More than one frequency for constructive backwaves?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am wondering if it is possible to tune an enclosure to have multiple constructive backwaves.

Consider the following situation:

Place a 15" subwoofer with long throw into a sonotube.

IF the back wave is directed through a large port of length equal to 1/2 20hz wavelength + some arbitrary distance X, and the front of the sonotube ahead of the woofer is equal to that distance X, then the two waves will add together constructively.

Could this be done with a system of more than one port set (front and rear) to create constructive addition of backwaves through the entire useable range of the subwoofer?
 
Question along the same line....

We took our prototype out into a field and measured SPL at one meter. We had almost 115 db at 5 hz, and less than 4% thd. We were hoping for over 125 db, however.

How much room gain can we expect in a listening room? We can't get the system into a house or even a garage at the moment... and if our current build @10kW can't hit more than 125db with in-room gain, then we'll call it quits.
 
I am wondering if it is possible to tune an enclosure to have multiple constructive backwaves.

Consider the following situation:

Place a 15" subwoofer with long throw into a sonotube.

IF the back wave is directed through a large port of length equal to 1/2 20hz wavelength + some arbitrary distance X, and the front of the sonotube ahead of the woofer is equal to that distance X, then the two waves will add together constructively.

Could this be done with a system of more than one port set (front and rear) to create constructive addition of backwaves through the entire useable range of the subwoofer?

This is actually a pretty popular question. I wondered the same thing when I was new - why not have a port (or a whole other ported enclosure) for each note, or each octave?

The problem is phase. With a single port you don't get a huge spiked port output that is very narrow bandwidth like WinISD suggests. You get a small spike that has lower amplitude output and spiky garbage resonances for several octaves above the port tuning like 1/4 wave software suggests. With multiple outputs you get vast overlap and consequently massive phase problems so they don't sum the way you would like.

When you put multiple ports in a single ported box it doesn't work like you suggest. See these links, in particular post 2 in the first link and post 4 in the second link. I've never experimented with this but in lieu of that I'll put my money on John Kreskovsky.
http://www.diyaudio.com/forums/car-audio/228026-multiple-ports-all-tuned-different-hz.html
What happens if you have different length ports in a cabinet?

While it is possible to fill in a single hole in response with resonant 1/4 wave enclosures (like a tapped horn or like a Bose Wavecannon which IIRC is the name of what you describe in post 1), trying to do more than fill in a single hole in response gets really tough, maybe impossible.

I was quite surprised to learn that this concept DOES actually work with sealed enclosures. A couple of years back I simulated a Hegeman subwoofer for a guy. That's a sealed box with 4 separate stubs connected at the throat chamber, and all 4 of the stubs are sealed.

An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


Each stub needed quite a bit of stuffing in the first few inches to settle the response spikes down to what you see above. Now compare that to a simple sealed box of the same volume with the same driver as shown below. As you can see the sealed stubs worked but did nothing except trade a squashed impedance spike for a slightly less flat response curve. No advantage for the extra complexity that I can see.

An externally hosted image should be here but it was not working when we last tested it.


So while that does work for sealed boxes I was unable to make it work for unsealed stubs. To be fair I didn't try very hard or for very long and I should probably try again.

BUT unless the goal is MORE than 3 octaves of bandwidth there's no point. That's what horns already do naturally.

Actually, since posting this, I found a military patent for a weaponized subwoofer.... and it used this concept.

So yes, it is possible.... we'll see if I can replicate it on a small scale.

JG

No, for the reasons listed above I don't think it's the same concept. But please post a link to this military patent so I can see it.

Question along the same line....

We took our prototype out into a field and measured SPL at one meter. We had almost 115 db at 5 hz, and less than 4% thd. We were hoping for over 125 db, however.

How much room gain can we expect in a listening room? We can't get the system into a house or even a garage at the moment... and if our current build @10kW can't hit more than 125db with in-room gain, then we'll call it quits.

Depending on the room you can expect a lot or a little. Please show some details of this prototype - either a picture, a drawing or at least a description.
 
Last edited:
Wow thank you for the detailed response.

Here is the military design. http://www.dtic.mil/dtic/tr/fulltext/u2/a399241.pdf

Applying the variable Helmholtz resonator to each port, we can limit the output of harmonics and other annoying distortions. Furthermore, the phase lag can be corrected in this way.

I can't disclose our design in full (because I am working on it with someone else), but I will say that it has the sum effect of a rotary subwoofer mounted in a sonotube (sp), but it has less chuffing, uses less power, and has the potential for greater volume.

Our design that we just tested in a field was simply to test the additive backwave, and our equivalent xmax was 12 meters through a 12" diameter pipe @ 5 hz.
 
Yeah, 5hz @ 125dB with less than 5% THD.

Our design is functional (producing frequencies) to about 600 hz, but phase issues there have not been considered for the upper frequency range.

Our design will be hardest to extend into ultra low frequencies, but the concept of multiple ports should allow high frequencies to be integrated much easier (shorter ports, narrower ports). If we can get a bandwidth from 5 to 20 hz, then it will be really easy to get a bandwidth from 5hz to 100hz
 
Please keep the tone here respectful. I am not an expert, which is why I came to the forum.

Trust me that I have a useful application from a port set with these characteristics.

If you have input WRT the integration of the the variable resonators, your input is welcome. Otherwise there is no need to mock or call me a liar because you don't understand how I will manage to oscillate the air through the ports.

JG
 
Yeah, 5hz @ 125dB with less than 5% THD.

Our design is functional (producing frequencies) to about 600 hz, but phase issues there have not been considered for the upper frequency range.

Our design will be hardest to extend into ultra low frequencies, but the concept of multiple ports should allow high frequencies to be integrated much easier (shorter ports, narrower ports). If we can get a bandwidth from 5 to 20 hz, then it will be really easy to get a bandwidth from 5hz to 100hz

Well, simulators make it pretty easy to test the acoustical possibilities WITH T/S PARAMETERS very quickly. You can test this type of thing with open stubs originating in the same spot:

An externally hosted image should be here but it was not working when we last tested it.


or you can offset the stubs like this (you can also use a combination of open and closed ended stubs):

An externally hosted image should be here but it was not working when we last tested it.


or you can simulate straightforward front loaded horn or horn shaped ports,

or you can use any combination of any of the above. Transmission line body of any size and shape with branched ports of any size and shape at any node. The problem is getting anything resembling a flat usable response for more than 3 octaves. As you state, if it was simple it would have been done. I really don't think we are going to change history here today.

But inputting t/s parameters into a sim is different than simulating a modulated pressurized airflow, I'm not even sure how to start with that.

As far as I know, the best that can be done is 3 octaves of flat(ish) bandwidth. The more acoustic gain you push out the narrower the bandwidth becomes with traditional designs like horns UNLESS you can increase the enclosure volume exponentially, and even then there's a 3 octave limit when using only controlled resonances as output devices (in other words not using the cone - which you don't have - to extend bandwidth). When you need more than 3 octaves you use more than 1 subwoofer, one covers the lower notes, the other covers the higher notes OR you use the direct cone output to extend the bandwidth.

Coming up with something that produces more than 3 octaves is hard enough, not to mention that this device needs to have the same frequency response AND phase response as the noise you are trying to cancel, otherwise it may cancel at some frequencies and ADD at other frequencies.

I'm not even going to ask why this isn't being done by a trained and qualified professional, I assume you can't even step foot into a nuclear power plant without some type of credentials. And while I'm not going to ask, I will freely state that I'm hesitant and very reluctant to even discuss a device that will be used in such an environment. I am reading over the patent you linked to and I am interested in the theory but really want no part (or responsibility) in helping to design something for a nuclear power plant.
 
Last edited:
Thanks for the link.

I noticed a few things:

"At any rate, a blower powered sub is not going to satisfy you in terms of fidelity, unless you do it the way the Thigpen does, or find some other way to instantly control the air flow of a high volume blower.
In which case, since you are not looking for output below 12 Hz, conventional higgh Xmax speakers will be far cheaper, whether one (or more) in a TH like Josh Ricci's Gjallerhorn, or several in an infinite baffle."

This indicates that the results here are high in distortion.

Also:
"A problem being outdoors was there is no room gain and to reach the target at 3Hz from 2 meters required a fan source that had about 50 cu mtrs sec and needed 12 X 5hp fans to produce. The only solution to the back wave I could think of was simply to place the house at one exit and move the opposite phase far enough away. To allow the flow to expand uniformly instead of forming a free jet within the exit, a tapered section following the graph I found in a Thermodynamics text book dealing with jet and rocket engines.
To be clear, if this had been coupled into the old test house at one end, I am absolutely sure we could have brought it down with very little effort, we did damage even from outside just fooling around."

My theoretical calculations indicate that for 5hz at 125dB my design only needs around 6 horsepower, and testing got us to 115dB in free space.

Danley also did not manage the backwave in any way. That is certainly room for improvement.

Furthermore, unless I greatly misunderstand his design, the venturi drive does work, but not at the efficiency that my design offers. With my design (already tested okay at 5hz), only 3 cubic meters per second Peak to Peak is needed. Not 50....


As I see it, my design is a better way to occilate air than either the rotary subwoofer or the Danley design; more laminar flow, better efficiency, and higher headroom. Last of all, we are working to make our design smaller... looks to fit in about the space of a small refrigerator.

Thank you for your constructive response weltersys.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.