Another Unity Horn

[Patrick Bateman]

If I'm not mistaken, my thread where I put a unity horn clone in my car was one of the most extensive examinations of the subject ever done.

You can see it here:
unity in my car

Last night I read the synergy horn patent and it got me thinking about what I did a year and a half ago.

There were some mistakes I made, so I thought I'd examine that here in this thread. If anyone wants to pipe in with their observations, it could be an interesting examination of the subject.

[Patrick Bateman]

For YEARS I tried to wrap my brain around what was going on in the unity horn. Read all the stuff Sheerin wrote on snippets, read what Danley wrote on audioasylum, etc... I even went so far as to build a unity out of particleboard, with four high efficiency mids bolted to the side.

One little problem - I didn't know how it worked. Results sounded muffled and weird, and I gave up for a year or so.

Later on I had a big "eureka" moment when I read some things Geddes wrote about bandpass boxes. Geddes argued that bandpass boxes and compromised horns are no different, that they're two points on a spectrum.

THAT inspired me to throw THIS together: bandpass vs horn

When someone on audioasylum said that the midranges in a Unity are no different than a bandpass box, IT FINALLY MADE SENSE. Up until that moment, I didn't "get it", that the tiny tiny amount of air in front of the midrange cones on a unity behaves very much like the front chamber of a single-reflex bandpass box.

Unfortunately, I think I got it wrong! (oops.)

I was wrong about the midranges, and so was the person on audioasylum. If the midranges were bandpass, then the volume of air in front of the cone and the length of the holes would be very important. When I made the unity, I spent a lot of time trying different hole diameters and lengths. It was frustrating because it didn't make a huge difference in the response. Yet it should have, if the "holes" are equivalent to the port in a bandpass box.

Here's how I think it actually works. As I type this it seems really obvious that this is the correct answer, but I missed it when I built my unity. The midranges AREN'T bandpass; their response is dominated by the horn that they're loaded in. So you could use the exact same front chamber volume and hole size, but if you loaded it into a unity horn that was 30" across it would have a completely different response than if you loaded it into a unity that was 10" across. I'm going to go further out on a limb and say that the depth of the holes, which I was obsessed with, isn't a big variable. The CROSS section is HUGE though.

I'm babbling here, so let me get to the point.

In a unity horn, the midranges are NOT behaving like a bandpass. They're behaving as if they're mounted on a conical horn, but they're NOT driving that horn at the apex. The volume of air in front of the midranges can be calculated with standard horn theory and is equivalent to an optimally sized coupling chamber. The diameter of the holes and the volume of air in front of the woofers must be sized carefully, but I do not believe that the depth of the holes should matter a whole lot, unless resizing the holes dramatically changes the volume of air.

If anyone goes and reads the synergy horn patent, they'll see where I came to this conclusion.

Attached is a pic illustrating some of these concepts from standard horn theory, thanks to Martin King.

[Patrick Bateman]

Thanks to the wonders of hornresp, I can take a stab at modeling midranges coupled to a conical horn.

PLEASE NOTE that I'm not aware of a way to mount the mids anywhere but the throat. Yes, I know Akabak can do this, but I don't have a Unity model for Akabak!

The way that I'm going to cope with this little problem is by taking my midrange, then determining what the parameters would be if I used four of them. For instance, VAS MMS and SD will all quadruple.

So the values I'm plugging into hornresp are for FOUR, not one.

Attached are two pics, one showing the parameters for a 2inch tangband midrange which I've measured, plus a pic of the response. The paramaters I used are for a textbook horn. The sealed volume, coupling chamber, size of the holes etc...

You'll note the response is kinda crazy.

[Patrick Bateman]

frequency response of above

[Patrick Bateman]

schematic diagram. It's a 90 degree conical horn. The mouth is 14 inches across.

[Patrick Bateman]

Here's the hornresp prediction of the response of the midranges in my unity horn from the thread a year and a half ago. Wow, looks TERRIBLE! 18 months ago I modeled the response using a bandpass modeling program, but it looks like the response goes to hell when you mount it in a horn.

YUCK!!!

I'll have to dig up my measurements to see how well they match up.

[Patrick Bateman]

So I dug up my old data from the unity in my car.
The model in the previous post is off-kilter; the actual mids I built have a smaller front chamber, shorter ports, and there are two ports per woofer, not one.

I'll post a pic of the mids next.

Here's an update response graph from hornresp.

You'll notice there are a lot of high frequency peaks, and the sensitivity is terrible, just 86db per watt.

This correlates to what I found when I built my unities. I had a hell of a time getting the mids to blend with the compression driver, and the level of the mids seemed to be very low.

[Patrick Bateman]

Here's a pic of my midranges from my unity horn of a few months back. There are two aurasound two inch midranges jammed inside a PVC pipe, with four ports hanging off of each enclosure (two per mid.) Each port has a diameter of 1/4".

[Patrick Bateman]

Here's the kicker.
This is actual measured data of my "bandpass" midranges, coupled to a oblate spheroidal horn with a mouth of 8" x 10" and a coverage angle of eighty degrees.

You'll notice that there's a peak at 2khz and 6500hz.
In the hornresp prediction there's a peak at 1.8khz and 6.8khz.

Hmmm...

[MaVo]

In a unity you have two mechanisms that filter the acoustic output of the midrange driver. First you have the horn, providing a highpass function. Second you have the front chamber and the duct which connects it to the horn, which forms an acoustic lowpass. Thus, the midrange is highpass and lowpass filtered, equating in a bandpass.

I modeled this in akabak, but i am still building those things, so i have no clue about the accuracy of my script. ill post it, perhaps you find it usefull. Its a bit messy, since it was only used for my application.

System 'Unity_Segment'

Def_Driver 'WN-520N'
SD=103cm2 dD1=5cm tD1=2.5cm |Cone
fs=172.8Hz Mms=10.61g Qms=0.912
Qes=0.256 Re=6ohm Le=0.14mH ExpoLe=0.618

Def_Const | Variables are cm
{
A_Len= 2e-2; | Volumen under membrane
A_dD= 5e-2;
A_dDb= 10e-2; |
B_Len= 0.2e-2; | Front Chamber
B_dD= 10e-2; |
C_Len= 3e-2; | Ports (2 pieces)
C_dD= 0.8e-2; |
C_dDb= 3e-2;

Hi = 2.54e-2; | Horn throat diameter
Mid = 6.8e-2; | Horn width & height where the ports are
Mouth = 70.7e-2; | Horn mouth width & height
Hi_Len = 6.1e-2; | Length of the hornsegment between compression driver and midrange ports
Mid_Len = 87e-2; | Length of the hornsegment between midrange ports and the end of the horn
}

Enclosure 'E1' Node=100 Vb=0.4L Lb=0
Enclosure 'E2' Node=101 Vb=0.4L Lb=0
Enclosure 'E3' Node=102 Vb=0.4L Lb=0
Enclosure 'E4' Node=103 Vb=0.4L Lb=0
Driver 'D1' Def='WN-520N' Node=2=0=100=120
Driver 'D2' Def='WN-520N' Node=2=0=101=121
Driver 'D3' Def='WN-520N' Node=2=0=102=122
Driver 'D4' Def='WN-520N' Node=2=0=103=123
Waveguide 'Du_A1' Node=120=130 dTH={A_dD} dMo={A_dDb} Len={A_Len}
Waveguide 'Du_A2' Node=121=131 dTH={A_dD} dMo={A_dDb} Len={A_Len}
Waveguide 'Du_A3' Node=122=132 dTH={A_dD} dMo={A_dDb} Len={A_Len}
Waveguide 'Du_A4' Node=123=133 dTH={A_dD} dMo={A_dDb} Len={A_Len}
Duct 'Du_B1' Node=130=140 dD={B_dD} Len={B_Len}
Duct 'Du_B2' Node=131=141 dD={B_dD} Len={B_Len}
Duct 'Du_B3' Node=132=142 dD={B_dD} Len={B_Len}
Duct 'Du_B4' Node=133=143 dD={B_dD} Len={B_Len}
Waveguide 'Du_C1a' Node=150=140 dTH={C_dD} dMo={C_dDb} Len={C_Len}
Waveguide 'Du_C2a' Node=150=141 dTH={C_dD} dMo={C_dDb} Len={C_Len}
Waveguide 'Du_C3a' Node=150=142 dTH={C_dD} dMo={C_dDb} Len={C_Len}
Waveguide 'Du_C4a' Node=150=143 dTH={C_dD} dMo={C_dDb} Len={C_Len}
Waveguide 'Du_C1b' Node=150=140 dTH={C_dD} dMo={C_dDb} Len={C_Len}
Waveguide 'Du_C2b' Node=150=141 dTH={C_dD} dMo={C_dDb} Len={C_Len}
Waveguide 'Du_C3b' Node=150=142 dTH={C_dD} dMo={C_dDb} Len={C_Len}
Waveguide 'Du_C4b' Node=150=143 dTH={C_dD} dMo={C_dDb} Len={C_Len}
Waveguide 'CompDriver' Node=301=300 dTH=1cm dMo={hi} Len=2.54cm
Waveguide 'W1' Node=300=150 wTh={Hi} hTh={Hi} wMo={Mid} hMo={Mid} Len={Hi_Len} Conical
Waveguide 'W2' Node=150=160 wTh={Mid} hTh={Mid} wMo={Mouth} hMo={Mouth} Len={Mid_Len} Conical
Radiator 'Rad1' Def='W2' Node=160 x=0 y=0 z=0 HAngle=0 VAngle=0