CNC files?
JMMLC,
I joined a co-op where we have access to the newest CNC fabrication equipment for personal use. I'd love to try to make a set of wood horns. I contacted Nicolas, and he did not want to release the CNC files. I'm not sure why.
Could you help me generate the data points for an ecliptic horn centered on 220Hz and 600Hz for CNC files? In the spirit of the forum, I would be happy to post any files I make for others to use.
Thanks.
JMMLC,
I joined a co-op where we have access to the newest CNC fabrication equipment for personal use. I'd love to try to make a set of wood horns. I contacted Nicolas, and he did not want to release the CNC files. I'm not sure why.
Could you help me generate the data points for an ecliptic horn centered on 220Hz and 600Hz for CNC files? In the spirit of the forum, I would be happy to post any files I make for others to use.
Thanks.
if your goal is a Tractrix round horn, you may want use Tractrix Calculator which gives you good CAD output you can use in Autocad to order your CNC 😉 I can uoload it for you
As long as I'm cranking up the CNC router, I'd like to try Itawa-JMMLC horns.
Points for Itawa
Can you make the calculated points available to the forum? I would like to CNC it as well.
Can you make the calculated points available to the forum? I would like to CNC it as well.
"other hypothesis"
JMLC,
Would you elaborate please; particularly in regards to the mapping of horn boundaries from the area of an assumed wave front surface.
Regards,
WHG
JMLC,
Would you elaborate please; particularly in regards to the mapping of horn boundaries from the area of an assumed wave front surface.
Regards,
WHG
update from Norway:
(without spending too much effort on the technical aspects of our journey)
a friend and I have spent the last year or so testing different alternatives to replace AG Trio satellites
we also replaced original Trio mid and hf drivers with TAD 4001 and BMS4538 with great success!
actually apart from the design, I think the kugelwellen horns of the Trios are their best feature and not the rather mediocre drivers
have mainly tested with JBL 2446 truextent and TAD 4001 since they´re the drivers we´ve homed in on from listening impressions
we´ve tested TAD TH4001, Autotech JMLC350, 200T and Iwata 300, a 40KG CNC´d 300Hz Iwata plywood horn and the AG Trio original midrangehorn.
my initial objective was to try and replace the Trio midbasshorn with a combo of higher lowpass of my 4x15/channel hi eff Onken style BR towers and lower hipassed driver in JMLC200T or equivalent to obtain a "seamless" 2 -2,5 way system.
we soon found out that all tested drivers even though measuring fine down to 300-350 in the big horns, sounded far better in the entire operating range when hipassed higher at about 600Hz
same in the other end of the range: both drivers went high and nice up to not really needing a supertweeter, TAD leading clearly, but both drivers benefited on being lowpassed at around 7-8K to a matching supertweeter
at the same time I realised that I could never live without hornloaded midbass
status today is:
Trio sold
4 straight conical midbasshorns ordered: 2 stacked per side with JBL2220B for 80-600Hz duty approx 110db to match TAD driver
(they´ll be double walled horns for sand-damping)
probably settling today on JMLC version for TAD with approx 600hipass
nothing to gain with 200T version anyway so I´m going for a size in between 200 and 350
to sum up
best sounding horns was JMLC and Trio kugelwellen
with 200T and Trio we got away with 15µF cap as only hipass component, which I consider a huge advantage
a couple of weeks and I´m up&running
best
Leif
Norway
(without spending too much effort on the technical aspects of our journey)
a friend and I have spent the last year or so testing different alternatives to replace AG Trio satellites
we also replaced original Trio mid and hf drivers with TAD 4001 and BMS4538 with great success!
actually apart from the design, I think the kugelwellen horns of the Trios are their best feature and not the rather mediocre drivers
have mainly tested with JBL 2446 truextent and TAD 4001 since they´re the drivers we´ve homed in on from listening impressions
we´ve tested TAD TH4001, Autotech JMLC350, 200T and Iwata 300, a 40KG CNC´d 300Hz Iwata plywood horn and the AG Trio original midrangehorn.
my initial objective was to try and replace the Trio midbasshorn with a combo of higher lowpass of my 4x15/channel hi eff Onken style BR towers and lower hipassed driver in JMLC200T or equivalent to obtain a "seamless" 2 -2,5 way system.
we soon found out that all tested drivers even though measuring fine down to 300-350 in the big horns, sounded far better in the entire operating range when hipassed higher at about 600Hz
same in the other end of the range: both drivers went high and nice up to not really needing a supertweeter, TAD leading clearly, but both drivers benefited on being lowpassed at around 7-8K to a matching supertweeter
at the same time I realised that I could never live without hornloaded midbass
status today is:
Trio sold
4 straight conical midbasshorns ordered: 2 stacked per side with JBL2220B for 80-600Hz duty approx 110db to match TAD driver
(they´ll be double walled horns for sand-damping)
probably settling today on JMLC version for TAD with approx 600hipass
nothing to gain with 200T version anyway so I´m going for a size in between 200 and 350
to sum up
best sounding horns was JMLC and Trio kugelwellen
with 200T and Trio we got away with 15µF cap as only hipass component, which I consider a huge advantage
a couple of weeks and I´m up&running
best
Leif
Norway
Last edited by a moderator:
Bjorn
Also is the OS in a baffle, or free? Baffled I hope as free is just adding problems where none need be. An OS should always be in as large a baffle as you can afford - because thats what the physics assumes!
I suspect that your model is having trouble at the upper edge near 8 kHz because the results that you show is not what happens. Below that is they are fairly good however. Based on the loss of control below about 2 kHz, I'm guessing the waveguide at about 10" across, maybe a little less. (Pardon the inches cr_p! Bad habit!)
Also is the OS in a baffle, or free? Baffled I hope as free is just adding problems where none need be. An OS should always be in as large a baffle as you can afford - because thats what the physics assumes!
I suspect that your model is having trouble at the upper edge near 8 kHz because the results that you show is not what happens. Below that is they are fairly good however. Based on the loss of control below about 2 kHz, I'm guessing the waveguide at about 10" across, maybe a little less. (Pardon the inches cr_p! Bad habit!)
Hi Earl,
I've put up the dimensions of the waveguides, and some more details on the mountings. It was a while since I had made the screenshots, and I discovered I had done a few mistakes when writing about them. It turned out that the waveguide with the comp. driver throat was in a "box" or cylinder, with sharp edges. I've made some comments about that. This is probably the cause of the HF trouble, but it may also be that 6 elements per wavelength is to little for HF detail, especially when diffraction from sharp edges is involved. (I've also noticed that when coming close to 6 elements per WL, the throat impedance may start to droop at HF).
I also found another simulation of an almost identical waveguide, but mounted in a baffle. It has a finer mesh (6 elm/WL @ 15kHz). Put up as #009.
You're close on the diameter, it's about 11.5in before the radius starts.
If you want me to do simulations that are closer to what you actually build, I'm happy to do that, provided I get the dimensions (confidential, if needed). The dimensions I list are those I need to generate the contour.
There is also a series of simulations showing the effect of the roll-back in LeCleac'h horns.
Regards,
Bjørn
I've put up the dimensions of the waveguides, and some more details on the mountings. It was a while since I had made the screenshots, and I discovered I had done a few mistakes when writing about them. It turned out that the waveguide with the comp. driver throat was in a "box" or cylinder, with sharp edges. I've made some comments about that. This is probably the cause of the HF trouble, but it may also be that 6 elements per wavelength is to little for HF detail, especially when diffraction from sharp edges is involved. (I've also noticed that when coming close to 6 elements per WL, the throat impedance may start to droop at HF).
I also found another simulation of an almost identical waveguide, but mounted in a baffle. It has a finer mesh (6 elm/WL @ 15kHz). Put up as #009.
You're close on the diameter, it's about 11.5in before the radius starts.
If you want me to do simulations that are closer to what you actually build, I'm happy to do that, provided I get the dimensions (confidential, if needed). The dimensions I list are those I need to generate the contour.
There is also a series of simulations showing the effect of the roll-back in LeCleac'h horns.
Regards,
Bjørn
Last edited:
BTW Earl, do you monitor my activity on this forum? 🙂 As soon as I post the link to the simulations, you pop up, after having been absent in those two threads for a long time... Nothing wrong, it just didn't seem like a coincidence 🙂
-Bjørn
-Bjørn
Thanks again for posting the BEM simulations, Bjorn. It gets them "out of the closet" of the private correspondence we've had over the past couple of years and into the light of day. I really feel these are a major step forward in understanding what goes on inside horns.
Pure coincidence. I lurk ever now and then, but seldom post as the topic is just not interesting enough. There are "real" questions in audio and there are those that just get beat to death because someone doesn't like the answer - like THD or phase coherence. I will discuss "real" questions, but not the other kind.
I did not realize that these threads were active because I did not get any notifications. Once you don't post for awile you don't get anything anymore. (which should not happen to someone who financially supports the forum!)
Thanks Bjorn
Your #009 simulations are very accurate for a small waveguide like that. Bigger is much better of course.
As to running simulations on my waveguides, that seems unnecessary. I mean simulations are great when you don't have the actual device or you want to know about changes, but if you actually have the device then simulations are not really of much interest. Then it's the measurements that matter, and I have a gizzilion of those and proprietary techniques that can reconstruct the mouth velocities, etc. As I said in other posts I was able to tell B&C that their compression drivers had changed when no one else had detected this. You have to put all the tools in their proper place.
What I would love to have is some simulations of different ways to do an elliptical waveguide, but I am not sure that your BEM can do that. Can you do 3D?
I did some tests on some elliptical stuff and it was pretty disappointing. Of course what I tested was not how I'd do it, but I'm not willing to disclose publicly how I'd do it. I'd like to know how much one throws away to get elliptical. If a good design is even possible. My waveguides now work better than anything else that I have ever tested (for what I am looking for) so I'm certainly happy with that.
My grand mother used to say "Self satisfaction is the beginning of wisdom."
My grand father: "Yes, but just the beginning".
Bjorn
I'm troubled by the data in "009 OSWG+DE250 in baffle, throat impedance" and "006 OSWG throat impedance". They just do not seem to be consistant. The one in the baffle should be smoother than the one in free space, but they are the other way arround. Any comments.
I'm troubled by the data in "009 OSWG+DE250 in baffle, throat impedance" and "006 OSWG throat impedance". They just do not seem to be consistant. The one in the baffle should be smoother than the one in free space, but they are the other way arround. Any comments.
You're welcome. Also good to hear that the sims are accurate. I have also compared the sims of AH425 to measurements, I may put those measurements up later. They were also quite close.
I understand that you don't see any need to simulate your waveguides when you have the data. I asked in case you felt your designs were misrepresented. For the AH425, I have enough information to make pretty accurate simulation setups, and I try to treat all cases the same.
I can do 3D BERIM (infinite baffle), and with some work, 3D ordinary exterior BEM too, but I have limited abilities to generate the mesh.
Agreed. But the DE250 throat section is slightly longer in the 009 case (IIRC, we didn't have the exact length when I set up the sims), thats my guess for the difference. I could check that by rerunning the sims without the throat section.
-Bjørn
Hello,
My friend Nicoals davidenko introduced new Iwata-JMLC horns that I calulated for him in his audio system :
See his blog :
Pavillons JMLC/Iwata 220, ça sonne ... - Le blog de audiobricolos
The Iwata-JMLC horns are made of wood. A 3D computer numerically controlled machine (CNC) was usin to obtain their beautiful shape.
Best regards from Paris,
Jean-Michel Le Cléac'h
My friend Nicoals davidenko introduced new Iwata-JMLC horns that I calulated for him in his audio system :
See his blog :
Pavillons JMLC/Iwata 220, ça sonne ... - Le blog de audiobricolos
The Iwata-JMLC horns are made of wood. A 3D computer numerically controlled machine (CNC) was usin to obtain their beautiful shape.
Best regards from Paris,
Jean-Michel Le Cléac'h