Spreadsheet for Folded Horn Layouts...

David's update to Hornresp and some "advances" I've recently made in emulating conical and exponential expansion via parabolic segments in workbooks led me to have another look at the workbook I put together for the "Keystone' fold. It looks like the S2-S3 section can be best sim'd as a conical section, so L12=Par, L23=Con, L34=Par. In fact, the workbook now shows why that short internal panel at the top is slightly off 90 degrees to the internal panel near the rear - it improves the approximation of the L23 section to a conical expansion. I might address that in a later version of this workbook. Otherwise the "error" line in the expansion graph is basically flat, showing that the Hornresp sim data in the workbook can describe the S1-S4 section of the Keystone pretty closely.

Of course it's the L45 section that makes a lot of the difference in the response of the Keystone, and I have to find a better way to approximate that in the model. Time to have a read of that long thread again - key details I need are the size of the mouth and an idea of how far from the center of the driver the approximate center of the mouth is located.
 

Attachments

  • 2017-11-03 (1).png
    2017-11-03 (1).png
    122.6 KB · Views: 462
Of course it's the L45 section that makes a lot of the difference in the response of the Keystone, and I have to find a better way to approximate that in the model. Time to have a read of that long thread again - key details I need are the size of the mouth and an idea of how far from the center of the driver the approximate center of the mouth is located.
Another "key detail" is the area difference between the upper and lower portion of the approximate area center of the Keystone mouth. Moving the equivalent exit area up and down on the exit panel does not have the same effect as the Keystone shape imparts. Hornresp simulates a circular exit, and though a square approximates a circle, a Keystone shape does not.

Happy reading!
Cheers,
Art
 
How can something be conical if the construction is not conical. Does this par,con,par sim actually going to comparable to the a non conical build measurement?

Hornresp sims are limited to four sections (L12, L23, L34, L45) of which each section can be Parabolic, Conical, or Exponential (ok, Hornresp can do Hyperbolic too, but there are certain limitations applied), and for a horn consisting of various segments folded up into a rectangular box, parabolic expansion typically applies, and that's the approach I take for segments in the horn that represent one of the Hornresp sections referred to above, particularly if there are no folds involved.

However, if several segments and folds are being grouped together under one Hornresp section (this typically happens for the "L23" section for a tapped horn), it is possible to emulate conical or exponential expansion in some cases by careful choice of the lengths of those segments - basically approximating conical or exponential expansion by using a series of connected parabolic segments.
 
An update to the "Cyclops" workbook....

I've "improved" the folding at the mouth end of the horn, using a proper flare rather than a stepped section. I've also in the sim used the same driver as that used in the original Cyclops and dimensions similar to the CBE model (I rounded off to the nearest 1/2").

I think the improved layout of those fold should give a slightly better low frequency performance than the original CBE 18, which IMO wastes a bit of valuable box volume in the front by using a mouth that's too wide and a fold that does not take best advantage of the available space.
 

Attachments

  • 2018-03-02 (1).png
    2018-03-02 (1).png
    117.5 KB · Views: 361
  • 2018-03-02.png
    2018-03-02.png
    22.8 KB · Views: 358
Boring straight TL

Layout for a boring straight TL. This is probably the worst way to build a TL, but it seems very popular with the car audio crowd, probably because of all the "ringing" that likely takes place at or near the resonance frequency when you use a car audio driver to drive it. Anyway, here it is.
 

Attachments

  • 2018-03-07 (1).png
    2018-03-07 (1).png
    80.8 KB · Views: 294
Hi Brian,

Once you have a lot of experience building and simulating, what would be you theoretical assessment for the "Cyclops"? the applications, pros and cons, etc.

A good question. It's a big box that's tuned high because of how the path is laid out in the box. I'm tempted to widen the box a bit more and sim it with a 21" driver to see if the lowest resonance frequency can be moved closer to 30 Hz with the response remaining relatively damped. Best use is probably permanent install because of the size and weight.