Hi Gedlee,
Yes, I would not expect to give us in the 'close form formula', but I believe that, may be, You have some calculated (or measured) value for some usual throat size ( say: 1", 1.5" and 2" throat DIAMETER) and the angles say 30 deg, and 45 deg.
regards
ivica
Hi Gedlee,
I have tried to do some calculations, and on the attached figure is the result,
but can You confirm.
Only what I have find is the presentation for 1-inch throat and 30 deg off-axis wall.
As shown in the post:
http://www.diyaudio.com/forums/multi-way/103872-geddes-waveguides-704.html#post4038689
Seems to me very comparable with my "work".
Bouncing on the curves is due to limited horn length, and mouth without proper "termination".
Regards
Ivica
Attachments
Last edited:
Ivica
Those certainly seem reasonable and if one of them matches my exact calculations then I would think that you are "good to go".
I have never given the impedance much interest. It was always the directivity that I was interested in. The device just gets EQ'd in the end so the impedance is just a secondary thing.
Those certainly seem reasonable and if one of them matches my exact calculations then I would think that you are "good to go".
I have never given the impedance much interest. It was always the directivity that I was interested in. The device just gets EQ'd in the end so the impedance is just a secondary thing.
Last edited:
Hi Gedlee,
I agree that in any case using CD-WG would need EQ, but lower frequency (applied to the horn-driver combo) limitation can be get from the acoustic impedance curve, in order not to overload the driver with too low frequency.
Regards
Ivica
The thing is that for home use this is usually not an issue. The one thing that I would check is that your curves seem to be pretty high in frequency. The 1" 45 degree, which is what I always use, is shown to have its peak at about 3 kHz. This seems high as I recall, but could be possible. I always have a cut of about 6-8 dB at 2500 Hz - over the natural HP cut at these lower frequencies. I can usually get down to 700-800 Hz in the larger waveguides with a 1" driver. I have never had a LF issue, but in pro applications I have seen some drivers burned out from HFs. Usually when the amps were clipping heavily. Nothing is there to protect from this HF noise that just heats up the coil.
ivicai
My book GedLee LLC Shows the 30° and 45° impedances for the first three modes plotted versus C = 2 Pi f a / c / sin(theta) - figures 6-9 and 6-11. With these two plots you can get all the curves you are looking for. I can't make it any easier than that. It does appear from this data that your curves are way too high.
My book GedLee LLC Shows the 30° and 45° impedances for the first three modes plotted versus C = 2 Pi f a / c / sin(theta) - figures 6-9 and 6-11. With these two plots you can get all the curves you are looking for. I can't make it any easier than that. It does appear from this data that your curves are way too high.
Last edited:
Hi Gedlee,
I have done rescaling my calculation concerning the define variable "C" = 2*Pi*f*Ro / (Co*sin(theta)), where Ro= throat RADIUS, Co, speed of sound in the air, f- is frequency.
I have attached the redraw figures for 2-inch driver ( Ro=1inch) with 30deg, and 45 deg off-axis angle. Looking at the mentioned figures 6-9 and 6-11 in Your book (Chapter-6.) it seems to me that my calculations are OK, but my previous attached figures are more practical for the DIY members as the acoustic impedance are shown as a function of the frequency.
I believe that on both figures You have mentioned (6-9, 6-11) from your book, are calculated for the 2-inch throat radius OSWG horn.
Regards
Ivica
Attachments
Last edited:
Since the figures in my book are functions of c which contains the throat radius, they are valid for any throat radius. But if you look at the calculations in the book - from c to f - you will see that the peak in the 45 degree 1" curve occurs between 1 and 2 kHz. Half power is about 500 Hz. That is much lower than your previous curves showed.
Last edited:
Hi Gedlee,
Yes, You are right, I have done the rest of the calculations ( for 1-nch driver) and after 'rescaling' the graph the results are the same, so I believe that we can use acoustic impedance graphs vs frequency as correct presentations of Your work on OSWG horns.
May be someone on the forum would be interesting, especially for lower frequency applications of the compression drivers with OSWG type horn.
Regards
Ivica
Hi Gedlee,
Yes, You are right (as usually), that in the books there have to be put more general solutions, but from time to time it would be useful (for the reader) to emphasize some practical suggestions, or conclusions. Such would make reading theoretical explanation to become more realistic. In mentioned Chapter.6, of the book, You are talking about HOMs, but as acoustic impedance here as derivative, is interesting thing to be noted, I think. So a kind of comparison can be made to the other known results.
So here forum member can use your results from the book, and in order to get real usual presentation impedance as a function of the frequency they have to recalculate all the data, or can use mine graphs presented in the:
http://www.diyaudio.com/forums/atta...waveguides-horn-impedance-vs-angle-throat.gif
for some most usual horn throat size and directivity angles.
Many thank for your cooperation and willing to help.
Regards
Ivica
Last edited:
Fortunately (or not) using the drivers with internal short horn, as JBL 2440, 2441, 2445, 2446, 2450, 2420/21/25/26 (and equivalent from the other producers) this short horn would improve LF loading. For example JBL 2446 with 46mm internal short horn (that would couple 1.5inch output from the phase plug to the driver 2-inch mouth)
regards
ivica
Attachments
How critical is a changing aspect ratio in a square/rectangular waveguide?
Eg: If I maintain the expansion rate of the equivalent 90 degree cone this is supposed to match, starting with a circular baffle cutout with a square circumscribed about it. The flat sides are 90 degrees apart, and the top/bottom 60 degrees.
The cutout has a radius of 8.5cm, and the CSA coincides with the target at an axial distance of 10cm. So the corners can be filled until this point, and the sides can be radiused over the next 10cm where the aspect ratio will be almost 1.6:1, then the top/bottom.
Would I be better reducing the height of the throat to make it rectangular as well?
That would seem to go against convention. I also plan to cross below 1kHz.
Eg: If I maintain the expansion rate of the equivalent 90 degree cone this is supposed to match, starting with a circular baffle cutout with a square circumscribed about it. The flat sides are 90 degrees apart, and the top/bottom 60 degrees.
The cutout has a radius of 8.5cm, and the CSA coincides with the target at an axial distance of 10cm. So the corners can be filled until this point, and the sides can be radiused over the next 10cm where the aspect ratio will be almost 1.6:1, then the top/bottom.
Would I be better reducing the height of the throat to make it rectangular as well?
That would seem to go against convention. I also plan to cross below 1kHz.
Bear in the Woods
The horn will behave differently in the horizontal and vertical planes. You can call it acoustical astigmatism if you like and it is not necessarily going to behave like you intend. You can get a frequency dependent rotation (flip) of the dispersion pattern with such geometries unless the bandwidth of the drive signal is appropriately limited. Any discontinuity in the throat area will encourage the generation of HOM's as well. Beyond these general comments a drawing showing horn geometry will be needed to obtain specific information. WHG
The horn will behave differently in the horizontal and vertical planes. You can call it acoustical astigmatism if you like and it is not necessarily going to behave like you intend. You can get a frequency dependent rotation (flip) of the dispersion pattern with such geometries unless the bandwidth of the drive signal is appropriately limited. Any discontinuity in the throat area will encourage the generation of HOM's as well. Beyond these general comments a drawing showing horn geometry will be needed to obtain specific information. WHG
Last edited:
Allen
I am not getting a very good idea of the question, but it is more the rate of change of the contour than the rate of change of the area that is important. Any discontinuities in either of these rates creates HOMs and internal reflections. I would also guess, although I am not as sure, that the aspect ratio should not change abruptly, although slow gradual changes are not likely to be a big problem. I would stick to as little aspect ratio and cross-sectional shape changes as possible however.
I am not getting a very good idea of the question, but it is more the rate of change of the contour than the rate of change of the area that is important. Any discontinuities in either of these rates creates HOMs and internal reflections. I would also guess, although I am not as sure, that the aspect ratio should not change abruptly, although slow gradual changes are not likely to be a big problem. I would stick to as little aspect ratio and cross-sectional shape changes as possible however.
As the last two comments illustrated, you want a slooooooow transition from one angle to another. IMHO, a reduction in HOMs makes a horn-loaded loudspeaker sound 'smoother' and less fatiguing. I've had SH50s and Summas in my home at the same time to learn this for myself.
An externally hosted image should be here but it was not working when we last tested it.
I posted a thread on Sunday which shows how to achieve that sloooooow transition with an elliptical waveguide. The method I posted works for cones and compression drivers. Note that the WG will look different in each case, because the shape is dependent on the exit angle of the radiator. (IE, a compression driver will have an exit of something like ten degrees, while a cone will have an exit around 120 degrees.)
Many of the rectangular waveguides used by JBL these days are simply an evolution of the biradial constant directivity horns. You can see this in the JBL M2; it still has a diffraction slot. More importantly, the transition from the throat of the horn to the mouth of the horn isn't perfectly seamless like it is in the Genelec. The problem is in the corners; because the PT waveguides are composed of four radial curves, there's a problem in the corners.
Basically the 'crease' you get in the corners of the rectangular waveguide probably cause some problems.
I know that this sounds like splitting hairs, but it really isn't. For instance, I designed and printed some waveguides in the style of the JBL M2, and they did NOT measure as good as the QSC rectangular waveguide.
If you look at the QSC, you'll notice that it's one loooooong seamless curve from throat to mouth. It's not perfect, but it's more seamless than the M2. (The M2 does some other things better though, and for that reason, I continue to use that curve the most.)
Yes, it does. I was asked to measure one of these more rectangular waveguides and along the diagonal (which no one ever measures) the response control was not very good at all. Far worse than either the vertical or the horizontal, neither of which was as good as a symmetric waveguide. This is not the way to make a great waveguide.
Another thing I've noticed with these plastic waveguides is high distortion.
I'm still trying to figure out what the problem is, but right now my theory is that the poorly damped plastic waveguide 'rings' when energy from the woofer causes it to vibrate.
The jury is still out, but that's my theory.
I noticed this recently, as I've been doing more distortion measurements, and I noticed that my midranges were distorting way more than the spec sheets indicated.
IE, the spec sheet indicated distortion that was twenty or thirty dB below the fundamental, and when I measured the distortion in a cabinet with a waveguide, the distortion was much much higher.
And then I got to thinking about what causes distortion, and it seemed that the motion of the midrange may be inducing vibration in the waveguide, which is then radiated.
Basically the plastic waveguide is ringing like a passive radiator.
It's ugly stuff and I'm having a heck of a time dealing with it. (Again, still not sure that it's the source of distortion.)
Patrick
I have no data on this, but my suspicions are with yours. As you know my waveguides are extremely rigid. Not at all like the usual fiberglass or injection molded devices, mine are several inches thick where it matters. (Which requires a unique mounting technique, as you know.) You cannot break one of my waveguides even with a sledge hammer - I know, I tried, and I once thought of posting a video of this but never did. Is this a factor in how good they sound? I am not sure, but I would certainly not be surprised if it turned out to be the case.
I have no data on this, but my suspicions are with yours. As you know my waveguides are extremely rigid. Not at all like the usual fiberglass or injection molded devices, mine are several inches thick where it matters. (Which requires a unique mounting technique, as you know.) You cannot break one of my waveguides even with a sledge hammer - I know, I tried, and I once thought of posting a video of this but never did. Is this a factor in how good they sound? I am not sure, but I would certainly not be surprised if it turned out to be the case.