This idea has been discussed in this forum before but didn't really go anywhere. The Graham Holliman Infrabass Speaker uses a new loading method to lower the natural resonance of a sub to the 5 - 25 hz range, allowing very efficient output in that range.
When I say this is a new loading method, I only mean that it is relatively unexplored, as the idea has been around for at least twenty years. Usually I like to do a fair bit of research before I build but it is hard to find any info on this design.
Below are links to the patent and the old thread from this forum. Also, the text and some pictures of an article/ad by a company selling drive units and crossovers for this design, including notes from a guy that made one. Some of the details are sketchy and sometimes contradictory between these sources but it's all I can find. Most of this stuff is quite old.
http://v3.espacenet.com/textdes?IDX=GB2037534&CY=gb&LG=en&DB=EPODOC&QPN=GB2037534
http://www.hi-fi.ro/fhifi/download.php?id=6317&sid=b0f4fb470065703326c3b3fff7bd6562
http://www.diyaudio.com/forums/showthread.php?s=&threadid=3238&highlight=
I am requesting more information, if anyone has any idea where to get some. I have been contemplating this design for weeks now and have some ideas but I don't want to get into it unless there is sufficient interest for a decent conversation. In my opinion this is a much better approach than using multiple 18 inch drivers in huge boxes tuned well out of their comfort zone trying to get below 10 hz.
Comments, opinions, ideas? Wanna talk about it?
When I say this is a new loading method, I only mean that it is relatively unexplored, as the idea has been around for at least twenty years. Usually I like to do a fair bit of research before I build but it is hard to find any info on this design.
Below are links to the patent and the old thread from this forum. Also, the text and some pictures of an article/ad by a company selling drive units and crossovers for this design, including notes from a guy that made one. Some of the details are sketchy and sometimes contradictory between these sources but it's all I can find. Most of this stuff is quite old.
http://v3.espacenet.com/textdes?IDX=GB2037534&CY=gb&LG=en&DB=EPODOC&QPN=GB2037534
http://www.hi-fi.ro/fhifi/download.php?id=6317&sid=b0f4fb470065703326c3b3fff7bd6562
http://www.diyaudio.com/forums/showthread.php?s=&threadid=3238&highlight=
I am requesting more information, if anyone has any idea where to get some. I have been contemplating this design for weeks now and have some ideas but I don't want to get into it unless there is sufficient interest for a decent conversation. In my opinion this is a much better approach than using multiple 18 inch drivers in huge boxes tuned well out of their comfort zone trying to get below 10 hz.
Comments, opinions, ideas? Wanna talk about it?
have you considered...
a dual driver "W" frame dipole? This is a velocity device and as far as I know it is the smallest volume sub box that you can make. If you reduce the vent area of the two drivers to between 1/4 to 1/3 the driver area then this increases the radiation resistance that the drivers will see and will decrease your driver Fs typically by 7 to 10 hertz (sometimes more). So if you were to start out with a driver (15 - 18 inch) with an Fs of say 18 Hz. (not all that uncommon these days for car subs) and the driver had lots of xmax and did not generate much rear noise (turbulence) that would get you right where you want to be. The key factors are driver noise and xmax. With dipoles you will need plenty of travel to make up your dipole roll off and you will probably require multiple units no matter what cabinet style you choose to make such low bass have impact.
You might try to see if you could get Tom Danley to comment as he is as far as I know the King of infrasonic bass reproduction.
You can take a look at this thread to see what kind of driver values you will want. The two options are to go with a driver Qts of about 0.35 and actively EQ the system or to use a driver with a Qts of about 0.5 to obtain more output at roll off without the need for EQ. Most of the really big hi xmax car drivers today would be about right to get the job done without EQ I should think.
Good luck and watch that you don't hurt yourself as such frequencies at level can do physical structural damage to your skeleton not to mention rearranging your intestines (seriously) and there is the house and neighbours to be concerned with. Tom has some great stories of this kind of bass. Best regards Moray James.
a dual driver "W" frame dipole? This is a velocity device and as far as I know it is the smallest volume sub box that you can make. If you reduce the vent area of the two drivers to between 1/4 to 1/3 the driver area then this increases the radiation resistance that the drivers will see and will decrease your driver Fs typically by 7 to 10 hertz (sometimes more). So if you were to start out with a driver (15 - 18 inch) with an Fs of say 18 Hz. (not all that uncommon these days for car subs) and the driver had lots of xmax and did not generate much rear noise (turbulence) that would get you right where you want to be. The key factors are driver noise and xmax. With dipoles you will need plenty of travel to make up your dipole roll off and you will probably require multiple units no matter what cabinet style you choose to make such low bass have impact.
You might try to see if you could get Tom Danley to comment as he is as far as I know the King of infrasonic bass reproduction.
You can take a look at this thread to see what kind of driver values you will want. The two options are to go with a driver Qts of about 0.35 and actively EQ the system or to use a driver with a Qts of about 0.5 to obtain more output at roll off without the need for EQ. Most of the really big hi xmax car drivers today would be about right to get the job done without EQ I should think.
Good luck and watch that you don't hurt yourself as such frequencies at level can do physical structural damage to your skeleton not to mention rearranging your intestines (seriously) and there is the house and neighbours to be concerned with. Tom has some great stories of this kind of bass. Best regards Moray James.
have a look at djk's assesment of the acoustic cannon...
Looking at the cross section drawing there are stricking similarities between the infra bass unit and the discription Dennis gives for a "half square antenna" even though the infra bass unit uses some horn loading of the initial section of the line. I hope that this is of some interest".
Quote of djk below.
I use a spacing of 1/4 the total tube length -- as does the Bose Acoustic Wave Music System"
The patent says:
"The separation between openings 28 and 31 is of the order of half the length of the shorter tube between the front of driver 22 and opening 31."
which is 1/8th the total tube length. In the SUMMARY OF THE INVENTION they claim:
"A preferred separation is within the range of one-eighth to one times the length of the path for pressure waves "
This is simply being expedient on their part in making the claim as broad as possible. It covers the Acoustic Wave at 1/8 exit spacing and the Sound Cannon at 1W exit spacing.
The optimum spacing for the smoothest fequency response and the highest broadband gain is 1/2W exit spacing.
"For those of us too dumb to know what the hell you're talking about, could you elaborate?;-)"
I ran a Yahoo search on 'half square antenna' (what a radical idea!) and came up with:
http://www.qsl.net/ka1ddb/20meterhalfsquaredwg.jpg
http://www.cebik.com/hs.html
http://kc5jk.tripod.com//sitebuilde...half-sq-ant.gif
"To get DX angle radiation and gain without the tower and the Yagi, use a HalfSquare. What's a HalfSquare? A HalfSquare is a simple, light, unobtrusive and inexpensive DX gain antenna that you can hang almost anywhere you can put a dipole.
How is it made? Start with a wavelength of wire supported from its ends, and feed it at a point one quarter wavelength from an end. In this form the antenna is called a long wire. It has a decibel or so of gain off its ends and its radiation angle is a little lower than a dipole or Yagi at the same height, but it's no killer DX antenna yet.
Move the support insulators in a quarter wave from each end and let the tails hang down. Bingo! What you have now is a HalfSquare-two quarter wave verticals separated by a half wave horizontal phasing section. By this simple trick you have pushed the takeoff angle down into the serious DX range and you suddenly have nearly an S-unit of broadside gain to boot. As a bonus the feed-point impedance drops to 50 +."
Notice the author mentions that the performance really increases when you have the exits 1/2W apart?
As noted in the Bose patent virtually any spacing from 1/8W (boom box) to 1W (straight cannon) will work, it will just require more equalization to smooth out, and be less efficent.
Assuming a loss-less tube: "providing a nominal 6 db increase in sensitivity compared to the same driver in the infinite baffle. "
Looking at the cross section drawing there are stricking similarities between the infra bass unit and the discription Dennis gives for a "half square antenna" even though the infra bass unit uses some horn loading of the initial section of the line. I hope that this is of some interest".
Quote of djk below.
I use a spacing of 1/4 the total tube length -- as does the Bose Acoustic Wave Music System"
The patent says:
"The separation between openings 28 and 31 is of the order of half the length of the shorter tube between the front of driver 22 and opening 31."
which is 1/8th the total tube length. In the SUMMARY OF THE INVENTION they claim:
"A preferred separation is within the range of one-eighth to one times the length of the path for pressure waves "
This is simply being expedient on their part in making the claim as broad as possible. It covers the Acoustic Wave at 1/8 exit spacing and the Sound Cannon at 1W exit spacing.
The optimum spacing for the smoothest fequency response and the highest broadband gain is 1/2W exit spacing.
"For those of us too dumb to know what the hell you're talking about, could you elaborate?;-)"
I ran a Yahoo search on 'half square antenna' (what a radical idea!) and came up with:
http://www.qsl.net/ka1ddb/20meterhalfsquaredwg.jpg
http://www.cebik.com/hs.html
http://kc5jk.tripod.com//sitebuilde...half-sq-ant.gif
"To get DX angle radiation and gain without the tower and the Yagi, use a HalfSquare. What's a HalfSquare? A HalfSquare is a simple, light, unobtrusive and inexpensive DX gain antenna that you can hang almost anywhere you can put a dipole.
How is it made? Start with a wavelength of wire supported from its ends, and feed it at a point one quarter wavelength from an end. In this form the antenna is called a long wire. It has a decibel or so of gain off its ends and its radiation angle is a little lower than a dipole or Yagi at the same height, but it's no killer DX antenna yet.
Move the support insulators in a quarter wave from each end and let the tails hang down. Bingo! What you have now is a HalfSquare-two quarter wave verticals separated by a half wave horizontal phasing section. By this simple trick you have pushed the takeoff angle down into the serious DX range and you suddenly have nearly an S-unit of broadside gain to boot. As a bonus the feed-point impedance drops to 50 +."
Notice the author mentions that the performance really increases when you have the exits 1/2W apart?
As noted in the Bose patent virtually any spacing from 1/8W (boom box) to 1W (straight cannon) will work, it will just require more equalization to smooth out, and be less efficent.
Assuming a loss-less tube: "providing a nominal 6 db increase in sensitivity compared to the same driver in the infinite baffle. "
Could someone give the patent number?
I'd like to experiment with this design and smaller drivers. I could probably load the whole thing into Akabak and simulate it that way. I wonder how the design would work with woofers smaller than 8". I suspect there's quite a bit of mass loading going on that pushes the driver's Fs lower. But what I really would like to explore is if adding a serious taper to the long vent (like 1:3) could introduce enough mechanical damping to make one of these things listenable. To the naked eye, it looks like the driver is simply loaded along the length of the tuned port, but thinking from an acoustic network perspective, it looks like the driver is actually connected to two separate air masses. Then the tube behaves as a single air mass for the infrasonic effect where the whole mass is just bouncing on the volume of air stored in the large chamber.
I'd like to experiment with this design and smaller drivers. I could probably load the whole thing into Akabak and simulate it that way. I wonder how the design would work with woofers smaller than 8". I suspect there's quite a bit of mass loading going on that pushes the driver's Fs lower. But what I really would like to explore is if adding a serious taper to the long vent (like 1:3) could introduce enough mechanical damping to make one of these things listenable. To the naked eye, it looks like the driver is simply loaded along the length of the tuned port, but thinking from an acoustic network perspective, it looks like the driver is actually connected to two separate air masses. Then the tube behaves as a single air mass for the infrasonic effect where the whole mass is just bouncing on the volume of air stored in the large chamber.
Regarding your first post, I am familiar with W dipoles, but only through research. They are a good idea but in the range of 5 - 25 hz high output is key. W dipoles are not known for sensitivity.
Also there are some claimed advantages to the Holliman box. In addition to great sensitivity, it claims low distortion and no practical lower limit. The design the inventor had the most success with only went down to about 7 hz but he claims it can be tuned as low as you like.
I like the idea of using a cheap 10 inch driver in a reasonable sized box to cover the lows, although I am considering the 15 inch version. I think there is lots of potential here.
About the second post - sorry, I didn't understand any of that.
BAM - the patent number is GB2037534, from the first link I provided. Someone in the old thread tried to model it in Akabak before. Also linked to in my first post. Anyway, I think there are reasons other than mass loading why the fs is lowered. I think that decoupling the driver from the resonant chamber is what accomplishes that.
Also there are some claimed advantages to the Holliman box. In addition to great sensitivity, it claims low distortion and no practical lower limit. The design the inventor had the most success with only went down to about 7 hz but he claims it can be tuned as low as you like.
I like the idea of using a cheap 10 inch driver in a reasonable sized box to cover the lows, although I am considering the 15 inch version. I think there is lots of potential here.
About the second post - sorry, I didn't understand any of that.
BAM - the patent number is GB2037534, from the first link I provided. Someone in the old thread tried to model it in Akabak before. Also linked to in my first post. Anyway, I think there are reasons other than mass loading why the fs is lowered. I think that decoupling the driver from the resonant chamber is what accomplishes that.
I'm into the idea of tying the speakers roll off into the rooms gain.
My tempests were pretty flat 10 - 30Hz without any EQ. If I was aiming for the 5Hz - 20Hz range right now, and had the space, I'd be looking at probably 24 tempest drivers in an IB or in Linkwitz Transformed boxes. They'd basically take up both front corners in the listening room, 12 per side. I'd build the manifold / box out of the corner so I could load them ' push push' .
Naturally you need the right room to do this, but it doesn't work out too expensive.
I used tempests as an example, but any decent 15" / 18" would do.
btw, I'd still cross them to my labhorns at around 25Hz - The labs kill my tempests for sound quality.
Cheers.
Rob.
My tempests were pretty flat 10 - 30Hz without any EQ. If I was aiming for the 5Hz - 20Hz range right now, and had the space, I'd be looking at probably 24 tempest drivers in an IB or in Linkwitz Transformed boxes. They'd basically take up both front corners in the listening room, 12 per side. I'd build the manifold / box out of the corner so I could load them ' push push' .
Naturally you need the right room to do this, but it doesn't work out too expensive.
I used tempests as an example, but any decent 15" / 18" would do.
btw, I'd still cross them to my labhorns at around 25Hz - The labs kill my tempests for sound quality.
Cheers.
Rob.
Anyway, since there is interest already, I'll get it started with some observations. I don't think any program (although I am not familiar with Akabak) can model this, as it is an unused loading method. It uses the "blowing across the mouth of a bottle" resonance and the driver is decoupled from the resonant enclosure.
There seem to be 3 main factors in this design.
1. The driver. I have no info at all about what type of driver is best, although lots of xmax seems preferable.
2. The resonant enclosure. This is the box at the heart of this plan, just a helmholtz resonator, a sealed box with a hole in it. Modelling each of the versions with a helmholtz resonator calculator shows tuning ranges from about 40 hz (15 inch version) to about 70 hz (10 inch version). So it would seem that the exact tuning of the helmholtz resonator is not incredibly important. On the other hand, the length of the port makes huge differences in the resonator's bandwidth and q, if changed.
3. The tapered ports. These start at the hole (X in the plans, the hole in the resonator) and extend to the sides, and then downward to the bottom. At first glance I thought these were two separate parts, but they form a steadily increasing port almost 4 feet long (15 inch version). The patent info says these ports should be as long as possible with a minimum of taper. I don't necessarily think that is sound advice.
More on all this later...
There seem to be 3 main factors in this design.
1. The driver. I have no info at all about what type of driver is best, although lots of xmax seems preferable.
2. The resonant enclosure. This is the box at the heart of this plan, just a helmholtz resonator, a sealed box with a hole in it. Modelling each of the versions with a helmholtz resonator calculator shows tuning ranges from about 40 hz (15 inch version) to about 70 hz (10 inch version). So it would seem that the exact tuning of the helmholtz resonator is not incredibly important. On the other hand, the length of the port makes huge differences in the resonator's bandwidth and q, if changed.
3. The tapered ports. These start at the hole (X in the plans, the hole in the resonator) and extend to the sides, and then downward to the bottom. At first glance I thought these were two separate parts, but they form a steadily increasing port almost 4 feet long (15 inch version). The patent info says these ports should be as long as possible with a minimum of taper. I don't necessarily think that is sound advice.
More on all this later...
The model that was done, if I remember correctly, showed very poor performance below 50 hz, strong above that. In reality, if one is built and measured, it should be exactly the opposite. I don't think these simulations are taking into account this type of loading method and I don't think they can possibly be accurate. Again, I am not familiar with the software but I don't trust the previous results. They go against everything else that has been published about this box.
I've had a look through the links, the patent and the word file, other thread and its comments. It seems hard to believe that this could work well as a part of an accurate hifi system.
I note that its seems unlikely that the driver excursion is damped, hence excursion will be out of control. Be careful. I recall hearing of guys with the AE speakers AV12 having excursions so high when operating undamped that the cone hit the back of the terminals and dented itself! That would mean an excursion of more than 60mm!
Still, I'm curious to see this sub. I doubt it would work in my room since I have no real gain to speak of. I suspect that the rooms that are suitable are those with very solid boundaries with major room mode issues, hence boomy bass is also a problem, not necessarily due to the alignment itself.
I can see how this could be frustrating. The output you will get down that low will most often be very little, and usually unintended. My system runs flat down to the low 20s and I have an RTA in my system all the time. I find there is nearly always response down that low that is worth reproducing. Below 20 Hz there isn't much. Watching TV I sometimes find that the recording of some commercials have quite loud and unintended boomy sounds that come from the person's breath going into the mic.
I note that its seems unlikely that the driver excursion is damped, hence excursion will be out of control. Be careful. I recall hearing of guys with the AE speakers AV12 having excursions so high when operating undamped that the cone hit the back of the terminals and dented itself! That would mean an excursion of more than 60mm!
Still, I'm curious to see this sub. I doubt it would work in my room since I have no real gain to speak of. I suspect that the rooms that are suitable are those with very solid boundaries with major room mode issues, hence boomy bass is also a problem, not necessarily due to the alignment itself.
I can see how this could be frustrating. The output you will get down that low will most often be very little, and usually unintended. My system runs flat down to the low 20s and I have an RTA in my system all the time. I find there is nearly always response down that low that is worth reproducing. Below 20 Hz there isn't much. Watching TV I sometimes find that the recording of some commercials have quite loud and unintended boomy sounds that come from the person's breath going into the mic.
The patent specifically states that the effects of the room are taken care of by design. Even if the design doesn't take care of it, I have some ideas to mostly take the room out of the equation.
But what I was hoping to discuss is the effect each part of this box has on the overall design. For example, anyone want to recommend a driver? Suggest a maximum length and expansion for the ports? Figure out the exact q and bandwidth of the helmholtz resonator?
As I stated before, lots of xmax may be good here, and since the driver does not seem very well damped by the enclosure, maybe a low qts driver?
I do not think the tapered ports should be any longer than about 4 feet. Shorter than this and they act as a wideband "exponential port", longer than that they should start acting like a horn and limit output based on it's horn characteristics. Regarding expansion, the patent states it should have a minimum of taper, but I think it could use a bit more taper, a bit more mouth area.
Using a helmholtz resonator calculator, I can find the tuning of the resonator part of this design, but not exactly. Because of the 45 degree chamfer on the port, I don't know how to guess at its length. Is the length of port the thickness of the wood, or is it longer like a flared port, or is it shorter, just a millimeter or so around the perimeter? I have tried different values for port length from 1 mil to 4 cm and while the tuning of the resonator does not change much, the bandwidth and q change DRASTICALLY. I think the porthole in the helmholtz resonator is a huge key in the tuning of this box, but I am unsure of it's current tuning because of the unknown port length of chamfered port X.
I've got a couple more little ideas still, but still looking for some input on the subject. More later...
But what I was hoping to discuss is the effect each part of this box has on the overall design. For example, anyone want to recommend a driver? Suggest a maximum length and expansion for the ports? Figure out the exact q and bandwidth of the helmholtz resonator?
As I stated before, lots of xmax may be good here, and since the driver does not seem very well damped by the enclosure, maybe a low qts driver?
I do not think the tapered ports should be any longer than about 4 feet. Shorter than this and they act as a wideband "exponential port", longer than that they should start acting like a horn and limit output based on it's horn characteristics. Regarding expansion, the patent states it should have a minimum of taper, but I think it could use a bit more taper, a bit more mouth area.
Using a helmholtz resonator calculator, I can find the tuning of the resonator part of this design, but not exactly. Because of the 45 degree chamfer on the port, I don't know how to guess at its length. Is the length of port the thickness of the wood, or is it longer like a flared port, or is it shorter, just a millimeter or so around the perimeter? I have tried different values for port length from 1 mil to 4 cm and while the tuning of the resonator does not change much, the bandwidth and q change DRASTICALLY. I think the porthole in the helmholtz resonator is a huge key in the tuning of this box, but I am unsure of it's current tuning because of the unknown port length of chamfered port X.
I've got a couple more little ideas still, but still looking for some input on the subject. More later...
Back to the antenna theory...
The reference made by Dennis Kleitsch (djk) in the acousit cannon thread would seem to apply (to some degree) in the infra bass design. Looking at the side view drawing and putting aside for the moment the rear cavity of the driver. You can see the distinct "U" shape duct that the driver fires into (all be it at a different physical position on the line). This is very similar in proportions to the 1/4 wave - half wave - 1/4 wave configuration prescribed by the antenna theory. I dont think that this is a coincident but may have been by chance rather than by design on Holliman's part or simply a trial and error discovery. The "half square antenna" sets up four distinct resonant nodes which give the antenna its gain over a wide band. This applies directly to the acoustic action of the design providing the resonant acoustic gain of the "U" shaped pipe section in the Holliman design (or would seen to to me). Please note I have no experience designing antennas and I am only expressing my limited understanding of the antenna theory as explained by Dennis.
Perhaps Dennis could provide some clairity here if he is following this thread. It seems to me that there are distinct similarities between the designs. The Holliman design is a lot smaller and perhaps represents a scaled version of the antenna theory at 1/8 - 1/4 - 1/8 and gains some additional resonant output via the air in the drivers back cavity? Regards Moray James.
The reference made by Dennis Kleitsch (djk) in the acousit cannon thread would seem to apply (to some degree) in the infra bass design. Looking at the side view drawing and putting aside for the moment the rear cavity of the driver. You can see the distinct "U" shape duct that the driver fires into (all be it at a different physical position on the line). This is very similar in proportions to the 1/4 wave - half wave - 1/4 wave configuration prescribed by the antenna theory. I dont think that this is a coincident but may have been by chance rather than by design on Holliman's part or simply a trial and error discovery. The "half square antenna" sets up four distinct resonant nodes which give the antenna its gain over a wide band. This applies directly to the acoustic action of the design providing the resonant acoustic gain of the "U" shaped pipe section in the Holliman design (or would seen to to me). Please note I have no experience designing antennas and I am only expressing my limited understanding of the antenna theory as explained by Dennis.
Perhaps Dennis could provide some clairity here if he is following this thread. It seems to me that there are distinct similarities between the designs. The Holliman design is a lot smaller and perhaps represents a scaled version of the antenna theory at 1/8 - 1/4 - 1/8 and gains some additional resonant output via the air in the drivers back cavity? Regards Moray James.
I don't believe it is possible with any box design to take care of room effects for all rooms. You could only achieve this for one room, and for rooms that are acoustically similar, and even then this would only be true below all room modes. For a room suitable for this subwoofer, any other subwoofers or woofers are likely to have problems because modes are likely to be severe, and require quite a lot of parametric eq and/or dipoles.
Below 25 Hz I don't expect any room modes, but you will still need to deal with room gain if present. I'd expect that eq is likely to help. I'd also expect that for this to work well, you would want the SPL level to be increased. If you have your system totally flat down to 5 Hz, I doubt you will be impressed - you may find that the extra depth does very little without being increased in level.
I hear comments regarding this type of subwoofer like "too much is dangerous," but that problem is easy to fix - turn the volume down. I think what is more challenging is achieving the kind of extension and SPL with accuracy where this will even get close to being an issue.
Looking at the design, I wonder what exactly is going on with the design. It's like a back loaded horn in that it lacks a sealed chamber to damp the cone, yet it's unclear whether the lines are working as transmission lines, or vents. I suspect more likely the latter, or their bandwidth would be too high. So perhaps this is a bandpass device connected to a helmholz resonator, essentially like blowing on the mouth of a bottle, then capturing the output and venting it to further increase output.
It could also be like a dipole. A few things could be happening. The output could be increased enough for the rear output of the driver to not be an issue in terms of cancellation. Or the phase could be altered such that the two outputs sum together.
I wonder if this design could be extended further, and something could be done with the undamped driver. Perhaps a sealed box, or a vented box to extend output up to 40 Hz. Possibly a bandpass. Now if this were done, and say the efficiency is not as much, you could then have a subwoofer which could both cover the usual 20 - 40 Hz range to integrate with typical mains, but then have the higher output sub 20 Hz output to create extra depth. Probably this would require more power to overcome the damping of the box. The main problem I see would be if the phase gets messed up, and the polarity needed to be reversed.
Every room will have problems, of course, but that is a whole other battle I don't really want to get into yet.
Just for reference, if you can equate the Holliman box to a ported, or vented box, then the link below is the "sealed" equivalent. It uses the same "blowing across a bottle" effect to achieve the required low frequencies to give this horn a lower cutoff and stronger bottom end than if it were front or back loaded. But in this case the driver is enclosed by the sealed chamber and the front radiation of the driver is blown across the throat of the horn, through the k-slot shaped port X (circular in the holliman plans) and into the room instead of Holliman's resonator. The sub in the link can be used full range, parallel with high efficiency mids.
http://www.decware.com/ImperialSO.htm
Just for reference, if you can equate the Holliman box to a ported, or vented box, then the link below is the "sealed" equivalent. It uses the same "blowing across a bottle" effect to achieve the required low frequencies to give this horn a lower cutoff and stronger bottom end than if it were front or back loaded. But in this case the driver is enclosed by the sealed chamber and the front radiation of the driver is blown across the throat of the horn, through the k-slot shaped port X (circular in the holliman plans) and into the room instead of Holliman's resonator. The sub in the link can be used full range, parallel with high efficiency mids.
http://www.decware.com/ImperialSO.htm
some observations...
"Quote"#1 : The patent info says these ports should be as long as possible with a minimum of taper. I don't necessarily think that is sound advice End Quote. Why?
"Quote"#2 : I do not think the tapered ports should be any longer than about 4 feet. Shorter than this and they act as a wideband "exponential port", longer than that they should start acting like a horn and limit output based on it's horn characteristics. Regarding expansion, the patent states it should have a minimum of taper, but I think it could use a bit more taper, a bit more mouth area. End Quote. Is this not a contradiction?
If you add more taper or flair to the port it will behave as a horn rather than a simple resonant duct.
It would seen to me that the duct is required to have peak resonant modes (as many as possible) which are as high as possible and over as wide a band as possible so as to achieve gain in the output at the frequency band of intrest . Those resonant peaks would then seem to be bounced off of the air in the cavity like a person bouncing on a trampoline to further boost the resonant peaks (additional gain). So as I see it the duct is required to be a resonant line where the line resonances are driven by the loudspeaker. I think that Dennis showed in the Acoustic Cannon thread tha the Half Square Antenna theory nicely provides the maximum number of resonant peaks in the line duct) over the widest range. The "U" shaped duct is almost exactly what I see when I look at the side view of the Infra bass drawing. So I think that what is required is an expert in Helmholtz resonators to crunch some numbers to make everything work in concert. Shorter duct lengths ie. 1/4 or 1/8 scale will probably work to help keep things more domestic.
In any event to make something like this work the resonances all have to be additive and nicely lined up. Can not see how random volumes and lengths will ever achieve your goal of high gain resonant output otherwise.
"Quote"#1 : The patent info says these ports should be as long as possible with a minimum of taper. I don't necessarily think that is sound advice End Quote. Why?
"Quote"#2 : I do not think the tapered ports should be any longer than about 4 feet. Shorter than this and they act as a wideband "exponential port", longer than that they should start acting like a horn and limit output based on it's horn characteristics. Regarding expansion, the patent states it should have a minimum of taper, but I think it could use a bit more taper, a bit more mouth area. End Quote. Is this not a contradiction?
If you add more taper or flair to the port it will behave as a horn rather than a simple resonant duct.
It would seen to me that the duct is required to have peak resonant modes (as many as possible) which are as high as possible and over as wide a band as possible so as to achieve gain in the output at the frequency band of intrest . Those resonant peaks would then seem to be bounced off of the air in the cavity like a person bouncing on a trampoline to further boost the resonant peaks (additional gain). So as I see it the duct is required to be a resonant line where the line resonances are driven by the loudspeaker. I think that Dennis showed in the Acoustic Cannon thread tha the Half Square Antenna theory nicely provides the maximum number of resonant peaks in the line duct) over the widest range. The "U" shaped duct is almost exactly what I see when I look at the side view of the Infra bass drawing. So I think that what is required is an expert in Helmholtz resonators to crunch some numbers to make everything work in concert. Shorter duct lengths ie. 1/4 or 1/8 scale will probably work to help keep things more domestic.
In any event to make something like this work the resonances all have to be additive and nicely lined up. Can not see how random volumes and lengths will ever achieve your goal of high gain resonant output otherwise.
Random values certainly won't work, that's why I wanted to talk about all this. As far as the length of the ports, the plans call for ports about 38 inches long. I'm just saying that if I were to make them any longer at all, I certainly would not go past 4 feet length. Looking at the taper of the ports it does not seem to expand as quickly as many of the other similar sized ports I have seen, but at this time I would be hesitant to make to many changes at all to the ports. The word "taper" in this case is used to describe the expansion of the ports, so more taper does in fact mean more mouth area, no contradiction there.
Moray, I am sorry, but I don't think the shape the ports take has anything at all to do with Half Square Antenna theory. I may be wrong, as I still know nothing about the subject but these ports are just ports regardless of shape.
Anyway, I have thought endlessly about this design and I feel I have it figured out to some extent. Definitely enough to build one and not be scared about the results. Unfortunately a proper description will not fit in a forum post, I wrote my description of the how and why down for someone that was interested and it was at least a couple of typed pages and still quite incomplete, so if you would like to continue the conversation through some other way than forum posts let me know.
Moray, I am sorry, but I don't think the shape the ports take has anything at all to do with Half Square Antenna theory. I may be wrong, as I still know nothing about the subject but these ports are just ports regardless of shape.
Anyway, I have thought endlessly about this design and I feel I have it figured out to some extent. Definitely enough to build one and not be scared about the results. Unfortunately a proper description will not fit in a forum post, I wrote my description of the how and why down for someone that was interested and it was at least a couple of typed pages and still quite incomplete, so if you would like to continue the conversation through some other way than forum posts let me know.
Well how do you see it?
I see a pipe about four feet long (lets ignor the small flair for this discussion) which will act as a resonant pipe with multiple predetermined resonant modes. That pipe is coupled at its centre point to a resonant cavity at which point there is also a driver positioned to excite (input energy) the system. Once the speaker does its job of playing some low notes (remember that the driver xmax and size are not major factors according to the patent) we are then depending upon the cabinet to amplify the notes introduced by the driver. The only way I see that happening is through a series of compound resonances which are built into the design of the cabinet (ie. the pipe and the cavity). The pipe if designed properly will have a series of broadly positioned modes available which is desirable. The large cavity will have a resonance (here I am not sure if there will be more than one [mass volume of intrapped air] {box dimensions may also come into play, width height and depth} which will be of a higher Q than those in the pipe due to the cavity volume. The pipe resonances will achieve a certqin amount of gain and this is then boosted or added to the gain provided by the cavity resonance. I will further guess that the cavity (due to its physical dimensions) probably has a fairly wide band resonance centred upon those of the pipe. The sum of both these two resonant systems provides the total gain to the notes played by the driver.
I cannot see any other means of gaining efficiency. If this is not how you view the system to work then I would be interested in knowing just how you do think it works.
I asked Dennis if he would provide some input regarding the Half Square antenna theory as he makes his living in RF design I believe.
I see a pipe about four feet long (lets ignor the small flair for this discussion) which will act as a resonant pipe with multiple predetermined resonant modes. That pipe is coupled at its centre point to a resonant cavity at which point there is also a driver positioned to excite (input energy) the system. Once the speaker does its job of playing some low notes (remember that the driver xmax and size are not major factors according to the patent) we are then depending upon the cabinet to amplify the notes introduced by the driver. The only way I see that happening is through a series of compound resonances which are built into the design of the cabinet (ie. the pipe and the cavity). The pipe if designed properly will have a series of broadly positioned modes available which is desirable. The large cavity will have a resonance (here I am not sure if there will be more than one [mass volume of intrapped air] {box dimensions may also come into play, width height and depth} which will be of a higher Q than those in the pipe due to the cavity volume. The pipe resonances will achieve a certqin amount of gain and this is then boosted or added to the gain provided by the cavity resonance. I will further guess that the cavity (due to its physical dimensions) probably has a fairly wide band resonance centred upon those of the pipe. The sum of both these two resonant systems provides the total gain to the notes played by the driver.
I cannot see any other means of gaining efficiency. If this is not how you view the system to work then I would be interested in knowing just how you do think it works.
I asked Dennis if he would provide some input regarding the Half Square antenna theory as he makes his living in RF design I believe.
Your theories on the principles of operation are very close to mine, with only a couple of differences. But the differences are quite substantial. I already have my theories typed out, as I know a guy that is also interested and I wrote it out to try to explain it as effectively as possible, as I understand it. I don't have any wild new radical ideas, everything I say is backed up by the patent info, but I find the patent is not well written and it took me weeks of studying the text and pictures to put it all together.
Unfortunately, the thing that I wrote is too big to fit here, and I am very hesitant to start cutting it up, as I think all the information is important. I was going to email you but the forum will not allow it. I saw that you have a website, I'm sure contact info will be there. Do you want me to contact you through email?
Unfortunately, the thing that I wrote is too big to fit here, and I am very hesitant to start cutting it up, as I think all the information is important. I was going to email you but the forum will not allow it. I saw that you have a website, I'm sure contact info will be there. Do you want me to contact you through email?
Ok, I will attempt to attach my text to this post. If it works, you will be able to see my explanation of how and why this box works, written with very little technical jargon so a dummy such as myself can understand it. There is a bit of chatter about ported boxes which probably everyone that reads it will already know, but understanding ported boxes is essential to understanding how this thing works.
Also added to the bottom of the document is a post I copied from a different forum (avforum, or maybe avforums.com?). It was written by a guy who built one of these in the 70's. There were 2 threads about this box in that forum but this single post is in my opinion the only one that really matters. Note that I do not necessarily agree with everything in this post. (I mean the post in the attachment, not the one I am currently typing.)
Please let me know if this explanation sounds reasonable and please let me know where I am wrong, if I am wrong. Also, this will probably open in Notepad and you may need to turn on word wrap to view it properly.
Also added to the bottom of the document is a post I copied from a different forum (avforum, or maybe avforums.com?). It was written by a guy who built one of these in the 70's. There were 2 threads about this box in that forum but this single post is in my opinion the only one that really matters. Note that I do not necessarily agree with everything in this post. (I mean the post in the attachment, not the one I am currently typing.)
Please let me know if this explanation sounds reasonable and please let me know where I am wrong, if I am wrong. Also, this will probably open in Notepad and you may need to turn on word wrap to view it properly.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.