Danley BC-subs reverse engineered

[just a guy]

I already posted the definition of diffraction, I know what it is. It's what happens when a sound source interacts with a boundary.

I don't think a reduced mouth area is required. Hornresp simulates diffraction (and the bubble) just fine without having to have a reduced mouth size, you can see that as you flip through the various Ang settings (no boundary, 1 boundary, 2 and 3 boundaries), but it assumes an infinite boundary size. Unfortunately it can't sim diffraction of finite size boundaries. Akabak doesn't consider the bubble as I've been told, so it has to be manually added to EVERYTHING (that has a port or mouth) that's simulated in Akabak. For all the difference the bubble makes, which is only a few inches at the mouth, and barely noticable in a sim. But Akabak can sim a finite size boundary.

I don't have a mic so don't wait on me.

There won't be any sim revisions, it will either match the measurement the first time or it won't.

 

[mark100]

Art, don't mean to steer away from main topic but in one of your posts....

"For instance, the LAB(Live Audio Board)Horn, Tom Danley's dual 12" FLH design that led to the development of the Lab 12" it uses, has around 315 degrees of phase "wrap" in the pass band of 35-100 Hz. If 14.5ms Offset Delay Correction is applied (as it can be in a FFT based measurement system that can measure acoustic phase), the phase response only deviates from "0" by 45 degrees through the same pass band- just like in "real life"."

.... For zero latency tops sitting flush on top of Labhorns, what delay would you give them? I've always just guessed at the length of the horn 9-10ft.

Thx, Mark

 

[sine143]

zero latency tops... pray tell where you found those

Even tops have group delay, esp if they are ported high (above 70).

 

[xrk971]

A direct radiator in a sealed box top can essentially near zero (very small, risetime limited) impulse response delay, and have minimal group delay (sub 2ms) if done right. Group delay and native delay as seen in the impulse response are two separate things. The delay seen in the impulse response will be caused by either phase lag in ported designs or by time-of-flight for horn designs.

 

[xrk971]

I already posted the definition of diffraction, I know what it is. It's what happens when a sound source interacts with a boundary.

I don't think a reduced mouth area is required. Hornresp simulates diffraction (and the bubble) just fine without having to have a reduced mouth size, you can see that as you flip through the various Ang settings (no boundary, 1 boundary, 2 and 3 boundaries), but it assumes an infinite boundary size. Unfortunately it can't sim diffraction of finite size boundaries. Akabak doesn't consider the bubble as I've been told, so it has to be manually added to EVERYTHING (that has a port or mouth) that's simulated in Akabak. For all the difference the bubble makes, which is only a few inches at the mouth, and barely noticable in a sim. But Akabak can sim a finite size boundary.

I don't have a mic so don't wait on me.

There won't be any sim revisions, it will either match the measurement the first time or it won't.

I think it may be a misconception that Akabak assumes a planar radiator that does not account for the curvature of the radiation from a horn mouth (or "bubble" as we call it). I know that David McBean has said numerous times that Akabak uses the "planar Webster" model for radiation. However, Akabak does have the capability to apply many small planar radiators with the curved convex geometry to approximate the output of a real horn bell mouth. One has to invoke the "Horn" radiator element rather than the standard "Radiator" element. When one chooses the "Horn" radiator, the sim takes about 20 times longer to run because it breaks the horn mouth into many smaller rectangular elements and calculates the effect of many radiators (kind of like a FEM) to provide the final answer. The horn geometry can be specified (radial horns, slit horns, continuous horns - page 188 in the Akabak manual for those who care). In my sims of the BC415, I have in fact, used the proper "Horn" element.

Here is an excerpt from the relevant section in the manual on how Aabak treats horn radiation:

For the simulation of the radiation in the mid and high frequency range of horns to be useful, you also have to take into account the flux in the direction of the two other dimensions. To a first approximation, AkAbak works with the 'radiation cone' of the horn: The Horn element is similar to the Waveguide element, but includes a special radiator. This radiator consists of multiple subradiators, each with the directivity of a rectangular aperture. These sub-radiators are placed on a sphere with vertical and horizontal convexities. The convexity of this radiation sphere can be entered manually or calculated automatically. In many cases, the radiation cone should be refined subsequently. In future versions of AkAbak, this model will be improved. Until then, this approximation will be adequate and will correctly show causes and effects with just a handful of parameter. The program calculates the radiation cone of a horn with exponentially flared walls automatically taking the point in the horn at which a ray starting from the center of the horn throat meets the horn wall tangentially (Fig. 102). Consider a cone whose tip lies in the center of the cross-sectional area of the horn throat and whose apex angle is equal to the angle of view when one looks into the horn from the horn throat. This apex angle can, of course, differ in width and height, depending on the horn shape. The convexity of the radiation cone is calculated from the curvature of the horn wall at the point at which the radiation cone contacts the wall (arc between Wc and Hc in Fig. 102). The surface is subdivided into numerous small rectangular areas. Each of these areas possesses the radiation characteristics of a rectangular surface. The greater the curvature of the radiation surface, the more radiation surfaces are installed. From this, the radiation characteristics can be ascertained. If the apex angle of the radiation cone is relatively narrow in width or height, or in both directions, and the curvature of the horn wall at this point is small, then the radiation characteristics of the horn resemble that of a single, small, rectangular surface or diaphragm. If the curvature of
the horn wall at this point is greater, the radiation surface is more greatly curved. The radiation then resembles more a small dome diaphragm. If, on the other hand, the apex angle is large, and the curvature of the horn wall
is small, the radiation characteristics resemble more those of a large rectangular diaphragm, so that their directivity frequency is at lower frequencies...

 

[weltersys]

.... For zero latency tops sitting flush on top of Labhorns, what delay would you give them? I've always just guessed at the length of the horn 9-10ft.

Mark,

For a top to have zero latency would require using FIR (Finite Infinite Response) filters, to get the FIR filters to have "zero latency" requires several ms delay to create a virtual (relative) "zero latency" (flat phase response) down to 100 Hz, more or less delay time depending on the top cabinets native phase response.
So the answer is I would measure, rather than guess what delay time to assign, but somewhere around 14.5 ms would be a place to start .

I will be purchasing DSP with FIR shortly, but presently I'm still in the "Dark Ages", using DSP with filter sets that introduce additional phase shift just like analog filters, then matching the phase shift through the crossover region for smooth phase response.

Art

 

[mark100]

zero latency tops... pray tell where you found those

Even tops have group delay, esp if they are ported high (above 70).

Maybe I should have asked "what delay should I add to the tops latency, when and if I ever figure out exactly what their latency is......"

Currently working on trying to understand REW.......

Also Art, I probably should have said Labhorn's x-over @ 70-80hz, currently using LR24. No filter on tops; just letting UPA-1p's do their thing with 9.4 ms delay....

 

[mark100]

Mark,

For a top to have zero latency would require using FIR (Finite Infinite Response) filters, to get the FIR filters to have "zero latency" requires several ms delay to create a virtual (relative) "zero latency" (flat phase response) down to 100 Hz, more or less delay time depending on the top cabinets native phase response.
So the answer is I would measure, rather than guess what delay time to assign, but somewhere around 14.5 ms would be a place to start .

I will be purchasing DSP with FIR shortly, but presently I'm still in the "Dark Ages", using DSP with filter sets that introduce additional phase shift just like analog filters, then matching the phase shift through the crossover region for smooth phase response.

Art

Thx Art, Yea, I've been looking at the miniDSP FIR boards, but they are waay over my head right now...

Like I said, trying to get a real grasp of REW.
Hey, quick question: I have an old Smaart 5.x license.....would you upgrade it or plow into the REW freeware?

 

[just a guy]

I think it may be a misconception that Akabak assumes a planar radiator that does not account for the curvature of the radiation from a horn mouth (or "bubble" as we call it). I know that David McBean has said numerous times that Akabak uses the "planar Webster" model for radiation. However, Akabak does have the capability to apply many small planar radiators with the curved convex geometry to approximate the output of a real horn bell mouth. One has to invoke the "Horn" radiator element rather than the standard "Radiator" element. When one chooses the "Horn" radiator, the sim takes about 20 times longer to run because it breaks the horn mouth into many smaller rectangular elements and calculates the effect of many radiators (kind of like a FEM) to provide the final answer. The horn geometry can be specified (radial horns, slit horns, continuous horns - page 188 in the Akabak manual for those who care). In my sims of the BC415, I have in fact, used the proper "Horn" element.

Here is an excerpt from the relevant section in the manual on how Aabak treats horn radiation:

Thanks for that. In my case I always start with a Hornresp sim imported into Akabak, that does most of the script writing work and I just have to tweak it at that point. So I have the radiator that Hornresp uses, which I believe is not the horn radiator.

I don't have the manual. I've looked for it but I can't find it. I've never seen it. I was going to ask if someone could send it but I don't even know if it's possible or what file type it is. All I have is the help file and it certainly doesn't have 188 pages.

 

[chrapladm]

I was running some sims at lower voltages and it is kind of mind blowing to fully grasp that one can get 125dB at 12volts. A small $20 class D TPA3116D2 in PBTL mode can drive this thing with some crazy SPL.

What simulation were you doing XRK971 for your results?

 

[Zero D]

@ just a guy

There should be 2 manuals in the Manual Folder. The big one is Manual Ak21 Main.pdf & Mine has 244 pages ! If it's not there ! you can DL it from their www

 

[xrk971]

I don't have the manual. I've looked for it but I can't find it. I've never seen it. I was going to ask if someone could send it but I don't even know if it's possible or what file type it is. All I have is the help file and it certainly doesn't have 188 pages.

When you install AkAbak the manual is automatically located in the "Manuals" sub directory I think as a PDF file. PM me if you still can't find it I can email it to you.

 

[xrk971]

What simulation were you doing XRK971 for your results?

I use AkAbak for the sims. I even use it instead of WinISD for bass reflex as I find the TS params input to be much less finicky.

 

[chrapladm]

Sorry to not clarify I meant what did you simulate to get 125db with only 12v's?

 

[camo888]

Sorry to hijack your thread but while you guys are talking about the BC415 design id love to hear peoples thoughts on what would be better 4 BC415's or 8 DBH218s? for mostly outdoor events.
DBH218s i figure will have more impact and punch and seem to dig pretty damn low too, fr graphs are very similar with the edge going to the BC415 for twice the power handling and slightly more sensitive so i figure 2 DBH218's = 1 BC415 they are easier to move around, 2 stacked will sit under stages better, 4 stacked high will be good hight for mains, or even 3, downside more boxes to move, more amp channels..
I've read everything I can on every forum and FB, just like to hear other thoughts on what would be a better setup and I can always ad more DBH218s to grow the stacks..


Only a few replies of ideas would be cool if this is not suited i can delete this post...
thanks

 

[David McBean]

I know that David McBean has said numerous times that Akabak uses the "planar Webster" model for radiation.

AkAbak uses the Webster plane wavefront model to calculate horn throat acoustical impedance, where the horn mouth impedance is assumed to be that of a plane vibrating piston. The throat acoustical impedance determines how much acoustical power is transmitted / propagated down the horn, and consequently radiated by the horn. A more accurate prediction can be obtained using an isophase wavefront model.

However, Akabak does have the capability to apply many small planar radiators with the curved convex geometry to approximate the output of a real horn bell mouth.

This is to calculate the directivity of the horn - a different issue to calculating the radiated acoustical power.

 

[David McBean]

Just to clarify:

the corner acts an as extended mouth to the horn and you get both gain and extension of bass.

The gain and extension of bass are due to the improved acoustical loading and reduced solid radiation angle at the normal horn mouth, as a result of the 0.5 x pi (eighth-space) conditions.

A flat wall or large baffle is sort of like this but relies on this mythical bubble to get gain and extension.

The gain and extension are due to the improved acoustical loading and reduced solid radiation angle at the horn mouth, as a result of the 2 x pi (half-space) conditions.

 

[David McBean]

So I have the radiator that Hornresp uses, which I believe is not the horn radiator.

Hornresp uses a Waveguide element plus a Radiator element, rather than a single Horn element.

This is done to standardise the specification of multiple-segment horns, and to make things consistent with how Hornresp analyses the system. For example, a four segment horn is specified as:

Waveguide1 + Waveguide2 + Waveguide3 + Waveguide4 + Radiator

Rather than:

Waveguide1 + Waveguide2 + Waveguide3 + Horn

 

[xrk971]

Sorry to not clarify I meant what did you simulate to get 125db with only 12v's?

The simulation I posted earlier showed SPL at max drive voltage of 108v, gave 142dB. Simply back off to 12v and SPL is 125dB. The sensitivity of a quad driver speaker with all drivers connected in parallel (2 ohm nominal) allows this. Even at 12v it is applying about 72watts. You need a 2ohm capable amp - which the TPA3116D2 can do when set up as parallel bridge tied load (PBTL).

 

[xrk971]

AkAbak uses the Webster plane wavefront model to calculate horn throat acoustical impedance, where the horn mouth impedance is assumed to be that of a plane vibrating piston. The throat acoustical impedance determines how much acoustical power is transmitted / propagated down the horn, and consequently radiated by the horn. A more accurate prediction can be obtained using an isophase wavefront model.



This is to calculate the directivity of the horn - a different issue to calculating the radiated acoustical power.

If the radiator at the mouth of a "Horn" element is indeed many smaller radiators with overall curvature, is the impedance at the mouth still that of a plane vibrating singular piston?

Regarding accuracy of isophase model - I am not sure how much more accurate the isophase assumption from the throat is in most practical horns where folds and bends are introduced which are then modeled in AkAbak using more segments of waveguides with expanding then contracting cross sectional area. For straight horns I will agree that isophase can be more accurate for systems where the wavefront has a lot of curvature. For relatively mild expansion ratio tapped horns built with flat boards in a conical expansion in 1-dim both approximations are good due to relatively low curvature while flow is still inside the horn. At the mouth the radiation models are treated differently and in AkAbak, the finite baffle size diffraction effects can be included, as well as nearby finite walls (up to three walls).

However, as most folks will agree that HR is superb for ease of use and user friendly features and for its ability to predict responses in a real time slider based wizard mode. I thank you for all you have done to give the DIY community such a superb tool. Frankly, most people here would not have been able to design all the TH's that are proliferating the threads without your software as they find AkAbak too daunting to overcome the learning curve.