Ever think of building a Cornu Spiral horn? Now you can!

Started with this, hung a coil and a cap on the original set up and put the two to the test.

Big improvement.
 

Attachments

  • IMG_1027.jpg
    IMG_1027.jpg
    57.4 KB · Views: 880
  • IMG_1028.jpg
    IMG_1028.jpg
    96.2 KB · Views: 735
Then I introduced the CSS VWR 126 to the mix to replace the Fostex WR125.

We've now gone from having fun to being serious. This seems to the next step. Again, not the end all but a big 'wow' is in order. I will work on the stuffing but this is something the Fostex was not able to do.

Lots of experimenting with the XO at this point, all 6dB, all by ear.

Wifey says turn it down...

... I love it. :D
 

Attachments

  • IMG_1030.jpg
    IMG_1030.jpg
    106.1 KB · Views: 693
Founder of XSA-Labs
Joined 2012
Paid Member
Very cool Cal! What do you think the crossover freq is about? The CSS VWR was what you were planning on using in the first place. You have me thinking now.... I have my wife's old JVC receiver with active crossover and built in bi-amp. Maybe connect my 20 in to the sub woofer and 14 in to the full range amp outs?

The large size makes me think this is like an OB with edge output. :)

Awesome experiment. Tell us more about the listening impression.

The last photo still shows the FF125wk?
 
Founder of XSA-Labs
Joined 2012
Paid Member
I built a 14 in cornu with Logitech 2 in drivers from s120 pc speakers. Works really well. Mine is 2 in deep. Assuming your driver Sd is about 11.5 square cm, The 2 in depth is about right to match the driver Sd.
The CHR70 is one I would love to see how it works out. The Sd on CHR70 is about 50 square cm so the depth will need to be at least 4 in for a 24 in cornu.
Good luck!
 
First Build

Here's my efforts for the weekend (and it's not over)! Throad area is about 2.5" sq., Sd for my speakers is about 5" sq. (Tang 3" Buyout). Speakers will arrive on Wednesday. I'll stuff the behind and channels out to about the division with pillow stuffing.

Does anyone see any glaring problems? I agree with Cal that these are a blast to build!
 

Attachments

  • Photo138.jpg
    Photo138.jpg
    154.1 KB · Views: 842
  • Photo139.jpg
    Photo139.jpg
    140.7 KB · Views: 826
Nice work j. The only caution I offer is the stuffing. You can use a fork [almost perfectly shaped] to add stuffing to the channels but it's a sonofagun to remove it if you have too much. I would recommend using something a little hardier like rockwool insulation and heavily damp the chamber and the opening of the throats. You can add more if you think it's necessary.
 
Founder of XSA-Labs
Joined 2012
Paid Member
Jsouder,
Looks great! :) Awesome job on the spirals. Make sure that the channel depth is uniform so that a good seal is made when you cap it. I sometimes use hot melt to build up layers on the channel edges that are too low. What is your channel depth, and I assume you are making the 20 incher? If doing a 20 incher, a 3.5 in depth should work well.
 
Jsouder,
Looks great! :) Awesome job on the spirals. Make sure that the channel depth is uniform so that a good seal is made when you cap it. I sometimes use hot melt to build up layers on the channel edges that are too low. What is your channel depth, and I assume you are making the 20 incher? If doing a 20 incher, a 3.5 in depth should work well.

Planning on using a caulk/glue for the top. I'll smooth down any of the edges where the pieces lined up. I think any changes will be < 1/16" which the cault should close.

The channel depth is 2.5", so the open area sum of both throuats is about 1/2 the Sd of the speakers.

I'm about to embark on a true spiral set. I'll try the 3.5" channel depth.
 
Founder of XSA-Labs
Joined 2012
Paid Member
Effect of Throat Area and Driver Chamber Volume on Freq Cutoffs

Hajj,

I am still in the midst of designing the next one, but it will probably be based on what I call the "True Cornu Function": 24 inches x 3.5 inches deep with the mouth size spanning the entire perimeter for an area of 24x3.5x4 in^2=336 in^2 or 2167 cm^2. The throat will be 31.3 cm^2. The driver chamber will have an area of 148 cm^2 and a volume of 1312 cm^3. The horn lengths will be 104 cm (Long path) and 49 cm (Short path) with a bifurcation. The path plan is attached.

If you can, please post the info on the resonator for the 70 cm case because many folks are interested in that size.

Thanks.

Xrk971

OK folks, I have been doing some studying of horn theory using MJK's nice papers. I have been simplistically scaling the horn's depth to achieve a matching Sd for the driver. That is only looking at half of the picture. The horn's upper cutoff frequency is determined by the relation (Eqn. 5.4 in MJK's Horn Theory paper):

fh = (c * S_throat) / (2 * pi * Vol)

where c is speed of sound, S_throat is throat area, and Vol is chamber volume. Since throat area is related to chamber volume when we increase channel depth, the only way to independently change volume yet still have a big enough throat is to either add depth to driver chamber but leave channel height alone, or to add volume filler in the chamber. Based on the sizes we are dealing with here for the typical 70 cm or 51 cm cornu, optimal performance means adding volume filler to the chamber. This may explain why heavily stuffing chamber works.

For example, the typical 70 cm chamber volume is based on a roughly 200 cm^2 cross sectional area x 11.4 cm = 2.28 liters. The throat is 2.3 cm x 11.4 cm x 2 = 52 cm^2. Using the upper cutfoff frequency equation we get:

fh = (342 m/s * 0.0052 m^2) / (2 * 3.1415 * 0.00184 m^3) = 124 Hz

Although, this cutoff freq is OK for some, it will highly attenuate the nice mid-bass which can extend up to 300 Hz to give the horn a more balanced and pleasing sound otherwise it sounds too bass heavy. If we pack the driver chamber with enough stuffing such that the void/volume goes up, this drives the cutoff frequency up into the 300 Hz range.

Alternatively, if one does not want to use too much stuffing because that kills the efficiency of the horn output, a solid object with volume can be placed in the chamber. Like the protrusion of the cup from a binding post mount, or in the case of the original cornu, a conical deflector at the bottom which serves to reduce the direct back reflections from the back wall and reduces the volume.

Now, take the example of my first cornu in which I heavily modded the plan so that I could achieve a 34 cm^2 throat area and keep only a 6.04 cm depth. I got lucky because my cutoff frequency ended up being:


fh = (342 m/s * 0.0034 m^2) / (2 * 3.1415 * 0.000623 m^3) = 300 Hz

So, I ended up with a pleasing upper cutoff freq to start with and adding some stuffing actually stretched it upwards of 400 to 500 Hz which may be too much upper bass to the point of phase delay is audible.

So basically, now that we are aware of the the effect of the chamber volume, and know that adding stuffing or maybe a cone deflector at the back will help extend the upper cutoff and also reduce back reflections.

Next is what is the effect of the throat area? From MJK's Horn Theory paper, the throat area (S_throat), coupled with the horn's mouth area (S_mouth) and the length of the horn (L) determine the lower cutoff frequency. The relation for an exponential horn is given by this equation:

fl = (m * c) / (4 * pi) where m is the exponential expansion coef

In the case of the cornu, the expansion ratio for a single horn (and we have to look at just one horn because the maximum wave that can be supported at the mouth of a single horn is what we are interested in) is about 15:1 (S_mouth/S_throat). Thus the lower cutoff freq fl is calculated as

m = ln(15)/2.2 m = 1.23 m^-1

Now

fl = (1.23 m^-1 * 343 m/s ) / 6.284 = 67 Hz

which is a good lower cutoff. Note here that in the case of the cornu, the depth will change the throat and mouth equally so it depends on only the expansion ratio and the length of the horn.

In summary, the typical cornu will require either lots of bulky stuffing in the driver chamber or a solid object to displace volume in order to extend the upper cutoff frequency to a pleasing range. This can be achieved also by highly modifying the channel plans so that a larger throat is achievable in a thinner channel depth. With a foam core build it is also easy enough to make squares or rings to glue to inside of chamber to take up volume. A stepped edge cone can also be made (like a zigurat) to provide both breakup of the back wave and also take up volume. Other things to try for stuffing might be foam peanuts in a mesh bag, mixed with stuffing. For the new true cornu designs I will definitely have to add volume fillers. So, if your cornu appears to have anemic mid bass - which is really where the good bass is, try experimenting with putting filler volumes in the driver chamber.
 
Last edited:
Nice work xrk,

Has Martin chimed in on these threads?

I've started working on a 20 incher this weekend. Will also use the Vifa 3.5. So, xrk, would you change anything based on the what you've learned? Is the 2-3/8" depth still ideal for the Vifa?

Has anyone tried a 2 sided spiral? like the original Cornu?

FYI, Michael's has the white Elmers 20" x 30" foam board on sale, 2 for $3.00 until 1/19/13. I picked up 4 to use for the outside pieces - insides will be the Dollar Tree stuff.
 
Founder of XSA-Labs
Joined 2012
Paid Member
Kec,

Thanks.

The 2-3/8 in size worked for me because I modified the channels to have a 1.1 in wide throat at the chamber. If you were to use the standard plan that everyone prints out, at a 20 inch scale, the throat is only about 0.6 in wide. So you will need a channel depth of about 3.5 to 4 inches to get enough Sd. But that increases the volume which will drop your upper frequency cutoff. To compensate, make a block of foam stacks inside and fill it up to reduce volume while leaving throats clear. Make a stepped cone like a zigurrat.

No, no one has done a 2 sided one yet. I have thought about it - basically apply spirals to both sides of the back and cut a hole in it. It can be done serially so basically it would be like making one at a time. This would simplify the messy step of gluing the face on.

MJK has not chimed in - I would love to hear from him though.

Regards,
X
 
Last edited: