A comparison of the phase-response from the THBP (black line) and the FLH from Just a guy.
If this is the real phase-response then i think it is safe to rule out FLH as a contender for the BC415.
It is also interesting to compare the Max spl from this FLH (from Just a Guy) to the max spl from the TH and the THBP...
The FLH has way to much compression. The hornpath is much smaller then a good bass-reflex port would be. The driver would not last long in the intended usage case.
Cheers,
Johannes
If this is the real phase-response then i think it is safe to rule out FLH as a contender for the BC415.
It is also interesting to compare the Max spl from this FLH (from Just a Guy) to the max spl from the TH and the THBP...
The FLH has way to much compression. The hornpath is much smaller then a good bass-reflex port would be. The driver would not last long in the intended usage case.
Cheers,
Johannes
Attachments
Last edited:
It's the right picture but it's the wrong phase. Notice Hornresp automatically sets the delay at 81 ms for the flh (not sure why but it's clearly not right) and 9.2 ms (or something) for the other design (which is much more reasonable based on path length).
If the response curve is the same and the path length (delay) is the same, the phase will be the same. Adjust the delay manually and see for yourself.
Don't rule out flh as a possiblility for the BC subs yet.
I'm not sure if you missed my point or not. I wasn't trying to make a great design, I was trying to copy what you had as exactly as possible with different alignments, this required a bit of compression in some alignments. The compression is clearly too high but that's not the point. Compression can be lowered, the design can be changed to have the same amount of bass but a different response curve shape. (My first example shows this, the simple tapped horn has same compression ratio as your design, they put out the same amount of bass but the response curve shape is different, IMO better, with the simple tapped horn.)
My point was that given the same size, no alignment is a clear winner, they all put out the same amount of bass.
Your original design is undersized (smaller than optimum), that forced some compromise in my examples, like high compression.
Last edited:
Also note that the published info is marketing brochures, not a scientific article, so don't believe anything you see. If you count up the number of misleading or just plain wrong info on one of their spec sheets it's just like any other form of advertising.
If you want to have another little experiment to find the best possible design in an undersized box (your original design is undersized) we can do that too. Name the driver, enclosure size and any other rules you want. BUT "best" is subjective and depends on goals of the designer. IMO, my first example (the one that didn't have the same response curve shape as yours) was a much better design than all the others, mostly because it didn't have such a weird response curve shape and it has a lot more bandwidth.
Anyway, I think at this point I've cleared up a lot of your claims, a straight 1/4 wave resonator is not more powerful than a tapered one, your method is not unique (it's a stepped design, no different than using taper with 2 flare rates), other methods and flare shapes produce the same results as your method, you don't need a lot of compression ratio to get the same results you show (see my first example).
I have to go to work now. Have fun.
The tapped horn using both sides of the driver does not add 3 db. Tapped horns are not more sensitive or efficient than transmission lines of the same size and shape. The extra mouth side tap does almost nothing except fill in the naturally occurring transmission line dip. Here's an example.
Graph 1 - a random tapped horn with a single flare rate
Graph 2 - same thing but set to Od so it's now a normal offset driver tl
Graph 3 - same thing (Od tl) but driver shifted down the line to the mouth side tap spot
You can see the mouth side tap does not add 3 db, it doesn't do much of anything except fill in the dip in Graph 2. It's a common myth but it's not true. Like I said yesterday, enclosure size determines resonant output potential. Flare shape (and driver location) determine the response curve shape. Horn shaped tapped horns are only louder than straight transmission lines because they have to be much larger to achieve the same tuning and the extra size allows more resonant output.
Last edited:
Hi Circloman, Art , Guy, all
It would take all the fun out of this stuff I posted drawings (and it would be frowned on at work) and besides, pretty often one or more of you eventually comes close or nails it.
The effective length of a horn is actually different than one measures with a tape measure because at the mouth of a normal symmetric horn, the pressure boundary is an arc where the center sticks out forward of the cabinet.
If I remember correctly, the old “rule of thumb” suggested the mouth bubble for a simple round horn is about 60% of the mouth radius, in other words, at the center of the horn, it protrudes that much forward of the wood. For a cd horn operating where it large enough to have directivity, that boundary is nearly perpendicular to the horn wall, following Keele’s pattern loss thumb rule.
Also, the “horn loading” effect has a “high pass” filter function based on the rate the area expands. For instance for a 30Hz exponential horn, the area cannot expand any faster than doubling about every 2 feet while for 300Hz, the area doubles about every 2.4 inches and so on.
The object of the BC horns is to have the rate of expansion proper so that the end of the bubble is formed by the 60 by 60 inch face dimension and have the inner part of the horn tie acoustically into that external bubble. The horn inside of the cabinet has a larger mouth than the opening in the side of the cabinet and but forcing the output to pass through a restriction puts it in the right place to tie into that bubble and in effect makes the mouth larger and horn path longer.
The restriction has a minor “low pass” effect and adds a little extra acoustic mass, neither of which are disadvantages. Internally, the BC speakers are regular front loaded horns, all of them are large enough where the Tapped design isn’t an advantage AND tapped horns “feel” the external loading less than FLH’s.
Several of you have mentioned an important thing and Art even mentioned making sawdust.
The strength of a modeling program is only as great as it’s ability to predict what you measure when you build from the model. I use Akabak which is a powerful, somewhat cumbersome program but it has real limitations and in bass horn land, should be viewed as a “ball park prediction”.
Often to get closer to what one measures, one has to add some unseen elements to the model. For example, a wood panel flexing at the throat can “look like” an air volume you didn’t account for and while a computer can predict high Q or sharp features in a bass horn, they hardly ever exist in real life as there are losses related to the layout which limit the Q and are not included in the simple model.
Best,
Tom
It would take all the fun out of this stuff I posted drawings (and it would be frowned on at work) and besides, pretty often one or more of you eventually comes close or nails it.
The effective length of a horn is actually different than one measures with a tape measure because at the mouth of a normal symmetric horn, the pressure boundary is an arc where the center sticks out forward of the cabinet.
If I remember correctly, the old “rule of thumb” suggested the mouth bubble for a simple round horn is about 60% of the mouth radius, in other words, at the center of the horn, it protrudes that much forward of the wood. For a cd horn operating where it large enough to have directivity, that boundary is nearly perpendicular to the horn wall, following Keele’s pattern loss thumb rule.
Also, the “horn loading” effect has a “high pass” filter function based on the rate the area expands. For instance for a 30Hz exponential horn, the area cannot expand any faster than doubling about every 2 feet while for 300Hz, the area doubles about every 2.4 inches and so on.
The object of the BC horns is to have the rate of expansion proper so that the end of the bubble is formed by the 60 by 60 inch face dimension and have the inner part of the horn tie acoustically into that external bubble. The horn inside of the cabinet has a larger mouth than the opening in the side of the cabinet and but forcing the output to pass through a restriction puts it in the right place to tie into that bubble and in effect makes the mouth larger and horn path longer.
The restriction has a minor “low pass” effect and adds a little extra acoustic mass, neither of which are disadvantages. Internally, the BC speakers are regular front loaded horns, all of them are large enough where the Tapped design isn’t an advantage AND tapped horns “feel” the external loading less than FLH’s.
Several of you have mentioned an important thing and Art even mentioned making sawdust.
The strength of a modeling program is only as great as it’s ability to predict what you measure when you build from the model. I use Akabak which is a powerful, somewhat cumbersome program but it has real limitations and in bass horn land, should be viewed as a “ball park prediction”.
Often to get closer to what one measures, one has to add some unseen elements to the model. For example, a wood panel flexing at the throat can “look like” an air volume you didn’t account for and while a computer can predict high Q or sharp features in a bass horn, they hardly ever exist in real life as there are losses related to the layout which limit the Q and are not included in the simple model.
Best,
Tom
A 60 x 60 inch cab with a centered 30 inch round sound source has a diffraction profile that looks like this.
If you could sim the cab internals (a very small horn for four 15s with a massively undersized AND mass loaded mouth causing very low sensitivity in the higher frequencies) and properly sum with this diffraction profile you should be right around the right ballpark.
I don't think a 60 x 60 inch cab face dimension is going to provide much more than 1 db boost at 30 hz. I've never played with boundary extensions but I don't think this issue is much more complex than the boundary effect of diffraction.
If you could sim the cab internals (a very small horn for four 15s with a massively undersized AND mass loaded mouth causing very low sensitivity in the higher frequencies) and properly sum with this diffraction profile you should be right around the right ballpark.
I don't think a 60 x 60 inch cab face dimension is going to provide much more than 1 db boost at 30 hz. I've never played with boundary extensions but I don't think this issue is much more complex than the boundary effect of diffraction.
An externally hosted image should be here but it was not working when we last tested it.
Thanks for giving us some more hints. Now that you say it is indeed a FLH as some of us have suspected all along, that doesn't leave too many different ways to fold it. In this particular horn, the mouths of 4 horns are essentially exiting from the same aperture - I assume AkAbak cannot capture the added bass extension that I have heard is provided by stacking bass horns next to each other. At least I have not noticed any additional bass extension when inset the horn radiators side-by-side in AkAbak.
Do you agree with Just a Guy's statement that a TH does not provide any additional efficiency gain compared to any other alignment for the same box volume?
I just showed 4 different designs (the OP's stepped design, a simple tapped horn, a tapped bowtie and a front loaded horn), all almost exactly the same size with almost exactly the same response curve shape and sensitivity. Then I showed a different tapped horn compared to an OD tl (same tl, same driver location, but without the mouth side tap) and they were also the same sensitivity.
That's not convincing enough?
You can do the sims to prove it to yourself or you can choose any tapped horn design you like and I can match it liter for liter, db for db (on average at least, if not a perfectly matched response curve as I've shown), with any other alignment you choose. (Simple ported boxes don't have as many useful impedance peaks to play with so I can only match the low knee region with those.)
Some alignments work better for certain jobs and are easier to design when using standard flares (tapped horn for low tuning in small boxes, flh for high tuning in large boxes) but when you equalize size you equalize output potential. With resonant enclosures Hoffman's law stands regardless of alignment.
Hornresp and Akabak both will show added bass extension when multiple cabs are simulated. It's not much, a couple extra hz and a slightly stronger output at the low knee, but it's there.
1 cab vs 24 stacked cabs, everything else equal except power input. (In other words the response curve shape is applicable but ignore the power level. I added more speakers but didn't change Eg to reflect the different impedance load.)
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Thanks for providing the sims as back up - it's not that I don't believe your sims, and maybe I never kept the volume fixed but when design horns of any kind, they are more efficient than straight boxes. You would think a TH is more efficient than a front loaded bass horn. The energy of half the FLH simply goes into heating the air in the back chamber. Similarly for a sealed cabinet vs ported cabinet, the ported is more efficient in getting higher levels of SPL out because both sides of the driver are used. I think we may be talking about different boxes here because I don't know what your topologies actually look like. I don't know how to read HR input screens. A diagram would be useful but perhaps you are doing an offset TL where sound energy from both sides of driver are used? That will have about same efficiency as a TH because both sides of driver are used. Compare a sealed back FLH and a TH and see which one is more efficient.
The horns will always be more efficient than straight boxes but it's not because of the alignment, it due to the fact that the horn will be much larger. It needs to be larger to get the same low knee as the straight tl.
Doesn't matter, the TH mouth side tap isn't doing much of anything either except filling in a dip that occurs naturally in transmission lines, as I showed. (Actually in both cases both sides of the driver are doing something very significant, half the flh energy is not just wasted heating the sealed box (it's working on annulling the reactance so the flh is tuned right), and the mouth side tap in a tapped horn is not adding 3 db - it's doing something important, but it's not adding any sensitivity or efficiency.)
Sealed isn't a resonant enclosure so it doesn't apply to this conversation. FLH does apply because it's got a huge resonator (flare) hanging off one end of it. Sealed vs ported is not an appropriate comparison, ported vs 4th order bandpass is a much more valid comparison, except in most cases the front chamber of a 4th order bandpass is specifically used as a filter to intentionally introduce losses.
All transmission lines have two output sources, the direct driver radiation and the output from the mouth. The only difference between Graph 1 and Graph 2 in post 88 is I changed TH (tapped horn) to Od (offset driver where the driver is in the same spot as the throat side tap but on the outside wall of the box, like any normal tl). There is no tl where both sides of the driver are not used, otherwise it wouldn't be a tl.
I already did that in post 79, same size box and sensitivity is almost identical. It's an unusual stepped tapped horn but it's a tapped horn. And my matching flh has a high compression ratio but it's valid to show that max potential sensitivity is restricted by size (not alignment), and liter for liter the alignments perform ~ the same. But if you want to choose another tapped horn design I can match it with a FLH and show you again. If the design you choose is quite undersized I might need a high compression ratio and possibly an odd flare shape to match your sensitivity and response curve shape but I can do it. Sensitivity and response curve shape will be the same.
Last edited:
I have played extensively with boundary extensions. I don't understand what you think diffraction has to do with the reinforcement of the horn (or any other radiating) bubble. Max potential sensitivity is restricted by enclosure volume if you ignore boundary potential.
Even without designing an enclosure to specifically take advantage of the boundary, doubling the frontal area with a 45 degree "waveguide" provides 3 dB or more gain when added to the front of the Keystone sub, all the way down to Fb (around 35 Hz):
http://www.diyaudio.com/forums/subwoofers/184986-horn-extender-wave-guide-th.html
Simply doubling or tripling the boundary area with flat "barn doors" also adds a similar amount down to the Fb of the BR cabs tested:
http://www.diyaudio.com/forums/subwoofers/204472-multiple-cabinet-combined-response.html
As Tom pointed out, the sims simply get one in to the ballpark, by comparing measurements to see where the sims don't represent reality, one can then make adjustments to the sim to correspond, similar to what you are doing with Le in another post. Tom designs cabinets on a more frequent basis, and is far better with math than me, so has dialed in the trade secrets, I just use my meat computer and "what looks like it will sound like I remember it should" to adjust to reality.
Between how far off Hornresp, Akabak, and whatever program that you last used that gave more incorrect sims of what an increase in baffle size does compared to the actual measurements I have provided, there is plenty of information that would allow one to start to make corrections from my back of the envelope drawing to something just a bit closer to the BC designs.
The BC are too big for me to make sawdust with to give you more breadcrumbs to follow though, my left wrist is bothering me just building a 35 pound cabinet, a BC would cause 10 dB more pain
. Maybe not, don't have to lift subs 2+ meters high to get valid measurements ;^).But if I built a BC as 4 cabinets sharing the boundary, slowly, maybe only 3 db more pain
... I'm retired, my last sub build, the 2x15" PPSL "Shoehorns" are " good enough" for me now.Hm, haven't tested them with wave guides/barn doors yet, have to give that a go after a few (dozen) other tests are completed....
Art
Last edited:
Diffraction is the effect on response of putting a boundary of a finite size behind a sound source. I'm not sure how you think that's inappropriate since this is EXACTLY what's happening. I think all this talk about bubbles is misleading (or at least leading attention away from the relevant issue which is the boundary loading). Yes, a horn mouth is a bit different than a driver, but barn doors surrounding a mouth(s) is not very unlike a baffle face surrounding a driver. Diffraction (how the sound source interacts with the boundary) is exactly what's going on.
The difference in radiation resistance between a driver on a baffle and a mouth surrounded by a barn door is the reason I said diffraction should get "right around the right ballpark", not "sum the diffraction profile and the Hornresp sim and it will be exactly perfect".
Look at it this way, if you countersunk a horn mouth into a large wall (at the floor/wall junction point), would you set your Ang to 1 pi in Hornresp or start from scratch without considering the functionality of available tools and look for a complicated way to simulate a horn bubble? 1 pi accurately represents the diffraction profile of a large flat wall, a solid 6 db increase across the entire bandwidth. Start chopping that wall into a finite size and you can't use 1 pi to describe the diffraction profile anymore, but you can use different software to simulate it.
I don't want to talk about the waveguide baffle extension, that should be accurately simulated easily enough in Hornresp, let's look at your barn doors.
You have a stack of 8 speakers surrounded by plywood that effectively triples the face area. I don't know how big the cabs are but the total frontal area looks quite a bit bigger than the 60 x 60 inches I created a diffraction profile for, also a different shape of frontal area and different driver location(s). Larger frontal area means more gains to lower frequencies and the other differences create differences in the diffraction profile.
And you specifically mentioned 2 or maybe 3 db of gain to the LF response, which is not much more than I've shown with the 60 x 60 frontal area diffraction sim.
At least look into the functionality of available software, look into what it is actually calculating and correlate the sims with the measurements. THEN look for discrepancies and try to adjust the sim's functionality.
I didn't sim your stack of BR boxes with barn doors so logically the diffraction curve I presented won't be the same as what you measured. Even if I simulated your stack as close as possible it still wouldn't correlate because your stack isn't even close to a round sound source surrounded evenly by a boundary. As you showed, even stacking the cabs in different ways gives different results. You have to be very careful to build exactly what was simulated and set it up exactly the same way the simulation assume it will be if you want more accurate results.
Last edited:
There's only so much I can do with this software and I didn't measure anything to get correct dimensions but I did try to recreate your frontal area a bit more accurately - larger and rectangular. I included your measurements, I believe the top left is your winged vs non winged measurement. You gained a couple of db from 70 to 200 hz gradually tapering off on both ends. A dip a ~ 250 hz and then a bit of gain above the dip all the way up to the chart limit at past 2khz.
I admit the sim doesn't correlate as well when I move the mic out to 30 feet and there are a few key differences between the physical sub stack and what the sim assumes, but look at this and tell me diffraction has nothing to do with boundary extensions.
I admit the sim doesn't correlate as well when I move the mic out to 30 feet and there are a few key differences between the physical sub stack and what the sim assumes, but look at this and tell me diffraction has nothing to do with boundary extensions.
An externally hosted image should be here but it was not working when we last tested it.
In AkAbak you can set the size of the baffle, and set the location of the radiator within the baffle as well as its area. The baffle edge diffraction effects are accounted for so you do not need to use a separate program. In addition to this the location of the nearby walls like floor or ceiling or back wall also play a part in the bass extension due to modifying the PI steradian angle of output. It is not simply 1, or 2 PI but actually calculated based on distance from boundaries. It will be interesting to see what a 60x60 in baffle does to a sub output in 4pi space.
Ok you have convinced me. Sensitivity of a TH is not better than another alignment given same volume.