Coupled Cavity Subs

TLS Guy

TLS Guy

Audioholic Jedi
Matt 34 has invited a thread on band pass, more correctly called coupled cavity subs.

This came into vogue in the 1980s. KEF produced a number of designs, like this.



These speakers certainly extended the bass of the 10 inch drivers especially because the bass roll of is second order rather than the fourth order roll off of the ubiquitous Qb4 ported box. They really did not reduce enclosure volume. However this is what they achieved. Not bad.

Bose of course started their long running series of bass modules.

An in depth discussion and theory behind coupled cavity subs appeared in speaker builder six/88.

Since the eighties interest in band pass bass sections for reference systems has declined. The technique has largely found its place in car audio, computer sound system and of course Bose systems.

The 1998 article encouraged me to build a number of these designs, of which only three were any good. However I learned a lot a long the way.

The eighties, was an era where pushing bass extension very much came into vogue. The B & W answer to the KEF was to make their flagship model a Qb6 box with a 15" woofer. A standard Qb6 box alignment has to include an active equalizer. The speaker consumed gobs of power, the driver could be made to bottom and this design was short lived and largely forgotten.

So what are the issues with coupled cavity designs.

If an isobarik design is chosen a low F3 can be obtained from a relatively small enclosure.

The designer has total control of Qt. For me that was a big attraction.

The low pass crossover from the bass unit can potentially be acoustic.

Bass roll off for the single tuned band pass box is second order and obviously helps extend the the bass response. The roll off is second order on the top and bottom end.

The down sides are decreasing efficiency as the band pass response is extended.

This is especially true for Single tuned Band Pass Box, like this.



This single tuned band pass box is like the KEF solution and like one of the successful boxes I built.



This improves efficiency because of the use of two drivers with their own sealed space.

This is the isobarik version.



My other two succesful designs were like this.

This obviously lowers efficiency but halves enclosure volume for the same F3.

These designs can work quite well. The phase shift is never more than 360 degrees and so is comparable to the vent output of a standard Qb4 box.

The big problem is efficiency, as high output can only be obtained over a very narrow range of frequencies. This leads to the charge so often leveled against these designs, of being "one note wonders."

If response is extended over at least one and a half octaves efficiency will be poor.

Drivers unload below tuning just like a Qb4.

All output comes from the port and so cone radiation is lost.

Port design is difficult unless you use a low pass filter and increase the order of the higher cut off. If you don't cut off the response an octave or so above the upper band pass roll off you will get port resonances stating around 900 Hz or so, or lower if the port is longer. So port artifacts are a problem. A large diameter or vent port makes the use of a crossover at the upper turnover frequency mandatory. This problem negates one of the advantages of a band pass design in my view.

So lets move on to the double tuned band pass box.

This comes in two forms, the parallel configuration.



And the series configuration.



Efficiency is improved by 6 db for the parallel alignment but only 3 db for the series, so the parallel arrangement is to be preferred.

The low frequency of the band pass rolls off at 24 db per octave and the upper end at 12 db per octave.

The big issue is transient response as the phase shift has to be at least 450 degrees and can be as high as 540 degrees. Distortion also rises as you sequence ports.

I'm not aware that any high end designs have used these higher order alignments. This brings me to the issue of nomenclature, where there is confusion. Different authors refer to the first series of enclosures variously as second or fourth order, and the second as fourth and sixth order. So I have used a less confusing and more descriptive nomenclature.

I personally favor the order to designate the order of the bass roll off.

If high sound quality is the goal, it is my opinion you need to stop at the first series of enclosures and reserve orders above that for car use, compact computer systems and Bose. That is pretty much the way it has turned out.

Narrow band pass units with the band pass centered around 50 Hz have proved popular for that high powered bass lick, but I'm sure that need not concern any of us.
 
GranteedEV

GranteedEV

Audioholic Ninja
Port design is difficult unless you use a low pass filter and increase the order of the higher cut off. If you don't cut off the response an octave or so above the upper band pass roll off you will get port resonances stating around 900 Hz or so, or lower if the port is longer. So port artifacts are a problem. A large diameter or vent port makes the use of a crossover at the upper turnover frequency mandatory. This problem negates one of the advantages of a band pass design in my view.
And what if a passive radiator is used instead?

The big issue is transient response as the phase shift has to be at least 450 degrees and can be as high as 540 degrees. Distortion also rises as you sequence ports.
In the "parallel" alignment, how relevant is phase shift or delay if port is tuned well in the subsonic range, away from audible frequencies? IE if a bandpass sub effectively operates from 12hz to 110hz. No electronic high pass filter is used... rolloff towards 12hz is gradual, not flat. Let's assume driver behavior is kept under control despite the low tuning(IE no overexcursion above tuning).
 
TLS Guy

TLS Guy

Audioholic Jedi
And what if a passive radiator is used instead?



In the "parallel" alignment, how relevant is phase shift or delay if port is tuned well in the subsonic range, away from audible frequencies? IE if a bandpass sub effectively operates from 12hz to 110hz. No electronic high pass filter is used... rolloff towards 12hz is gradual, not flat. Let's assume driver behavior is kept under control despite the low tuning(IE no overexcursion above tuning).
I have never seen a design use a passive radiator, but there probably is one out there. However I would bet the high frequencies would pass right through the cone, and the need for a steep low pass filter wold be greater.

You can't tune the port subsonic. It is is the sealed side that that is the high pass filter that limits the low frequency response and the vented side is the low pass filter that limits the upper frequency response.

You would have a hell of time making a band pass sub go from 12 Hz to 110 Hz, that is over three octaves, 1 to 2.5 octaves is the best to hope for. 1.5 octaves is what you usually get, or less. Remember efficiency drops as you extend bandwidth, and that takes a lot of power and leads to driver bottoming.

The KEF 107 had a pretty wide bandwidth with 3 db point of 20 and 150 Hz, if I remember correctly. They used drivers that would not be considered good sub drivers now. They were light coned high efficiency drivers, to get the sensitivity into a reasonable range. I don't know what the sensitivity of the drivers were, but I bet around 100 db 1 watt 1 meter.

For the double tuned box delay is highly significant, a wave length and a half. Transient response has been reported as poor by all investigators, Bullock and Dickerson among others.
 
Stereodude

Stereodude

Senior Audioholic
For the double tuned box delay is highly significant, a wave length and a half.
I guess it's a good thing receivers and pre/pros let you delay the other channels to compensate. ;)
Transient response has been reported as poor by all investigators, Bullock and Dickerson among others.
All?.?.?. Really?

I do find it fascinating that you chose to use examples of bandpass designs in full range speakers as an indictment against them as subwoofers too.
 
TLS Guy

TLS Guy

Audioholic Jedi
I guess it's a good thing receivers and pre/pros let you delay the other channels to compensate. ;)
All?.?.?. Really?

I do find it fascinating that you chose to use examples of bandpass designs in full range speakers as an indictment against them as subwoofers too.
I use only single tuned types and the phase shift is comparable to the phase shift of a port from a Qb4 box, as I pointed out.

For a double tuned bass box the delay would be 12.5 msecfor an 80 Hz tone 25 msec for a 40 Hz tone and 50 msec for a 20 Hz tone. So really once phase shift exceeds 360 degrees time smear becomes increasingly significant.

I don't think I would build anymore coupled cavity boxes, as I think there are better bass loading systems. Quite honestly I don't believe their advantages, attractive though some of them are outweigh the disadvantages over all, but that's my opinion.
 
M

Mark Seaton

Junior Audioholic
Hi TLS Guy and others,

I just caught up with this thread as I was out of town the week prior and on vacation last week. My argument to the above is that most objections are related to viewing in isolation or related to poorly executed examples. A few quick comments to start:

I have never seen a design use a passive radiator, but there probably is one out there. However I would bet the high frequencies would pass right through the cone, and the need for a steep low pass filter wold be greater.
Actually quite the opposite. Admittance of high frequencies through the higher frequency or front side port is related to resonances due to size and length up to much higher frequencies where noise could make it through the open airway, depending on the physical configuration. A PR follows the predicted high frequency roll off (without port resonances) very closely. James Loudspeakers had their EMB products which are 4th order systems (sealed rear, PR front), albeit crammed in excessively tiny enclosures for the parts used. Many had found them to have plenty good subjective sound quality.

You would have a hell of time making a band pass sub go from 12 Hz to 110 Hz, that is over three octaves, 1 to 2.5 octaves is the best to hope for. 1.5 octaves is what you usually get, or less. Remember efficiency drops as you extend bandwidth, and that takes a lot of power and leads to driver bottoming.
If the primary design goal is not simply output gain, 12-100Hz is most certainly possible from a 6th order BP (dual reflex, front/rear, separate tuned chambers).

For the double tuned box delay is highly significant, a wave length and a half. Transient response has been reported as poor by all investigators, Bullock and Dickerson among others.
While the group delay IS increased, the why is very important. If you take a look at the operation, a 4th order / single reflex BP exactly follows the magnitude and phase response of a simple sealed subwoofer with a 2nd order electrical low pass. Similarly, a 6th order / dual reflex BP exactly follows the magnitude and phase response of a simple reflex design with a 2nd order electrical low pass. I view these as more the conventional sealed or reflex design with an acoustic low pass added to the system. There are more benefits than just the acoustic low pass, but if we take a look at all of the studies on distortion audibility, it is clear that higher order distortion is more audibly offensive or identifiable, and the front chamber effectively filters out of band distortion harmonics by ~12dB/octave.

More importantly, when using a PR which will not allow the passing of resonances at higher frequencies, you find that if you first match the magnitude response, the magnitude and phase response of a BP is the same as a woofer of the same rear chamber design with a 2nd order low pass applied. If measured as a black box within linear limits, you can not clearly differentiate between the two.

I guess it's a good thing receivers and pre/pros let you delay the other channels to compensate. ;)
All?.?.?. Really?
I use only single tuned types and the phase shift is comparable to the phase shift of a port from a Qb4 box, as I pointed out.
The difference in phase/group delay is entirely due to the front side acoustic low pass, and is there from any equivalent response low pass. So long as you can integrate the higher frequencies with this additional 2nd order low pass in the chain, it is no different than switching to a steeper slope on your subwoofer low pass filter. Obviously I don't expect everyone to agree with my conclusions on the benefits, but I would strongly suggest that the primary issues have been ones of execution rather than inherent to the design.
 
newsletter

  • RBHsound.com
  • BlueJeansCable.com
  • SVS Sound Subwoofers
  • Experience the Martin Logan Montis
Top