Would someone adept w. Akabak3 sim the "Dutch K12" for me ?

- with many thanks for CAD work by TB46 (2D) and Pelanj (3D models), I present for examination 3D files for the popular Karlson type cabinet known as “The Dutch K12”, which was introduced In a June 1963 article titled “Karlson Luidsprekerkasten voor Kleinbehuisden”

https://www.diyaudio.com/forums/full-range/313227-dutch-karlson-12-a-2.html#post6473774

This cabinet’s front chamber resembles that of Karlson’s 1st K12 (introduced fall 1954) and rear chamber that of the 2nd K12 (fall 1956 plus featured in July 1958 Popular Mechanics.

It sports 3 horizontal slots as a distributed port.

I and other Karlson fans would be very grateful for a thorough Akabak3 examination of this cabinet. Items of interest would be 2pi frequency response on and off central axis, frequency response at some distance back and “up” from the cabinet’s central axis, input impedance, cone excursion, and group delay.

Drivers of interest = Fane’s twin whizzer 12-250TC and Eminence’s 12LTA.

Details of just the Eminence 12LTA would be plenty - I would like to see what 12-250TC does for peaking on the low end.

Also, it could be interesting to see what things look like with the port shut off.




Def_Driver ‘12-250TC’ | Fane 12” full range driver with extended highs, 3.5mm xmax
Sd=530.9cm2
Fs=50Hz
Mms=52g
Qms=7.6
Qes=0.72
Re=7.2
Le=0.156mH
Vas=78.06L


Def_Driver ‘12LTA’ | Eminence 12” whizzer cone speaker, xmax 3.2mm
Sd=532.4
Fs=45Hz
Mms=36g
Qms=6.44
Qes=0.55
Re=7.73
Le=0.83mH
Vas=135.3L
 

Attachments

  • DutchK12.stl.zip
    10.8 KB · Views: 79
  • DutchK12.step.zip
    12.4 KB · Views: 66
As I mentioned in a thread below an acoustic BEM code will miss some relevant physics if there is a Helmholtz-type resonance involved. I intend to perform a set of simulations on the subject in the future when I have completed developing some software but that won't be soon. I would be happy to use your problem as an example so long as it doesn't involve too much of my time. Your model doesn't appear to have a driver which will be required. If you want to include details like whizzer cones the 3D driver model will be complex and I'm not sure Fane or Eminence would be prepared to supply theirs for a bit of messing about on the web.
 
I also doubt the availability of that type model. Perhaps you may have sometime a suitable 12 inch woofer model with relatively low mass , low L which would suffice to a few KHz.

I suspect you may not have wholly appreciated the issue. You want to simulate a particular type of loudspeaker for purposes that are currently unclear. The purpose will influence the level of modelling assumptions that can be introduced to reduce the size and complexity of the model. Determining what this is will require significant time from someone who knows what they are doing talking to you, collecting information, checking the information is reliable against measurements and initial simulations of sub-models.

For example, in the neighbouring thread on python code I refer to a 0D sub-model of a driver developed to provide the boundary conditions for the vibration analysis of a small subwoofer cabinet. It took a good day to collect the equations, write the code and check against measurements to establish that the level of modelling was sufficiently for the purposes wanted.

[How do I insert an image that is not on the web?].

This aspect in your simulation will be more complicated because you have a light resonant cone with a possibly asymmetric acoustic loading. It may be possible to derive boundary conditions or it may be required to include the complex resonant motion of the large light cone either as a separate simulation coupled one way, iterated one way coupling or included with the main simulation of the motion of the air. It will take more than a day to sort out.

As always with these types of simulations damping is the tricky thing to get right (enough). We will need sufficiently accurate properties for the damping material in the cavities with the correct form of empirical coefficients to match the damping model used. Driver manufacturers tend to be less than forthcoming about the composition and properties of their proprietary "paper" cones. Whatever is adopted it is likely to take several more days to work into acceptable shape.

I will almost certainly have missed some other things that will require time.

After collecting all the required information and creating the relevant submodels then we would be in a position to simulate the sound in the air (and damping material) using a linear acoustic BEM approach and a more complete CAA approach. Setting up and running these simulations was what I was offering to do in the future when the software currently being developed is in a good enough state. I was not volunteering to develop and prepare the model but if you wait long enough perhaps more of it will get done in other projects like the subwoofer cabinet project mentioned above.
 
purpose to see how good a correlation could be done for what I think would normally be a difficult case to examine.

I would have to have that particular cabinet made for real measurements. FWIW I've measured some Karlson enclosures in the past but not in great detail.

I can fudge some things in hornresp. XRK971 uses Akabak 2.1 for Karlson
cabinets.

There were a couple of limited sims at Diyaudio of two Karlson type using Akabak3 - the original 1951 Karlson enclosure known as "K15", and an ~1980 115BK cabinet by Acoustic Control.

There are no commercial interests involved, nor do I have money to pay anyone for their valuable time.

Karlson's 2nd and final "Acoustic Transducers" patent, 3540544, used something like that "Dutch K12" with port panel parallel to the wing's surface as an example of prior art and what "not to do" for smoothest response.
 
Last edited:
Thanks for the response which has confirmed this is a hobby interest for you. I am semi-retired and the software and associated consulting is the work part of my semi-retirement. Speaker design is a hobby interest which, perhaps unwisely, I am allowing to leak over into the work side to provide some realistic examples.
 
Is there a "Kalson" mode's code left over from various discussions which could be dropped into akabak3 ? I doubt if I could figure out even how to run a sim. What kind of pc power does akabaklike? Is it more dependant upon GPU than cpu? (FWIW have 28 slow cores on one recent build)

What I'd really like would be to investigate wheter there's certain front chamber shapes and associated aperture which play smoother than others and especially for small
K-couppler using 5- 6-8 inch pro driver as would like that size for midrange duty