Hello
Currently I am trying to design and build prototypes similar to danley BC218, divided in half for easier transport and multiple possible deployment configurations. I will be using 18SW115 8ohm version as i already own them and it seems Danley also uses them a lot in his designs. My goal is not to make a identical copy of his design but to make a sub with similar preformance and same design benefits. Aiming for same dimensions and about same frequency response (-3db@ 28hz).
I've been following @weltersys folding prediction and made a 3d model, but 18SW115 sadly doesnt fit and i am not exactly sure how to model in hornersp with decent accuracy.
I am relatively skilled in Solidworks, Catia and cabinet building (designed and built my own version of TH118 and 4 way synergy horn), but lack experience for simulating such advanced horn sub.
Any help would be appreciated, as i said looking forward do design 3d model first from usable simulations, then build 2 or 4 prototypes - I have a huge free area near my house and I will keep this thread updated with build progress and measurements.
Looking for support in simulating this folding or exploring possible different horn foldings
Attaching picture of a very rough model
Currently I am trying to design and build prototypes similar to danley BC218, divided in half for easier transport and multiple possible deployment configurations. I will be using 18SW115 8ohm version as i already own them and it seems Danley also uses them a lot in his designs. My goal is not to make a identical copy of his design but to make a sub with similar preformance and same design benefits. Aiming for same dimensions and about same frequency response (-3db@ 28hz).
I've been following @weltersys folding prediction and made a 3d model, but 18SW115 sadly doesnt fit and i am not exactly sure how to model in hornersp with decent accuracy.
I am relatively skilled in Solidworks, Catia and cabinet building (designed and built my own version of TH118 and 4 way synergy horn), but lack experience for simulating such advanced horn sub.
Any help would be appreciated, as i said looking forward do design 3d model first from usable simulations, then build 2 or 4 prototypes - I have a huge free area near my house and I will keep this thread updated with build progress and measurements.
Looking for support in simulating this folding or exploring possible different horn foldings
Attaching picture of a very rough model
Last edited:
Looks like the 18SW115 "fits" in your rough model 😉 .
The "half moon" exit shape won't model in Hornresp, you would need to use Akaback to get closer.
I have no experience using Akaback.
If you are just looking for a start in Hornresp, these inputs would get you going, they would all need adjustment in length and area to conform with whatever your horn's cross sectional areas and lengths are:
Two (2P) 18SW115-4 were used in those models, you would have to substitute a single 18SW115-8 driver parameters.
Art
Thank you for detailed answer @weltersys, I am only slightly experienced with hornresp and have no experience with Akaback, As far as my understanding goes the ''half moon shape'' extends low frequency by few hz, so for what reasonable low corner should I aim in my hornresp model? in advance I apologise for potential use of wrong expressions or silly questions. Do you still think that this folding of horn is suitable for this cabinet or there might be a better option? I have seen a video on Youtube posted by Ivan and it seems they use this driver orientation (we can see driver through plexiglass). Attaching a picture of driver mounted in 3d model, you can see there is a few milimeters missing, I followed your prediction which uses lower driver height.
Will try to modify overall shape a bit to get it to fit comfortably
What do you think about overall folding of a horn? I will add a reflector at the top left corner but i suppose adding other reflectors in early sections of horn will reduce horn volume and make response worse? Also extensive bracnig will be added.
Will try to modify overall shape a bit to get it to fit comfortably
What do you think about overall folding of a horn? I will add a reflector at the top left corner but i suppose adding other reflectors in early sections of horn will reduce horn volume and make response worse? Also extensive bracnig will be added.
Last edited:
Yes, since I have been following the Xoc1 thread for TH18 build and built a few of them I already know about benefits of not adding corner refectors, my plan was only to add a reflector on top left and good bracing. Planning to use this refector also as place for handles
Currently I am exploring different horn folding approaches to eliminate wasted space next to horn mouth, if any of you has any suggestions I will be happy to 3d model and sim them.
Currently I am exploring different horn folding approaches to eliminate wasted space next to horn mouth, if any of you has any suggestions I will be happy to 3d model and sim them.
"BC" =boundary coupled.
The increased boundary area increases forward gain by as much as 4dB around 55Hz, about half that below the low corner.
Shoot for an impedance response similar to the BC218, and an F3 of 26Hz.
Obviously some slight changes are required to fit the driver.
Looks OK to me, more twists in the throat end could make the path length longer, but it's close to the correct 1/4 wave length already, assuming 60" x 30"x 30" outer dimensions.
Those dimensions are made for (USA) truck pack and standard 5'x5' (60" x60") Baltic Birch.
Danley seldom uses "reflectors", other than for wheels or handle placement as you plan.
Art
Thank you for response, will follow your impedance response advice
Here in europe we have plywood mostly available in 3000x1500mm sheets (pretty close to 120x60 inches), so outer dimensions I plan to use are 1500x750x750mm, which closely resemble original dimensions (one half).
I think i will stick with current folding, only adapt it to fit driver and then adjust it to simulations.
Will keep you updated and thanks for detailed advice.
Here in europe we have plywood mostly available in 3000x1500mm sheets (pretty close to 120x60 inches), so outer dimensions I plan to use are 1500x750x750mm, which closely resemble original dimensions (one half).
I think i will stick with current folding, only adapt it to fit driver and then adjust it to simulations.
Will keep you updated and thanks for detailed advice.
Here are a couple of good threads on the BC concept, worth a read to align assumptions:
https://www.diyaudio.com/community/...t-reverse-engineering-a-danley-design.262490/
https://www.diyaudio.com/community/threads/danley-bc-subs-revisited.408245/
https://www.diyaudio.com/community/...t-reverse-engineering-a-danley-design.262490/
https://www.diyaudio.com/community/threads/danley-bc-subs-revisited.408245/
Yes I have seen these 2 threads. there seems to be quite a bit of useful info, but both threads kinda go offtopic fast. There isn't really a lot of hornresp sims or actual models with dimensions. Also some predictions seem to get way too low tuning in my opinion. Art's prediction for horn folding seems to be the best and closely resemble original. I started this thread with intention to exploring BC subwoofer concept and to build actual prototype, then post real word measurements. Making 3d model and building good quality cabinet is easy part for me, but to get horn dimensions right was the main issue, so I seek help and advice from experienced horn designers.
Very cool project. Maybe I can help you with some more accurate simulations involving the shape of the mouth of the horn etc, if you are close to your first prototype design. The acoustic impedance of the horn we can get from FEA, and is pretty straightforward once you have the 3D model (and FEA software provided by your employer). FYI: I'm no audio or acoustic engineer, but I think this is a good example where I can try out some things and maybe your first prototype can be a first-time-right, and we have enough acoustic experts available here to warn us if something is too good to be true ;-).
Thats exactly what I am looking for. So do you think this would be possible to do FEA for horn speakers directly in solidworks? I normally use it only for 3d modeling parts and for making .stl files for 3d print, but never actually touched FEA and know pretty much nothing about it . Since this is a very large sub I definitely hope to get it right in first try. Will try to get 3d model done (with approximate simulations from hornresp), then we can sim it with FEA and see if we are close. Also noticed we can get a lot of info from official spec sheet as @weltersys mentioned, it should be very helpful for designing process.
Unfortunately it takes a bit more than simply do some FEA in SolidWorks (a lot of years ago I thought it was that easy as well ;-)). Software like WinISD and Hornresp combine the electrical, mechanical and acoustical part to calculate sound pressure level as function of generator (amplifier) voltage and frequency. The electrical and mechanical part are determined by the driver (TS parameters) and the acoustic part consists out of acoustic impedances (relation between volume velocity and sound pressure). By using FEA it is possible to calculate the acoustic impedance (and directivity) for any shape/dimension of a horn (and include for example a wall or groundplane for a subwoofer). What I do is implement the acoustic impedance simulated by ANSYS (you must be able to work with 'fluid' elements) in Matlab, and it calculates the overall response.
What it can be is a confirmation that the simplified horn model in Hornresp (don't get me wrong, Hornresp is great!) is correct, or maybe we find out that the lower cutoff point is too low/high due to the shape of the mouth (for example).
What it can be is a confirmation that the simplified horn model in Hornresp (don't get me wrong, Hornresp is great!) is correct, or maybe we find out that the lower cutoff point is too low/high due to the shape of the mouth (for example).
Last edited:
Yes I also agree with you, I think shape of the mouth definitely changes things a bit. In attached picture we can see how different configuration change frequency response. I believe that in ''flat'' and ''stacked'' configuration, it behaves as a standard front loaded horn. So i think we should take that frequency response for our goal and then just mimic moon shape from pictures. Is my thinking correct?
About simulations, what would be logical next step (or complete process of simulating design from start to finish?) If I understand correctly, first i should make a 3d model with hornresp sims and then verify it with methods you have mentioned above?
About simulations, what would be logical next step (or complete process of simulating design from start to finish?) If I understand correctly, first i should make a 3d model with hornresp sims and then verify it with methods you have mentioned above?
As a start I would begin iterating between Hornresp and 3D, trying to find the best compromise between response and what you can fold into the enclosure 🙂. Weltersys has shown some nice examples. If you are going for similar response as the 'flat' response shown above, beware that you compare 2 of your 1x18" boxes with 1 Danley sub.
Now I see that the datasheet also contains the response for the 'flat' position, Hornresp will probably do just fine to simulate this, and indeed apply the 'half moon' as close to the original. It would be cool however to see if we can simulate the same response differences for different stacking configurations by using a finite element method. So you can always send me the model (*.stp) when you're done iterating.
Now I see that the datasheet also contains the response for the 'flat' position, Hornresp will probably do just fine to simulate this, and indeed apply the 'half moon' as close to the original. It would be cool however to see if we can simulate the same response differences for different stacking configurations by using a finite element method. So you can always send me the model (*.stp) when you're done iterating.
Will try to get a decent result with hornresp, then send you .stp file for further simulations. I am also curious how cabinet behaves in diferent configurations and also with half moon placed towards corner and similar options.
I think this will actually be the most difficult part of design process. I wonder how exactly do you begin with model process? Do you make a rough 3d model first, then enter values in hornresp and then fine tune model or?
Yes.
Since you are building the cabinet in two parts, shoot for the frequency and phase response shape of two cabinets to be similar to that of the blue trace (1XBC218-Flat).
I think the Hornresp sensitivity will still come in lower than DSL's, but I've given up chasing that windmill😉
Also, as Tom Danley wrote:
The strength of a modeling program is only as great as it’s ability to predict what you measure when you build from the model. I use Akabak which is a powerful, somewhat cumbersome program but it has real limitations and in bass horn land, should be viewed as a “ball park prediction”.
Often to get closer to what one measures, one has to add some unseen elements to the model. For example, a wood panel flexing at the throat can “look like” an air volume you didn’t account for and while a computer can predict high Q or sharp features in a bass horn, they hardly ever exist in real life as there are losses related to the layout which limit the Q and are not included in the simple model.
Best,
Tom
That behavior is going to depend on what corner you put it in...
To get you started in Hornresp, my workflow would be:
1. Enter driver parameters
2. Use 1 exponential horn section
3. You know your global box dimensions, so you can enter the horn's mouth area at 'S2'.
4. Your 3D design gives you an idea what horn length can be obtained, enter at 'L12', or 'exp'.
5. Get your available horn volume (volumetric) from your 3D design (L*W*D of your total cabinet)
Now you can fiddle around with:
And check if your horn volume calculated by Hornresp ('Window' -> 'Schematic Diagram') does fit in your box, and if the response shown ('Window' -> 'Acoustical Power') does meet your requirements.
If all is well you need to discretize this 1 exponential horn section to 3 or 4 parabolic sections (because that is wat you are actually building), and see if you can still obtain your desired response.
I see Weltersys is using the Offset Driver (OD) model in Hornresp, but I think as a start you could stick to the Single Driver (ND) model.
I got too excited and tried the above already 😉, and it seems promising:
1. Enter driver parameters
2. Use 1 exponential horn section
3. You know your global box dimensions, so you can enter the horn's mouth area at 'S2'.
4. Your 3D design gives you an idea what horn length can be obtained, enter at 'L12', or 'exp'.
5. Get your available horn volume (volumetric) from your 3D design (L*W*D of your total cabinet)
Now you can fiddle around with:
- The horn's throat area 'S1'
- The rear chamber volume 'Vrc'
- Driver's front chamber volume 'Vtc'
And check if your horn volume calculated by Hornresp ('Window' -> 'Schematic Diagram') does fit in your box, and if the response shown ('Window' -> 'Acoustical Power') does meet your requirements.
If all is well you need to discretize this 1 exponential horn section to 3 or 4 parabolic sections (because that is wat you are actually building), and see if you can still obtain your desired response.
I see Weltersys is using the Offset Driver (OD) model in Hornresp, but I think as a start you could stick to the Single Driver (ND) model.
I got too excited and tried the above already 😉, and it seems promising:
So after some modeling I came up with this. Horn length is 362cm. @Vermond I cannot send you .stp file here, can you give me mail or something in PM? so you can simulate it further.
