Hi a3cervo,
I asked a while ago whether Marcel publicly released it, but judging by the answers, only the smaller CE360 is publicly available. Please, let me know if I am mistaken and you will find it.
Kindest regards,
M
Hi danibosn,
can you please amplify?
As understood one would have to first somehow convert the *.stl file to a 3D drafting software, e.g., Solidworks do derive the profile curve. Now what? One cannot just re-scale the entire drawing because everything, including the throat would be re-scaled.
Kindest regards,
M
can you please amplify?
As understood one would have to first somehow convert the *.stl file to a 3D drafting software, e.g., Solidworks do derive the profile curve. Now what? One cannot just re-scale the entire drawing because everything, including the throat would be re-scaled.
Kindest regards,
M
He means to use the CE360 configuration file and make a change to the length parameter to scale the waveguide up. That may not be exactly the same as the different length might want slightly different parameters, that mabat optimized with the CE460.
https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-6767343
Hi fluid,
well, we do not know what danibosn meant. However, considering, arguendo, that you are correct, if it were that simple, an optimized set of parameters (for a given requirement), could be used for different sized devices, which I do not believe is the case based on my limited experience wit the ATH.
Kindest regards,
M
well, we do not know what danibosn meant. However, considering, arguendo, that you are correct, if it were that simple, an optimized set of parameters (for a given requirement), could be used for different sized devices, which I do not believe is the case based on my limited experience wit the ATH.
Kindest regards,
M
Dear M.,
this is correct, just as fluid and danibosn told you before, upscaling the overall device length does usually create very similar results compared to the original source file. The same goes for changing the throat diameter. The results may or may not be optimized. It is a very simple process.
this is correct, just as fluid and danibosn told you before, upscaling the overall device length does usually create very similar results compared to the original source file. The same goes for changing the throat diameter. The results may or may not be optimized. It is a very simple process.
Hi sheeple,
thank you for the response.
I agree that the process per se is simple, after all, even I can do it. What is not simple is the time for the calculation, and reviewing the results. It would be nice, if the application had some command line interface, so that one could write a simple script, that would change some of the parameters, e.g., for example not the throat diameter and angle of the driver, and length of the device, let it run and then look at the curves and refine the most promising looking ones. If one started form a design that Marcel has already generously provided, this would quickly yield an optimized design.
Please do not misunderstand, I am not complaining that such feature does not exist, I am just responding to your term "simply".
Kindest regards,
M
thank you for the response.
I agree that the process per se is simple, after all, even I can do it. What is not simple is the time for the calculation, and reviewing the results. It would be nice, if the application had some command line interface, so that one could write a simple script, that would change some of the parameters, e.g., for example not the throat diameter and angle of the driver, and length of the device, let it run and then look at the curves and refine the most promising looking ones. If one started form a design that Marcel has already generously provided, this would quickly yield an optimized design.
Please do not misunderstand, I am not complaining that such feature does not exist, I am just responding to your term "simply".
Kindest regards,
M
It literally is a command line application?
What you probably mean is a batch routine. However, with an axisymmetric device, calculation time can be optimized to a few of seconds and you will have to review results and make decisions frequently, anyhow. If you don’t, you will not learn how to control the parameters and how to reach your goals.
Only @maiky76's optimizer might automate this away, but then you'd have to learn how to control this program.
I do not see what could be gained from automation with such a computation friendly device as the acisymmetricals. A different thing with 3D meshes.
What you probably mean is a batch routine. However, with an axisymmetric device, calculation time can be optimized to a few of seconds and you will have to review results and make decisions frequently, anyhow. If you don’t, you will not learn how to control the parameters and how to reach your goals.
Only @maiky76's optimizer might automate this away, but then you'd have to learn how to control this program.
I do not see what could be gained from automation with such a computation friendly device as the acisymmetricals. A different thing with 3D meshes.
Hi sheeple,
Thank you for the reply.
Thinking about it more, the command line invoking the ATH calculation reads the configuration file and carries out the calculation. So, it could be feasible to write a script, changing the parameters in the configuration file and repeatedly call it from the command line, saving the results. Unfortunately, I am only a dumb user of Windows, all my scripting/programming is done in UNIX-like environment. Perhaps I look into Windows scripting, when I have some time.
Kindest regards,
M
Thank you for the reply.
I do not want to get into semantics, but when I wrote command line, I meant that the options, in this case the parameters, may be specified on the command line.
Thinking about it more, the command line invoking the ATH calculation reads the configuration file and carries out the calculation. So, it could be feasible to write a script, changing the parameters in the configuration file and repeatedly call it from the command line, saving the results. Unfortunately, I am only a dumb user of Windows, all my scripting/programming is done in UNIX-like environment. Perhaps I look into Windows scripting, when I have some time.
Well, that is a matter of opinion. I can clearly see the advantage. Consider, that one would like a certain shape of DI. One could then write an error function and optimize that.
Kindest regards,
M
Hi
I've been couple a days away but I meant exactly what fluid and sheeple mentioned. Several people, here on forum, did big amount of tries (Marcel specially) and we found almost the best shape (sometimes on different way with with different parameters but very similar). That shape you can scale only by changing the lenght in script. Of course you are welcome to try to optimize and share if you are willing but per experience the differences will be minimal.
I've been couple a days away but I meant exactly what fluid and sheeple mentioned. Several people, here on forum, did big amount of tries (Marcel specially) and we found almost the best shape (sometimes on different way with with different parameters but very similar). That shape you can scale only by changing the lenght in script. Of course you are welcome to try to optimize and share if you are willing but per experience the differences will be minimal.
I can't follow. Did you add "lua54.dll" to the ath folder and then it worked again? For me this does nothing.
Hi mabat, you wouldn't still have those dimensions, would you?
Adding an enclosure to a waveguide seems to be missing something. I added the lines from 6.12.1 to demo1.cfg and it doesn't create an enclosure for me. https://pastebin.com/KnYJ4Rzq
You have your ABEC sim type as 1 which is infinite baffle, try changing it to 2 for Freespace. Otherwise have a look at this example to see if you can spot any differences.
https://www.diyaudio.com/community/...-design-the-easy-way-ath4.338806/post-6996527
I am not very proficient with the slot feature. Can someone help me with what the required modifiers to the OSSE parameters are, to insert into an oval waveguide only on the horizontal axis -> wide and deep cheeks?
something like this would be the the starting point:
And I want to incorporate cheeks to widen the pattern, like here:
something like this would be the the starting point:
And I want to incorporate cheeks to widen the pattern, like here:
A great deal of it is just trial and error.
Take a look at something like this: Slot.Length = 1 + 24.4*cos(2*p)^2
Add that line to nearly any of the waveguides will basically create a slot where the depth of the slot will be long on the X and the Y axis, and short on the diagonals.
Do a search here and you'll see other examples of adding a slot: https://www.google.com/search?q=site:diyaudio.com+"slot.length"
Fair warning: I've never come up with a waveguide that performed better, as far as polar response and frequency response goes, than one WITHOUT a slot.
In the case of the JBL speakers, I think they're largely using the slot for marketing reasons and also to create a waveguide that has wider beamwidth than a conventional waveguide of equivalent size. The JBL design basically "robs from Peter to pay Paul." It widens the beamwidth on the X and Y axis by narrowing the beamwidth on the diagonal axis, at the expense of overall smoothness.
Take a look at something like this: Slot.Length = 1 + 24.4*cos(2*p)^2
Add that line to nearly any of the waveguides will basically create a slot where the depth of the slot will be long on the X and the Y axis, and short on the diagonals.
Do a search here and you'll see other examples of adding a slot: https://www.google.com/search?q=site:diyaudio.com+"slot.length"
Fair warning: I've never come up with a waveguide that performed better, as far as polar response and frequency response goes, than one WITHOUT a slot.
In the case of the JBL speakers, I think they're largely using the slot for marketing reasons and also to create a waveguide that has wider beamwidth than a conventional waveguide of equivalent size. The JBL design basically "robs from Peter to pay Paul." It widens the beamwidth on the X and Y axis by narrowing the beamwidth on the diagonal axis, at the expense of overall smoothness.
Another thing that trips me up is that I often forget the results are in a new directory
I'll sit there poring over the config file on the command line and then realize the freespace results go in a different directory
Yes, I know the basic formula and removing the multiplier for p results in purely horizontal bulges. Multiplier for cos seems to define depth / dimension and a smaller exponent makes them less slotty abd more bulgy. But I never get as far as that they become pronounced, and most times the waveguide looses shape. I applied a guiding curve but then, when I try to make the cheeks, ath fails to mesh.
This is not about slots for HF, it is for widening horizontal directivity. My results for a pattern wider than 50 degree where not as good as I had hoped, so I want to try this now.
How to overcome ath / gmsh limitations here?
Oval waveguide, working fine
Slot with wide, horizontal cheeks. Is getting meshed when Morph.TargetShape=2 is set active (; removed)
For oval waveguide, shape must be set with guiding curve however, is meshed.
But when Slot feature is enabled ..
Oval waveguide, working fine
Code:
Throat.Angle = 10.08
Throat.Diameter = 25.4
Throat.Profile = 1
Coverage.Angle = 53 -14*sin(p)^2
Length = 56.5
Term.s = 3 -0.04*sin(p)^2
Term.n = 3.5 +0.7*sin(p)^2
OS.k = 0.28 +0.4*sin(p)^2
Term.q = 0.87674Slot with wide, horizontal cheeks. Is getting meshed when Morph.TargetShape=2 is set active (; removed)
Code:
Slot.Length = 25*cos(p)^10
;Morph.TargetShape = 2For oval waveguide, shape must be set with guiding curve however, is meshed.
Code:
GCurve.Type = 2
GCurve.SF = 1.6,1.5,4,4.3,5.6,2.2
GCurve.Width = 256
GCurve.Dist = 1But when Slot feature is enabled ..