Some questions on horn design

[e-side]

Hi,

Could someone help me with the following questions?

1. Currently I'm experimenting with Leach's math for low frequency horns. It seems that when the horn's bandwidth gets wider, the system efficiency decreases and the throat gets smaller. But according to 'classic' horn theory the efficiency increases when the throat is made smaller. How come these theories predict different results?

2. Does the compression ratio play an important role in horn design? What are the consequences of a CR that is too high?

It seems that Leach's math results in quite small throat areas, but the simulated response is flat in most simulations. Also, in the "System design from specifications"-section Leach uses a CR of 4:1, which is considered as being high. Why did he do that?

3. For the driver I'm using/simulating, it seems that I can get a flat response only when I use a high compression ratio. Increasing throat area results in dips and peaks. Should I conclude that this driver is not suitable for horn loading?

Are there any other horn design methods/approaches, other than Leach's? Or is Leach the only correct approach for calculating system parameters for a given bandwidth?

4. It seems that the throat on midrange horns is made equal to Sd (CR 1:1) in many designs. Is this just a 'rule of thumb', or can it be calculated?

5. How important is the voice coil inductance corner frequency? Does it play a major role on the HF response of a (midrange) horn?

6. Is a throat chamber (for HF rolloff) required for a midrange horn?

7. It seems that some people substitute Qes for Qts in design formulas. Maybe because Qts ~ Qes is assumed. But in case of a dual voice coil driver with one coild shorted, the difference between Qts and Qts is significant (Qts ~ 0.5Qes); So it becomes important to know which one I should use. Could someone explain this?

I greatly appreciate every reaction.

Erwin.

[e-side]

Another question came up to my mind.

In his Loudspeaker Design Cookbook, Dickason mentions that a driver's dustcap made of textile causes air leakage. Could this become a problem when the driver is used for horn loading, because of high pressure at the throat section?

[planet10]

Cloth surrounds can leak to. Yes this can be a problem (or a feature)

dave

[bear]

Since I have CRS, and have not looked at the horn design equations in a long time, I may get some things wrong or have forgotten them...

Originally posted by e-side
Hi,

Could someone help me with the following questions?

1. Currently I'm experimenting with Leach's math for low frequency horns. It seems that when the horn's bandwidth gets wider, the system efficiency decreases and the throat gets smaller. But according to 'classic' horn theory the efficiency increases when the throat is made smaller. How come these theories predict different results?

Something is funny here - a higher compression ratio implies more efficiency, iirc. BUT IF the bandwidth is increased, the efficiency decreases, so this may account for what you are seeing. Keep the bandwidth the same, and the efficiency should increase.

2. Does the compression ratio play an important role in horn design? What are the consequences of a CR that is too high?

Distortion. Changes in the polar response. (horn geometry changes)

It seems that Leach's math results in quite small throat areas, but the simulated response is flat in most simulations. Also, in the "System design from specifications"-section Leach uses a CR of 4:1, which is considered as being high. Why did he do that?

I presume ur talking about cone drivers, not compression drivers?
It means that the forces acting on the cone are high, and that there is potential for added distortion. Not to mention destroying the cone by rips or tears... depending on the levels the thing is played at. Keep in mind most horns were used for PA/SR where the levels are near or at max for the drivers.

3. For the driver I'm using/simulating, it seems that I can get a flat response only when I use a high compression ratio. Increasing throat area results in dips and peaks. Should I conclude that this driver is not suitable for horn loading?

Not sure.
There is a figure of merit for cone applications.
Also, there is an implied relationship between the driver and the horn - namely that the horn supplies the necessary boost in LF response to make the overall response look flat. Above some freq the horn really doesn't do much of anything. Below some freq, the horn adds gain.

Flatness in a horn usually depends on adequate mouth size to support the LF without "comb filter" looking dips, and the proper flar e rate for the freq of interest - but the mouth size is a big deal.

Are there any other horn design methods/approaches, other than Leach's? Or is Leach the only correct approach for calculating system parameters for a given bandwidth?

Yes.
Usually the methods are based on Webster's equations.
Many variations and references, including online.

4. It seems that the throat on midrange horns is made equal to Sd (CR 1:1) in many designs. Is this just a 'rule of thumb', or can it be calculated?

This is imho a kludge in reality - as above the idea is likely to "extend" the LF response of a given driver. So, you don't want the horn to have much effect except where the driver starts to drop.

Think about what a standard compression driver sounds like without a horn on it? All highs... rolls off going down in freq... the horn pulls that back up. Ok?

5. How important is the voice coil inductance corner frequency? Does it play a major role on the HF response of a (midrange) horn?

It can.

6. Is a throat chamber (for HF rolloff) required for a midrange horn?

Only if you want it to roll off at some HF by acoustic means.
Reasonably common technique in bass horns, though...

7. It seems that some people substitute Qes for Qts in design formulas. Maybe because Qts ~ Qes is assumed. But in case of a dual voice coil driver with one coild shorted, the difference between Qts and Qts is significant (Qts ~ 0.5Qes); So it becomes important to know which one I should use. Could someone explain this?

Consider what this means in practical terms.
How does the excursion of a standard driver look wrt to frequency?
IF you kept the excursion "constant" wrt frequency, what would the freq response look like?
Now, if you took that second driver and hooked it up to a horn, what would the freq response look like with and without the horn?
Qe dominates in the horn loaded situation... I think that is why Qt is disregarded.

Btw, I do not think you want to short one coil.
That will create extra heat, as the coil will try to act like a brake.
Reduce output too...
Maybe there is a good reason to short a VC, but I'm not sure what that would be...

I greatly appreciate every reaction.

Erwin.

_-_-bear

[e-side]

Hi bear,

Something is funny here - a higher compression ratio implies more efficiency, iirc. BUT IF the bandwidth is increased, the efficiency decreases, so this may account for what you are seeing. Keep the bandwidth the same, and the efficiency should increase.

That makes sense. I thought that there's a relation between throat area, efficiency and upper cutoff frequency. I came to this thought when I simulated several equally efficient designs, but with different lower cutoff frequencies. It appeared that throat size, system efficiency, upper cutoff and front cavity volume remained unchanged, no matter what lower cutoff I chose.
Because of this observation/relation it seemed reasonable to me that efficiency can changed only when bandwidth does too.

That's also why I wasn't sure about the 1:1 throat on a midrange horn; Such a throat would result in a very narrow bandwidth. However, Bruce Edgar mentions in his Midrange Horn article that the upper end response increased by making the throat equal to Sd...

I presume ur talking about cone drivers, not compression drivers?

Yes, I meant cone drivers.

Yes.
Usually the methods are based on Webster's equations.
Many variations and references, including online.

I have searched quite a long time but I haven't found any other methods except Leach's and Keele's. Could you provide some links?

Thanks for your help.

Originally posted by e-side
Another question came up to my mind.

In his documentation on Front Loaded Horn Design, M J King discusses high Q drivers in horn systems. His simulations show a flat frequency reponse for these drivers, usually considered as being unsuitable for horn loading.
Although I certainly don't question the accuracy of his simulations, HornResp shows different results for these high Q drivers. Which tool is more 'reliable', and why?

regards
Erwin

[e-side]

anyone?

[Juergen Knoop]

don't know if it is more reliable or more accurate, but ajhorn seems quite popular
regards

[Ed LaFontaine]

http://www.diyaudio.com/forums/showt...60#post1217060

[Ed LaFontaine]

How does Leach compare with LeCleac'h?

http://ndaviden.club.fr/pavillon/lecleah.htm

click on: "Telechargement"

http://ndaviden.club.fr/outils/axial.zip

[GM]

Greets!

Hmm, increasing the CR implies an increasing acoustic power/watt, ergo decreasing the CR implies an increasing acoustic efficiency and wider BW.

ML's math is for compression loaded horns while JMLC's is more along the lines of low diffraction termination waveguides.

I don't use Hornresp, but MJK's sims of ML's math have shown good enough correlation with my 'adventures' in using non-optimal drivers with Qes as high as ~3.3! to satisfy me, though as always YMMV.

Bottom line, in any line of work (audio app), you want to use the right tool (driver) for the job (speaker alignment), so using non-optimal drivers requires trade-offs that you'll have to choose which ones are acceptable.

GM