I have tried a few a few years back. None were satisfactory at the higher frequencies. This particular package I had asked for more information from a rep in this region. Never got a response regarding price or capabilities.andy19191 said:
andy19191 said:
The problem is that FEA is very inefficient at the radiation problem since one of the boundary conditions is at infinity. There isn't an ideal solution to the HF acoustic radiation problem even today.
gedlee said:
Remarkable. 90 years after the Webster equation, and with personal computers more powerful than the IBM and Control Data mainframes of the Seventies, horn and waveguide design still has many unanswered questions to be resolved. I am very appreciative that Geddes and others are "working the problem" and coming up with new information.
Considering the primitive state of the art in the Seventies, we should probably cut JBL and Altec some slack on their first-generation Constant-Directivity horns.
Lynn Olson said:
Absolutely, people like Don Keele were no slouchs. Don and others did some great work. I know that Don is very aware of the same problems that I adress. We used to talk often, but I haven't talked with Don since he left Harman.
Lynn Olson said:
Ummm. As someone that was working in the "high tech" industry in the 70s I can assure you that BEM/FEM was essentially a solved problem for linear structural and acoustical problems by this time. There was a fair amount of development activity but the vast majority of research activity had shifted to nonlinear problems. Linear FEM was essentially done in the 1960s as expressed in codes like NASTRAN.
I do not know the environment in which Earl performed his PhD all those years ago but it would seem to have been outside the high tech industry/research lab community which seemed to cooperate more in those days than it does now despite all the Gnu, GPL, open source stuff. Perhaps it was Microsoft earning billions of dollars for rubbish software that tipped off the management but things tightened up a lot in the 80s.
andy19191 said:
It was at the US Navy's lead underwater lab in sound radition at Penn State (The Garfield Underwater Tunnel). While the problems were mosty "solved" at that point, the code was not that far along and acoustic radiation was NOT being done with FEA or BEM (some codes like CHEIF were available, but those were quite limited). FEA as applied to room acoustics was exceedingly rare. Only one Phillips Labs study that I knew of.
Below some "hard facts" correcting for all the wrong theories about BEM floating around the web.
(my initial take on the matter included – nothing beats first hand experience 😉 )
The simulations shown, I calculated with the excellent BEM software AxiDriver (recommended to me some earlier).
The corresponding visualisation software is called Vacs.
AxiDriver and Vacs are an excellent package for the DIYer who deals with horns - almost intuitively to handle at a really fair price IMO (non commercial Vacs Euro 41.- / AxiDriver Euro 69.- excl. VAT ).
For this bargain price I bought it right away – to me its kind of "must have" if you are serious in horns. 😉
Even better- to check out is easy as demo version is not limited despite storing.
🙂
http://www.randteam.de/Vacs/Index.html
http://www.randteam.de/AxiDriver/Index.html
This package is incredible in its functionality – I haven't even scratched at the surface so far.
Asking the developers about the limitations what's *exactly* being calculated by AxiDriver I got the following answer:
"AxiDriver ist ein finite Elemente Simulator für die Akustik und ein Lumped Element Simulator für den elektro-mechanischen Teil. Auf der akustischen Seite ist die Grösse der finiten Elemente die einzige Approximation, ansonsten ist die Berechnung vollständig und beinhaltet alle Reflektionen und Interferenzen und sogar die Nahfeldeffekte."
"AxiDriver is a finite element simulator for acoustics and a lumped element simulator for the electro acoustic part.
For the acoustic part the width of the finite elements are the only approximation, despite that, the calculation is complete, including all reflections and interference and even near field effects."
Meaning the restrictions I related to BEM don't hold here – AxiDrive provides a *complete* solution for the wave equation.
You could calculate even detailed compression drivers attached to a horn - if you'd like and have some time to spend.
What I was most interested at first hand was to do the sanity check with what W.M. Hall has measured in 1932 (referenced in Bjorns paper).
Excellent correlation – no? (despite its not at all about "wave fronts")
Simulation time needed – guess what – less than a second (on my several years old laptop) !!
My second interest was, to what extend the irregularities at the top end of horns (not captured by soongsc's BEM simus) can be simulated / visualised.
I set up for a 100 points / octave high resolution simulation of the LeCleach contour – as I have had the co-ordinates available already and measurements I've shown before.
Simulation time I don't know – was sleeping – got the results in the morning.
We clearly see in the directivity plot that there is an irregularity captured by my BEM simu almost exactly at the point where it shows up in my measurements too.
Same is seen in the overlay of FR traces for 0 10 20 30 40 deg at 1m
The driver was modelled as a flat piston for simplicity – maybe if I adjust for some "surrounding" loss of the AMT diaphragm, I would even be able to more precisely sim / visualise the angle dependency shown up in my measurements.
Anyway the development of that dip over the angles of 0 10 20 30 40 deg is quite clearly demonstrated.
Bottom line – you *can* simulate "eternity / infinity" – it just takes some time...
😀
Michael
PS:
Earl is facing hard times to come – a loooot of pretty pictures popping up everywhere – pure anarchism of colours mixed by some crazy monkeys spread all over the world.
🙂
(my initial take on the matter included – nothing beats first hand experience 😉 )
The simulations shown, I calculated with the excellent BEM software AxiDriver (recommended to me some earlier).
The corresponding visualisation software is called Vacs.
AxiDriver and Vacs are an excellent package for the DIYer who deals with horns - almost intuitively to handle at a really fair price IMO (non commercial Vacs Euro 41.- / AxiDriver Euro 69.- excl. VAT ).
For this bargain price I bought it right away – to me its kind of "must have" if you are serious in horns. 😉
Even better- to check out is easy as demo version is not limited despite storing.
🙂
http://www.randteam.de/Vacs/Index.html
http://www.randteam.de/AxiDriver/Index.html
This package is incredible in its functionality – I haven't even scratched at the surface so far.
Asking the developers about the limitations what's *exactly* being calculated by AxiDriver I got the following answer:
"AxiDriver ist ein finite Elemente Simulator für die Akustik und ein Lumped Element Simulator für den elektro-mechanischen Teil. Auf der akustischen Seite ist die Grösse der finiten Elemente die einzige Approximation, ansonsten ist die Berechnung vollständig und beinhaltet alle Reflektionen und Interferenzen und sogar die Nahfeldeffekte."
"AxiDriver is a finite element simulator for acoustics and a lumped element simulator for the electro acoustic part.
For the acoustic part the width of the finite elements are the only approximation, despite that, the calculation is complete, including all reflections and interference and even near field effects."
Meaning the restrictions I related to BEM don't hold here – AxiDrive provides a *complete* solution for the wave equation.
You could calculate even detailed compression drivers attached to a horn - if you'd like and have some time to spend.
What I was most interested at first hand was to do the sanity check with what W.M. Hall has measured in 1932 (referenced in Bjorns paper).
Excellent correlation – no? (despite its not at all about "wave fronts")
Simulation time needed – guess what – less than a second (on my several years old laptop) !!
My second interest was, to what extend the irregularities at the top end of horns (not captured by soongsc's BEM simus) can be simulated / visualised.
I set up for a 100 points / octave high resolution simulation of the LeCleach contour – as I have had the co-ordinates available already and measurements I've shown before.
Simulation time I don't know – was sleeping – got the results in the morning.
We clearly see in the directivity plot that there is an irregularity captured by my BEM simu almost exactly at the point where it shows up in my measurements too.
Same is seen in the overlay of FR traces for 0 10 20 30 40 deg at 1m
The driver was modelled as a flat piston for simplicity – maybe if I adjust for some "surrounding" loss of the AMT diaphragm, I would even be able to more precisely sim / visualise the angle dependency shown up in my measurements.
Anyway the development of that dip over the angles of 0 10 20 30 40 deg is quite clearly demonstrated.
Bottom line – you *can* simulate "eternity / infinity" – it just takes some time...
😀
Michael
PS:
Earl is facing hard times to come – a loooot of pretty pictures popping up everywhere – pure anarchism of colours mixed by some crazy monkeys spread all over the world.
🙂
I've recommended this one before, it's good enough to get a decent prototype, but is not enough for fine tuning. Driver diaphragm being a rigid body and limited shape options is are the most limiting factors I can think of.
Why do you have the mic location to one side instead of on-axis?
Why do you have the mic location to one side instead of on-axis?
Soongsc - mic position is due to the polars I wanted to show as FR overlay.
Are there better ways to do this?
Any advice welcome as I don't have the time to skim though all options possibly available.
Its the same package you useed ? - no wonder that your contour works like heaven fo me!!! Why didn't your sims capture the top end irregularities we were talking about - any clue?
The available driver optins of dome, cone, piston and various combinations and free scalability are not enough for your requirements? In addition there seems to be the possibility to add more or less complicated phase plug contours.
And all of this where price is no longer of any severe concern - I really was smashed by this offer.
Michael
Are there better ways to do this?
Any advice welcome as I don't have the time to skim though all options possibly available.
Its the same package you useed ? - no wonder that your contour works like heaven fo me!!! Why didn't your sims capture the top end irregularities we were talking about - any clue?
The available driver optins of dome, cone, piston and various combinations and free scalability are not enough for your requirements? In addition there seems to be the possibility to add more or less complicated phase plug contours.
And all of this where price is no longer of any severe concern - I really was smashed by this offer.
Michael
AxiDriver provides good capability for the price. I think there is no doubt about it. However, due to the limitations I mentioned, there are lots of options that cannot be explored.
I edited the graphics to get all the curves on to one chart.
I asked the author about the off-axis curves whether they were constant distance from the baffle or from a point. Never received a response.
I edited the graphics to get all the curves on to one chart.
I asked the author about the off-axis curves whether they were constant distance from the baffle or from a point. Never received a response.
Its pretty easy to do overlays for different off axis angels - though you have to do it in Cartesian system (no polar entry for this one as far as I can see).
In the tab "Vacs" you edit "r-coordinate" and "z-coordinate" for a specific mic position and send results automatic or manually to the Vacs visualisation tool by prompting.
You have to duplicate the trace in Vacs first (by double clicking) to not get overwritten the original one.
IIRC I have read that all coordinates have its origin at the drivers diaphragm origin.
Michael
In the tab "Vacs" you edit "r-coordinate" and "z-coordinate" for a specific mic position and send results automatic or manually to the Vacs visualisation tool by prompting.
You have to duplicate the trace in Vacs first (by double clicking) to not get overwritten the original one.
IIRC I have read that all coordinates have its origin at the drivers diaphragm origin.
Michael
I'm used to looking at data about a single point source. This is the way most people measure.
open baffle edge radius
I'm making new small open baffles for full range 12 inch speakers that will set on top of sub woofers.
It boils down to using a router for a 3/4 inch edge radius or use a quarter section of a PVC tube to get a much larger edge radius.
Is there any benefit to larger radii or is there a practical limit?
Frank
I'm making new small open baffles for full range 12 inch speakers that will set on top of sub woofers.
It boils down to using a router for a 3/4 inch edge radius or use a quarter section of a PVC tube to get a much larger edge radius.
Is there any benefit to larger radii or is there a practical limit?
Frank
dobias, still looking for the calculations, but as per my memory there was an appreciable benefit for frq-resp even out to 8" diameter quarter rounds (where I believe the estimates stopped).
open baffle edge radius
Ang,
That info will be great.
I wonder if the back of the baffle would also benefit with an edge treatment. In other words, maybe the (8 inch) PVC tube should wrap around to the back surface.
Two years ago, after I made the present baffles, I added 4 inch diameter cardboard carpeting rolls to the sides that stick out in front as well as in the back. It made a big improvement to the sound stage but I never knew if it could be improved upon or if it was overkill.
Thanks for looking up those calculations for me.
Frank
Ang,
That info will be great.
I wonder if the back of the baffle would also benefit with an edge treatment. In other words, maybe the (8 inch) PVC tube should wrap around to the back surface.
Two years ago, after I made the present baffles, I added 4 inch diameter cardboard carpeting rolls to the sides that stick out in front as well as in the back. It made a big improvement to the sound stage but I never knew if it could be improved upon or if it was overkill.
Thanks for looking up those calculations for me.
Frank
A feb 2003 audioxpress article apparently has the calculation I was thinking of... will email a scan if I can find the (physical) issue.
Your question of whether to give the baffle a smooth profile on the back would seem to depend on if you're using a dipole tweeter.
Your question of whether to give the baffle a smooth profile on the back would seem to depend on if you're using a dipole tweeter.
open baffle edge radius
Ang,
Dipole tweeter ? As in one tweeter facing the front & another (out of phase) tweeter facing the rear ?
That's a possibility, but I planned on facing all four 3 inch tweeters forward, open to the rear. That many are needed to get the SPL up to the level of the full range. If I had my 'druthers' I'd prefer a single tweeter that was up to the task, but I have the cone tweeters & they work OK from 7.5K. The 95dB full range operates without any XO . I don't want to pad the volume down to use a single tweeter. The 93dB sub helps below 120Hz.
Frank
Ang,
Dipole tweeter ? As in one tweeter facing the front & another (out of phase) tweeter facing the rear ?
That's a possibility, but I planned on facing all four 3 inch tweeters forward, open to the rear. That many are needed to get the SPL up to the level of the full range. If I had my 'druthers' I'd prefer a single tweeter that was up to the task, but I have the cone tweeters & they work OK from 7.5K. The 95dB full range operates without any XO . I don't want to pad the volume down to use a single tweeter. The 93dB sub helps below 120Hz.
Frank
Hi Frank - my point was more that without rear radiation, there's nothing flowing out from the rear of the baffle to the front - and you were considering how to treat the rear of the baffle. Your mid may be emitting frequencies who's wavelengths are on the order of the baffle width, so in that case I guess your proposed diffraction optomizations would still be useful.
PS
There are dipole tweeters (neo3) or ones that can be converted (some compression drivers) or you could have a second in or out of phase (advantages to each).
PS
There are dipole tweeters (neo3) or ones that can be converted (some compression drivers) or you could have a second in or out of phase (advantages to each).
Ang said:
Not easy to have 95dB eff, though. The beyma tpl 150 comes to my mind. It is 99dB and can be padded without being hurt. It costs like 4 or 5 of the cheaper ones 😉