The construction of a multicell horn

PCgab,
The Masonite exterior worked out well for me, along with filling the internals with the sand -granular rubber mix. Although the entire assemble now has weights 60 + pounds, I need to get an accurate weight measurement It took a little effort to fit the pieces to the horn assembly, I used thin cardboard as a templet for fitting each piece. If you can see in the pictures, I made a dado – groove at the half round molding around the horn side and a similar groove at the driver adapter side to slide the Masonite into the groves. The Masonite was cut long so when install they would “snap” into place and remain under slight compression. Hope that makes sense.

Wesayso
My concern is born out of inexperience interpterion of REW graphs:
How fast will a good titanium compression drive settle down after initial impulse?
What are the vertical irregularities in the X direction on the spectrograph?
In the waterfall what are the irregularities below 30dB and between 500 – 7k Hz?
Is that ringing?
I put some effort into bringing down the + 5dB frequency rise centered at 1300Hz with not
much success, every iteration of notch filter I tried, did very little to reduce the rise.
I want to make sure I am not chasing a base line issue I cannot correct without replacing the diaphragms.
 

Attachments

  • 2446h w gap.jpg
    2446h w gap.jpg
    316.7 KB · Views: 329
  • JBL 2446H Impulse NO HORN.jpg
    JBL 2446H Impulse NO HORN.jpg
    155.7 KB · Views: 336
  • JBL 2446H Driver NO Horn Spectrogram.jpg
    JBL 2446H Driver NO Horn Spectrogram.jpg
    175.8 KB · Views: 336
Administrator
Joined 2004
Paid Member
Great work on the horns! It's inspiring. :up:

Nothing other than FR jumps out at me on your plots. But it is odd that you can't tame the 1300 Hz hump. Could it be an artifact of measurement? Do you have DSP or other active EQ between REW and the amp with which to make drastic cuts that would be easy to see?
 
Pano,
No, I do not have a Mini DSP - something I can make direct frequency adjustments. I am using a Marchand MX44 two way electronic crossover, allowing XO point and levels between the line array woofers and the multi cell horns. The multi cell horn are turned down all the way and the woofers are turned up all the way - closed to balanced. been thinking about a mini-DSP, but heard they are a little noisy on high efficient speakers.

Just to be clear, you are referring as the throat, as the entrance to the horn section - where the driver adapter bolts to the horns ?

If yes, I squared off the multi cell assembly, using the Durhan's wood putty, shimmed up with MDF to create a flat surface on all sides. Made a flange adapter out of cardboard as a model, then reproduced with 1/2" ply that fit over the horn throat with lots of wood glue to fill void.Yes it did take some time - but the final product is worth it to me

Looks like I have been offered a set of used 2446H diaphragms in good shape for a good price. I will purchase then and have a bass line to measure.
 

Attachments

  • Mid Horns Home System.jpg
    Mid Horns Home System.jpg
    589.4 KB · Views: 363
  • Durham's.jpg
    Durham's.jpg
    815.9 KB · Views: 371
Pano,
No, I do not have a Mini DSP [...] been thinking about a mini-DSP, but heard they are a little noisy on high efficient speakers.

I've used a mini DSP with sensitive gear - JBL 2445 on horn + 12" on horn (2.4m long rough outdoor rig), and the noise floor was fine. It is also very little money to risk.

The only time the noise floor has been any trouble is when plugged into a desktop PC.

Maybe earlier versions were noisier.

I have rejected several bits of cheap PA gear (power amps, an active crossover, some powered speakers) for having too high a noise floor.

Gutting / rebuilding the powered speakers is my current project.

Almost forgot to mention - grand project you've made :)
 
Last edited:
Maybe you want to replace the stock diaphragm by a Radian 1225-8? The JBL 2425 and 2426 titanium drivers are more durable for stage use, compared with the preceding 2421 aluminium driver, and may sound somewhat harsh, while the Radians are supposed to sound better/softer in HiFi and studio applications.
Best regards!
 
Administrator
Joined 2004
Paid Member
Thanks Tom. :up: I was curious as to how you made the transition from round exit of the compression driver to the larger rectangular opening on the horn. Can't have been easy.

Have you tried some passive attenuation on the JBL drivers? 6-8dB could help give you more range on your Marchand. Some people claim that compression drivers sound better with some resistance between them and the amp. I can't say that I've found that, but I have normally used it just because it helped the gain structure.
 
Pano,
admittedly I can only speak of JBL compression drivers. But judging by how they implemented them in their own speaker systems, you may be quite right. Anyway, I've always thought that the L-pads and voltage dividers simply served to align the different sensitivities of the CD/horn combo vs. the woofers and squawkers.
Best regards!
 
To all,
Thanks for the suggestion - all good ones...

Did not think about an L-Pad on the horns to reduce the level

Will look into a replacement diaphragm options.

Will also consider the mini DSP option, was trying to keep the system simple with one amp, but I would give me more flexibility


Tom in Lafayette
 
Administrator
Joined 2004
Paid Member
12 cell 3x4 horn

Here is a spreadsheet that will help you with the dimensions and profiles of a 12 cell horn 60cm wide by 45cm tall by 47cm long. This makes it approximately a 320 Hz exponential flare. This is close to the size of an Altec 511, but a little taller. The horn will be a little wider than 60cm because of its geometry.

The cells are each 37cm long with a mouth of 15cm including a 3mm wall thickness. If you alter the cell wall thickness, the dimensions of the horn will change accordingly.

The throat block is 16cm long and the cells should fit down into it 6cm. That gives an overall horn length of 47cm. You can make the overlap shorter if you wish. The small ends of the cells should be sanded to a thin wall as shown in the photos early in the thread. Currently the throat is calculated to fit at 10cm if cell wall thickness is 1mm at the small ends.

In the spreadsheet you will find 2 pages. One is the profile of the cells, the other is the profile of the throat section, starting with a 1"x1" square opening. The dimensions you want to trace out are marked in yellow. These are in mm. The only values you can change are for wall material thickness. See the green cell.

You will need to make the square to round transition at the throat opening as seen in the original photos.

Of course you'll want to make a cheap prototype or two before you ever cut wood. Perhaps poster board or cardboard. Beware, I have yet to use this for building.

I will be building a 46cm wide prototype from foamcore that is about 4.5mm thick. I will let you know how that turns out. :up:
 

Attachments

  • 12 cell 60cm.zip
    8.1 KB · Views: 159