The HOMster! (or How I Learned How to Fix a Horn)

[Patrick Bateman]

Originally Posted by gedlee

Either I don't understand or your numbers can't be correct. The most loss that I have seen from foam is about 3 dB and its usually more like 2 dB. The HF fall in a CD device because of the CD and this adds another 6 dB or so resulting in about a 10 dB boost required at 10 kHz to yield a response that is flat to 1 kHz. But only a few dB of this is foam, the rest is the fact that the energy is going over a much larger polar angle at HF than it is a typical system.

Earl,

Remember how some people didn't believe that HOMs existed, because they were difficult to measure? And subjectively, I have personally noticed that the foam makes a more audible improvement in horns than in waveguides.

My hypothesis is that bad horns generate more HOMs, and therefore the foam is more effective on bad horns. And the response plots reveal the story - there's a greater attenuation in a bad horn than in a waveguide, because a greater percentage of the horns energy is due to HOMs.

One way to visualize this is to compare it to a device where distortion is a significant percentage of output, for instance an amplifier that is being driven to it's limits, and a significant chunk of the output is harmonic distortion.

Anyways, that's the hypothesis. Measurements will tell the whole story. To demonstrate my hypothesis, I will measure four devices:

• A HOMster - a horn that's far far from ideal. The USD horns are an excellent example of a horns that's severely undersized, and whose dimensions are skewed. (IE, it's height been reduced by about 80% from the ideal.)
• A HOMster that's been treated - to demonstrate the effect of diffraction treatment on a horn that's far FAR from ideal.
• An OS waveguide - to demonstrate how a properly sized waveguide behaves, loaded with the exact same compression driver.
• An OS waveguide that's been treated - to demonstrate the effect of diffraction treatment on a properly sized waveguide.

Stay tuned for the results.

[kenpeter]

Have you considered breaking up HOMs with absorptive septums (balsa or foam??)
rather than non-directional foam full plug that absorbs -2dB of the dominant mode.

I'm not suggesting a cellular horn of exponentially shaped septums, just a plain
conical waveguide divided by thin flat septums. Where cutoff frequency of the
lowest possible HOM has been pushed slightly above the audible limit. Does this
make any sense???

[Patrick Bateman]

Originally Posted by Patrick Bateman

It's interesting to consider that at many frequencies, an untreated horn has more energy from reflected energy than from the primary wavefront. No wonder they can sound so poor!

Originally Posted by pos

How can you conclude that??

How can I conclude that an untreated horn has more energy from reflected energy than from the primary wavefront?

With a bit of math:

In the pic above, we have a graph of the same horn. One is fully treated, one is not. The bottom graph is shifted by ten db. At 4khz, we see that the treated horn is 6db less efficient. This demonstrates that at least 50% of the energy at 4khz in the untreated horn is due to reflections. The difference is particularly noticeable in the purple curve, which is 45 degrees off axis. At 2800hz there is an extraordinarily strong reflection, and the untreated horn is fully 10db more efficient at that frequency.

Keep in mind that this horn would not behave in such a strange manner if it was properly sized. Due to the strange dimensions, it's all but impossible to simulate in software, and conventional horn theory just isn't designed to predict how this thing will behave.

It is my hypothesis that this horn has particularly severe HOMs, and that is why the diffraction treatment is so noticeable in the plots.

As noted in the previous post, I intend to explore if that is true by comparing it to a waveguide.

[kenpeter]

You wanna break up HOM in a really bad horn??? Plug up a Karlson.
Gonna need even more Eq I think... They used to Laquer inside of
them things to deliberately sustain said reflections.

Wide waveguide, parallel walls, what better worst case can you ask?
Though it may be a HOM that builds up behind the central slot that
makes the thing work at all, foam might just kill it completely...

[Patrick Bateman]

Originally Posted by kenpeter

Have you considered breaking up HOMs with absorptive septums (balsa or foam??)
rather than non-directional foam full plug that absorbs -2dB of the dominant mode.

I'm not suggesting a cellular horn of exponentially shaped septums, just a plain
conical waveguide divided by thin flat septums. Where cutoff frequency of the
lowest possible HOM has been pushed slightly above the audible limit. Does this
make any sense???

I've considered that.

But I have another solution for the problem. Here's what I am working on at the moment.

A conventional horn that plays to 150hz needs to be very big. For instance, the length should be at least 23 inches, and the mouth needs to be very large... A diameter of 92 inches isn't "out of the question."

Now we can reduce the mouth size, but the frequency response goes to hell. The typical solution is to find a "happy medium", where ripple isn't too bad, and the box size isn't too big.

Danley stumbled across a novel solution to the ripple problem, which is basically to "tune" the ripples to smooth out the response. It works quite well for subwoofers, but not so great at high frequencies.

The reason that it doesn't work at high frequencies is that a "notch" is introduced into the passband, and that sets the upper limit.

For over three years I've been trying to figure out how to get a tapped horn to play full range, and I think that I'm very close now.

To demonstrate how close I am to the goal of a "full range" tapped horn, here's a distortion measurement of an Aurasound NS4 in a sealed box, along with my "full range" tapped horn. In the measurement you'll see some interesting things:

• This horn's response is +/- 3db for three octaves, from 200hz to 1600hz.
• Above 500hz, the horn's distortion is over 40db below the fundamental. To get this kind of performance in a sealed box requires very expensive woofers. There are other horns with comparable distortion performance, but none that are this small.
• Probably the most impressive aspect of this horn is that it has the best of both box types. It has low frequency performance that's comparable to a sealed box, a box size that's competitive, but distortion that's much lower, with a wider bandwidth than a tapped horn.

Like the HOMster, the device above has lots of HOMs. The difference is that this device is generating reflections in a very calculated manner, and then using polyfill to "soak them up." So you wind up with the low distortion of a horn, in box which would normally be far too small. (The entire enclosure is less than two liters.)

[kenpeter]

I've seen that shape before... What was it "Jensen Laboratory Standard"???

Drawing ripped from Radio Electronics April 1955 pg 31.
"Horn Type Speaker Systems" by George L. Augsperger
There was a PDF of that article floating about recently,
but I've lost track of the link.

The text of the article describes just what you were saying.

[kenpeter]

The Lab Standard was model RS-100, there was also an Imperial PR-100.
I'm having trouble digging up anything more informative...

[kenpeter]

OK, not the link I was looking for, but plenty of infos here...
The second one takes a while to load.

http://www.geocities.ws/footstony/audio54.pdf
http://www.superbadcat.com/jensen/imp_diy.pdf

[pos]

Originally Posted by Patrick Bateman

How can I conclude that an untreated horn has more energy from reflected energy than from the primary wavefront?

With a bit of math:

In the pic above, we have a graph of the same horn. One is fully treated, one is not. The bottom graph is shifted by ten db. At 4khz, we see that the treated horn is 6db less efficient. This demonstrates that at least 50% of the energy at 4khz in the untreated horn is due to reflections.

The foam also affect the direct wave, so this is unconclusive.
You should mesure the attenuation caused by the foam plug when a wave passes through it once (direct wave), and conpensate your fuly threated curve to really show reflections attenuation. Only then you could deduce the amount of energy from reflections.

[gedlee]

Originally Posted by kenpeter

Have you considered breaking up HOMs with absorptive septums (balsa or foam??)
rather than non-directional foam full plug that absorbs -2dB of the dominant mode.

I have considered this as conceptually it makes sense. Its the implimentation that is problematic. Compared to foam its a nighmare to invision how to make such a thing.