Really good performance in the top octave from dome tweeter inside a non-diffracting waveguide.
Ok so I need to separate throat angle from flare. @docali I think maybe you translate flare incorrectly? I get what you mean by "Drag curve" but your comment lead me to some readings suggesting that "friction" isn't necessary to calculate a tractrix curve per say?
How do I incorporate T and throat angle into the tractrix formula?
I think I finally backed chatgpt into a corner on one of these aspects lol. T to be exact.
Last edited:
OK, I don't follow. What does "friction" have to do with a mathematical equation? "Flare" simply means an expanding cross section.
OK, that part of the comment was directed at Docali who brought the idea that Tractrix was a "drag curve".... I probably wanted to figure out the friction part of it, being that, how do you "drag" without the presence of friction, so I likely googled tractrix horn and friction
I found things like this
But as I continued to read different sources, I came across the information suggesting that friction is not apart of the equation.
"The mathematical description of a tractrix does not even mention friction.
The tractrix is defined as a curve with the property that the distance from any point on the curve to a fixed straight line measured along the tangent of the curve is constant.
This can be interpreted in classical mechanics as the path of an object that is pulled by a string in certain circumstances -- and for this interpretation one probably needs to assume that the intertia of the object is negligible compared to the friction, such that its direction of movement at any point is exactly the direction of the string that pulls it. However, this is purely a matter of interpretation and not part of the definition of the curve, which is entirely independent of such physical concepts."
I found things like this
But as I continued to read different sources, I came across the information suggesting that friction is not apart of the equation.
"The mathematical description of a tractrix does not even mention friction.
The tractrix is defined as a curve with the property that the distance from any point on the curve to a fixed straight line measured along the tangent of the curve is constant.
This can be interpreted in classical mechanics as the path of an object that is pulled by a string in certain circumstances -- and for this interpretation one probably needs to assume that the intertia of the object is negligible compared to the friction, such that its direction of movement at any point is exactly the direction of the string that pulls it. However, this is purely a matter of interpretation and not part of the definition of the curve, which is entirely independent of such physical concepts."
Last edited:
Posting to my own post, because I've owed everyone a response
Ok, the four 3/4" dia mid ports form a rectangle 5.25" wide, and 3.25" high, with diagonals between corners of 6.2".
Using the diagonals as the max c2c distance, that equates staying with 1/4 WL to about 550 Hz.
So if 1/4 WL summation really holds, I would expect no change in response under rotation up to 550Hz, rotation which puts one side's mid ports closer to mic than the other sides ports..
And when I finally ran this test properly....that's what I got !!!! Summation held up fine.
Trying to present the best data possible, and wondering if any mid port polar response variations I had seen previously were due to larger ports' interferences, I blocked all low and sub ports on syn11.
Used sheet metal and duct tape. (Have found in past tape only works for very low levels.)
So the mid ports were essentially measured on a 75x60 conical, with no ports.
Raw response looked quite usable from 150Hz to 1kHz.
I normalized on-axis response, both mag and phase, via FIR, with about 1/12th oct resolution.
Below are 0, 10, 20, 30, and 40 degrees horiz.
In my book, they are dang equivalent, surely to 500Hz if not a bit beyond.
(Angles are not really correct because POR (point of rotation) is forward of the CD's acoustic origin. My turntable isn't big enough to get that right on syn11.
But I don't think that matters at all for testing 1/4 WL summation.)
Anyway, it appears to me I stood in error thinking 1/4WL does not inseparably sum under rotation. Port interferences were blinding me.
Oh, I should add...on seeing great far-field summations with all ports blocked, I moved indoor to test up close where rotation would exacerbate any issue.
At 40 degrees, the closer side's ports were 18.5" from mic, the farthest 20.75". Sure seems like 1/4 WL holds to me ! (Below is the indoor, close up set.)
This test has really turned out to be a blessing for me......
After seeing how well all driver sections hold up after processing, with no ports in play, I now use this as the first stage in quasi-anechoic tuning.
On axis EQs hold up very nicely horiz, vert, and even in pattern corners.
So I now tune each section this way, starting with the CD, with all lower frequency sections ports blocked..
Then, I unblock all the ports and use a second stage of IIR EQs to smooth any anomalies on each section-by-section, caused by port interferences.
It's relatively easy to quickly identify what port interferences can be corrected, when they have been split apart from the driver section corrections on the unported horn.
The previous 'lump sum method', correcting both the 'match to conical horn and handle port interferences at the same time', worked;
but was so much harder with a lot of trial and error experimentation.
The new two-stage approach helps to quickly bin what should and shouldn't be corrected at each stage, and what holds up over the pattern.
And I think it helps better visualize the prime factors needing correction, and match appropriate corrections.
Waaay less juggling
Last edited:
Hi Mark
Have been waiting and checking out your SYN11 thread for a while, and here we go. Very interesting.
It seems that you have changed your angle of tuning from 10 degrees of axis to on axis, is that correctly understood?
I hope you copy your post over in the SYN11 thread, so that the discussion can be continued there, as that thread already has summed up a lot of good synergy/MEH-knowledge in one place.
Thanks for sharing
Steffen
Have been waiting and checking out your SYN11 thread for a while, and here we go. Very interesting.
It seems that you have changed your angle of tuning from 10 degrees of axis to on axis, is that correctly understood?
I hope you copy your post over in the SYN11 thread, so that the discussion can be continued there, as that thread already has summed up a lot of good synergy/MEH-knowledge in one place.
Thanks for sharing
Steffen
Hi Steffen, you are very observant, catching my use of different angles used for a tuning reference.
My longstanding use of a reference angle somewhere close to degrees, has been because it best represented the mean of averaged responses in a 0-30 degree window. Using it for tuning minimized the +/- amplitude of ripple around 0dB, for the set of 0-30 polars.
When I said to the forum, that I've been wondering if 1/4 WL summation is a rule of thumb for good acoustic coupling, or does 1/4 WL summation actually give inseparable summation regardless of angular rotations.......it was due to seeing ripple, when I thought there shouldn't be any if inseparable summation occurs.
To keep my speculations minimized, and trying to come up with the best test I could think of for testing 1/4 WL summation, it hit me I HAD to remove ports from the equation, as a source of potential ripple.
Covering all the ports quickly delighted me, how relatively ripple free each driver section behaved.
There is no significant difference between 0 degrees and 10 degrees for either mid or low sections.
The coax CD (shown below) has the most ripple, but the 75x60 horn doesn't have secondary flares, and I haven't bothered to carefully match Cd exit to horn flare, yet.
(note the 40 degree orange trace below is outside the 75 degree horn pattern; but due to point-of-rotation, trace is really more like 35 degrees.)
Anyway, for this 1/4WL summation test with ports covered, I could use either 0 degrees of 10 degree (or almost any thing below 30 degrees) for reference .
What's really cool, and I'll elaborate on the syn11 thread, is that I've found using FIR for this stage like always, allows identifying and separating the port disturbances from any basic driver section response rolloffs.
I'm optimistic that a new level of quasi-anechoic tuning has been stumbled on....
And yes, fully moving back to the syn11 thread, other than responding here as needed re this 1/4 WL summation test.
B&C dcx464 on 75x60 conical; 500Hz 16th order LR lin phase high-pass
My longstanding use of a reference angle somewhere close to degrees, has been because it best represented the mean of averaged responses in a 0-30 degree window. Using it for tuning minimized the +/- amplitude of ripple around 0dB, for the set of 0-30 polars.
When I said to the forum, that I've been wondering if 1/4 WL summation is a rule of thumb for good acoustic coupling, or does 1/4 WL summation actually give inseparable summation regardless of angular rotations.......it was due to seeing ripple, when I thought there shouldn't be any if inseparable summation occurs.
To keep my speculations minimized, and trying to come up with the best test I could think of for testing 1/4 WL summation, it hit me I HAD to remove ports from the equation, as a source of potential ripple.
Covering all the ports quickly delighted me, how relatively ripple free each driver section behaved.
There is no significant difference between 0 degrees and 10 degrees for either mid or low sections.
The coax CD (shown below) has the most ripple, but the 75x60 horn doesn't have secondary flares, and I haven't bothered to carefully match Cd exit to horn flare, yet.
(note the 40 degree orange trace below is outside the 75 degree horn pattern; but due to point-of-rotation, trace is really more like 35 degrees.)
Anyway, for this 1/4WL summation test with ports covered, I could use either 0 degrees of 10 degree (or almost any thing below 30 degrees) for reference .
What's really cool, and I'll elaborate on the syn11 thread, is that I've found using FIR for this stage like always, allows identifying and separating the port disturbances from any basic driver section response rolloffs.
I'm optimistic that a new level of quasi-anechoic tuning has been stumbled on....
And yes, fully moving back to the syn11 thread, other than responding here as needed re this 1/4 WL summation test.
B&C dcx464 on 75x60 conical; 500Hz 16th order LR lin phase high-pass
Hi Mark
Well, I guess I am a bit obsessed with MEH´s. Interesting finding, to separate rav driver-response in the horn and port-disturbances.
I have come up with an idea for a MEH-design with secondary flares to put in a corner and have all drivers on removable top- and bottum-plates! Hope it works, I am in the proces of building a 80x60 testhorn in 16 mm MDF, approximately 55x73 cm. Completely arbitrary dimensions, just what the pile of old MDF material could provide. I will share some pictures for inspiration in the SYN11-thread when I have it finished.
Steffen
Well, I guess I am a bit obsessed with MEH´s.
Interesting finding, to separate rav driver-response in the horn and port-disturbances.I have come up with an idea for a MEH-design with secondary flares to put in a corner and have all drivers on removable top- and bottum-plates! Hope it works, I am in the proces of building a 80x60 testhorn in 16 mm MDF, approximately 55x73 cm. Completely arbitrary dimensions, just what the pile of old MDF material could provide. I will share some pictures for inspiration in the SYN11-thread when I have it finished.
Steffen
8" fullrange should be able to deliver 95-100dB in a 500Hz horn and closed back. Of course not everybody like whizzer cone but you can use coax instead.
I think this combo should be perfect for this thread. I'm working on novelity horn with acoustic filters built in, no electrical that is easy.
I think this combo should be perfect for this thread. I'm working on novelity horn with acoustic filters built in, no electrical that is easy.
Attachments
Last edited:
No, technically its a three way as the four midranges will act like a single source.