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

[xrk971]

Originally Posted by chazdrumzalot

XRK's dream is a car in every garage, a chicken in every pot, and a cornu horn in every living room.

Yes, touche.

[Einric]

The Home Depot has 5.2mm Lauan plywood in 2x4 sheets for $6 or 4x8 sheets for $12.
It usually has a bad side and the good side is a pretty nice looking variant of Mahogany.
This is what I'm thinking of using.
Still using the foam core for the spiral.
I should be able to build 2 cornu's from one full sheet.
I also have some nice 3/8" hickory to make a "Supra Baffle" for flush mounting the driver.

[xrk971]

Einric,
The Lauan ply should work. What size were you thinking of making?

[Einric]

24" also, I may try having two mouths firing forward.
That might help the speakers integrate into the room better.

What do you think about some of that really porous (think speaker grill) foam in the mouth opening.

What do you suppose the throat height should be for the RS100?

[xrk971]

Einric,
The sound firing out the sides is better because it uses the wall as an extension of the mouth. The sound in the side firing config integrates superbly in the room and is what I believe is part of the cornu's ability to sound bigger than it is. If you redirect to forward facing horn mouth, you will lose this effect,. I think you can modify to forward emitting later if you still desire. I would recommend you build and hear it side firing first. The RS100 can work nicely at 24 in size with 4 inch depth. Porous (reticulated) speaker grill foam should be fine.

[hajj]

Thanks for the great threads Cal and xrk971 (not done with the foam core thread yet )

While going through the threads I was constantly looking for a methodology concerning the sizing of the chamber and the horn paths.

Gpapag addresses some of the limitations in post #102, page 11 of this thread by pointing out the phase cancellation issues that might arise.
This limitation can be vastly mitigated by addressing the design of the "compression chamber" at the back of the driver allowing it to work as a properly adjusted low pass filter and also adjusting the horn length so as to better match each driver's specific T/S parameters.
A rule of thumb is to limit a horn's output to two octaves, while using stuffing to kill the higher frequencies that might travel through the horn, and thus minimizing the phase cancelling that might occur higher up.

I think that a more systematic approach for the most popular drivers would make more people embark on the endeavor.

Being completely incompetent with AKABAK, I'll try to whip something up using MJK's worksheets, but this might take some time, and will be limited to one horn length (the bifurcation being impossible to model using Martin's worksheets).

I currently have a pair of MA CHR-70, a pair of CHP-70 and a pair of FF125wk all looking for an enclosure.

I'll try to model the CHR-70 first as I think it's a good fit with its peaking response.

Regards

Nick

[hajj]

The chamber's size might seem to be directly related to the overall size, throat opening sizes, as well as horn length, but this does not have to be the case should one consider having an extra baffle under which the extra chamber volume could hide.

[xrk971]

Hajj,
That's a great idea to do a grass roots study of how to optimize the cornu based on driver and real phase cancellation. I am trying to do the same thing right now but do not have MJK. I have been using combination of HornResp and Akabak. One approach to using MJK is to model each horn length independently and redesign the cornu to have two lengths starting from throat chamber. Adjust lengths so that when response of both horn lengths are laid on top of each other, you get one horn's peaks filling in the other horn's dips. Horst Moeller (diyaudio 'hm') uses a double horn approach in his Kornet design but he has two independent drivers. The driver compression chamber can also be adjusted by expanding the central cavity and then wrapping the horn around that. Basically to do this approach we start with a new flow path that is custom designed but driven by the results of the simulation. Thanks for taking this on and I look forward to your MJK sims. One things that would be very useful is to see where and how much stuffing to put in channels to smooth things out.
Regatds,
Xrk971

[hajj]

Originally Posted by xrk971

Hajj,
Adjust lengths so that when response of both horn lengths are laid on top of each other, you get one horn's peaks filling in the other horn's dips.

My initial impression is that this approach would not take into consideration phasing at the horn's mouth, and thus could lack accuracy.

Horst Moeller (diyaudio 'hm') uses a double horn approach in his Kornet design but he has two independent drivers.

Should the dips be high enough in frequency, a front horn could tackle some of the problems (it has been brought up already in the foam core thread I think)

The driver compression chamber can also be adjusted by expanding the central cavity and then wrapping the horn around that. Basically to do this approach we start with a new flow path that is custom designed but driven by the results of the simulation.

Problem is, I am mathematically challenged (don;t be fooled by my clever-looking posts ), so this part will need to be tackled by someone better fit than I am

Thanks for taking this on and I look forward to your MJK sims. One things that would be very useful is to see where and how much stuffing to put in channels to smooth things out.

You're welcome! It's just another step in my learning curve, but at least it's a fun one, these speakers are definitely next on my list, after i'm done with the Silbury currently lying in pieces on my work bench

Stuffing placement and density should be easy to figure out using MJK's worksheets.

[chazdrumzalot]

There are a couple of other nagging issues with this design which could/should be addressed, if possible. One thing on my mind is the problem of putting the driving smack in the middle of a large square. Baffle diffraction, anyone? Anyway, enough talk; time to mount some drivers.