlier where I described it. And another, as just discovered, the driver troat radius is not transferred to the field that is used for calculation.
Thus we have bug fixes so far:
1) Cell D10 should read =C10*PI()/180 (calc radians of driver exit angle)
2) Cell C14 should read =C3 (transfer throat radius from input field)
- Klaus
there is one little error in the sheet, look a few post earlier where I described it. And another, as just discovered, the driver troat radius is not transferred to the field that is used for calculation.
Thus we have bug fixes so far:
1) Cell D10 should read =C10*PI()/180 (calc radians of driver exit angle)
2) Cell C14 should read =C3 (transfer throat radius from input field)
- Klaus
A few comments. I changed the spreadsheet quite quickly late one night after having had enough math for the day. As such I didn't take many steps to make it real readable.
C3 is a2 as I'd defined it in the spreadsheet. This is the exit radius of the CD. a1 is the original throat radius of the OS at X=0, and is in C15.
As Klaus points out, I had hard coded two cells when playing around with the spreadsheet. They should be referenced to the inputs.
I also added another column J, so that X could be shifted over and the contour drawn from where the CD would attach. This, however, would be done adhoc by making column K reference column J + xc.
It needs some cleaning up. However, this is a lot like my day job and this is a 3 day weekend, so I am not in a hurry to do so right now. 😉
I like the Cad realizations.
C3 is a2 as I'd defined it in the spreadsheet. This is the exit radius of the CD. a1 is the original throat radius of the OS at X=0, and is in C15.
As Klaus points out, I had hard coded two cells when playing around with the spreadsheet. They should be referenced to the inputs.
I also added another column J, so that X could be shifted over and the contour drawn from where the CD would attach. This, however, would be done adhoc by making column K reference column J + xc.
It needs some cleaning up. However, this is a lot like my day job and this is a 3 day weekend, so I am not in a hurry to do so right now. 😉
I like the Cad realizations.
Dr. Geddes,
Does the math predict how the mouth roundover influences the waves? Is there a best type of mouth roundover profile? How about at baffle edge with your large roundovers?
I was just kicking around ideas in my head about adding the Le Clea'ch style mouth roundover at the OS's exit, where it goes 180º back.
Does the math predict how the mouth roundover influences the waves? Is there a best type of mouth roundover profile? How about at baffle edge with your large roundovers?
I was just kicking around ideas in my head about adding the Le Clea'ch style mouth roundover at the OS's exit, where it goes 180º back.
soongsc said:
Except that is not what I said. I said that OS waveguides lose control when they are made wide. There are waveguides that work for wide patterns, just not OS waveguides.
JoshK said:
The math does not say that one needs to round the mouth edges, but it does tell you that if you don't you will get a large reflection from this termination. I don't think that the "roundover" going back is particularly effective, since the waves aren't going to go around this sharp of a bend very well. I radius all of my sharp edges just on principle - they look better and diffract less. They do make for difficult cabinet construction however.
That is why I am developing a new line of speakers. These will be made in wood as it is considerably less expensive to fabricate. I am building the prototypes now. These will not have as large a corner radius as that is what I have to give up to go to wood. Hence I will soon know to what extent the corner radius is an effect.
Incidentally, the concept is to offer these designs as kits where everything that one needs is supplied. You just put it together and paint it and you have a truely first quality loudspeaker. Cost are not firm yet, but I'm expecting to start at around $500 for a 10" system.
Back to the mouth radius. There is a mathematical discussion about the effect of this in my book if you are interested. Basically the radius tends to spatially smooth the polar response. This polar response effect can be simulated by a spatial convolution with a Gaussian function.
You are suggesting that the Theta of 45º is too wide (90º wall-to-wall)? You used 30º on your summa right?
JoshK said:
No I use 45°, but it appeared to me that the drawing was more like 60°. Excuse me if I'm wrong, I didn't actually measure it, but I have no idea if the scale is right anyways (screens are not always 1:1 aspect either).
Let me be more precise; 45° is just about the limit of what I have found works for an OS waveguide; 30° works a lot better and 60° a lot worse. But if the one shown is 45° then with a good driver it will work fine. It appears to be shorter than what I use and this is also a big factor.
Member
Joined 2003
Klaus,
Thanks!
Josh,
The LeCleach roundover seems to make sense...decreasing radius with increasing distance from the driver would seem to result in the lowest mouth reflection for a given overall diameter. When playing with the DDS ENG-90 waveguides last year, I could clearly see significant baffle diffraction artifacts on a baffle about 19" wide with 1/2" roundovers.
It would be really great if something like HornResp would allow the addition of various mouths on the OS. The directivity tool shows the pattern abruptly narrowing just before it collapses on the low end...which I suspect is the lack of any sort of mouth treatment.
Earl,
What are the largest OS waveguides you've worked with (depth and diameter)? What are the largest you considered functionally successful? Doesn't matter if they were commercially practical or not.
Thanks,
Paul
Thanks!
Josh,
The LeCleach roundover seems to make sense...decreasing radius with increasing distance from the driver would seem to result in the lowest mouth reflection for a given overall diameter. When playing with the DDS ENG-90 waveguides last year, I could clearly see significant baffle diffraction artifacts on a baffle about 19" wide with 1/2" roundovers.
It would be really great if something like HornResp would allow the addition of various mouths on the OS. The directivity tool shows the pattern abruptly narrowing just before it collapses on the low end...which I suspect is the lack of any sort of mouth treatment.
Earl,
What are the largest OS waveguides you've worked with (depth and diameter)? What are the largest you considered functionally successful? Doesn't matter if they were commercially practical or not.
Thanks,
Paul
gedlee said:
Considering that the AI website is up and running, I had the impression that the 12in version of the Summa would be available for purchase soon?
gedlee said:
This is an excellent news. It is possible to have a kit that just have the baffle and not the rest of the panel? For the rest of us who in NZ/OZ/EU sending the full box kit will cost an arm and legs.
Sam
Dr geddes,
I was looking on the AI website, and I have a question. What is causing the on axis dip at 6Khz on the 10 and 12in versions? the 15in seem to have a much more gentle dip lower in frequency.
I imagine you use smaller radii on the enclosure and horn terminus, but I would think that would show up at lower frequencies.
I was looking on the AI website, and I have a question. What is causing the on axis dip at 6Khz on the 10 and 12in versions? the 15in seem to have a much more gentle dip lower in frequency.
I imagine you use smaller radii on the enclosure and horn terminus, but I would think that would show up at lower frequencies.
Well, I never intended to agree only with what you said. But wide patterned wave guides really don't give much options for control to very specific distribution pattern desired without other tradeoffs.gedlee said:
Patrick Bateman said:
The web site is good news, but it reamains unclear if speakers will ever come out of Thailand. SO rather than wait, I decided to see if I could get something going locally.
Paul W said:
Or simply an inaccurate model.
Paul W said:
the Summa 15 has the largest that I have built in the US. I built some larger ones in Brazil, but they did not get in to production - the waveguide was too large.
I could make any size work within the right frequency range. The concept scales quite nicely, but then so does the frequency range. A larger throat has a much more limited high fequency response, while a larger mouth is only an advantage.
SamL said:
Sorry, I am focused only on the US market as the only one capable of supporting my business model. If you wanted to do something "down-under" that would be great. I would give you a license for my designs.
mbutzkies said:
This is an interesting question and one I argued some time back with Mr. Declerq.
In an axisymmetric waveguide there is a diffraction at the baffle edge. At some frequency the direct wave and diffracted wave will be exactly out of phase because of path length differences. This will cause a cancellation, but only directly on-axis. As the waveguide gets smaller, the diffraction gets greater and the path length difference gets smaller. So the hole moves higher in frequency and deeper in level. The data completely substantiates this, but Mr. Declerq did not seem to agree.
Soongsc,
"How accurate can paint tolerance be held?"
heh heh...Yeah, it wouldn't take much to obliterate it.
Earl,
"No I use 45°, but it appeared to me that the drawing was more like 60°. Excuse me if I'm wrong, I didn't actually measure it, but I have no idea if the scale is right anyways (screens are not always 1:1 aspect either)."
There may be some distortion going on there, but it's a 90 degree (45 per side). I resized the screenshots and that's a 4 inch mouth radius - larger than you would probably use on a 10 inch wg, and that gives it a disproportionate look. I was just checking the math and the process and cut the loop off at 5 inches in Y (r), slapped a 4 inch mouth radius in there for effect and surfaced it. Consider it a demo. 🙂
"How accurate can paint tolerance be held?"
heh heh...Yeah, it wouldn't take much to obliterate it.
Earl,
"No I use 45°, but it appeared to me that the drawing was more like 60°. Excuse me if I'm wrong, I didn't actually measure it, but I have no idea if the scale is right anyways (screens are not always 1:1 aspect either)."
There may be some distortion going on there, but it's a 90 degree (45 per side). I resized the screenshots and that's a 4 inch mouth radius - larger than you would probably use on a 10 inch wg, and that gives it a disproportionate look. I was just checking the math and the process and cut the loop off at 5 inches in Y (r), slapped a 4 inch mouth radius in there for effect and surfaced it. Consider it a demo. 🙂