Ask an Acoustic Engineer (me)

[Holmz]

How did you get into car audio acoustics? I am majoring in mechanical engineering with an acoustics concentration, going into my senior year. Car audio acoustics is certainly an interest of mine. After I graduate I might consider going into that, or to grad school.

The stuff that is being done with active cancelation in exhausts and also "headers" which use speakers to have a higher bandpass of tuning is pretty slick.

[Holmz]

Hi Justi,
A couple of questions in 6th and 8th order bandpass boxes...

in a 6th order "dual bass reflex" can design the lower frequency side using normal 4th order ported software tools, and then consider the higher frequency side the same way?

the two questions for the 8th order...
1)
In the 8th order that is a 6th "order dual bandpass", blowing into a common chamber where are those cambers tuned?
Let's say it is 30, 45 and 60 Hz... and the chambers are D, E an F... (just to make it different than ABC.)
So D is at 30 Hz, and the E is to the right... and the chamber F sit on top with D and E blowing into F.
If the E chamber/port is tuned to 45 Hz, and does then F just pass all notes below its tuning freq of 60 Hz?
Is that third chamber just to add some gain at the highest frequency, and also to limit/attenuate frequencies above that chamber and port's tuning frequency?

2)
Then in the ABC design, with a port between the larger and smaller chambers... most of those call for the same port dimensions in all three sections. How does the lower frequency work? and the port that connects the two chambers seems to not be tuned to 30 Hz?
it seems like if that is for the 30 Hz side, then by the time the 30 Hz tunes come out of the other two ports that they would be 180 out of phase?
And why isn't each port tuned to a specific frequency, along with the two chamber volumes differing?

[geolemon]

I used to do a lot with ABC boxes right around the time I was doing sub engineering - I even have a 15 year old prototype XBL^2 ten in a prototype compact ABC box... which I'd generally call a failure, because Hoffman's Iron Law still holds true. The frequency response wasn't bad, but didn't have much bottom octave, and wasn't all that loud because I really wanted to see how small I can push things... this was my "that's enough, then" smallest box, end of experiment. Still have it...

If you think of how a port works, you can see how both the coupling port and the rear port work (and why all the group delay). It might be easier to picture how passive radiators work - they work the same as ports.
I won't step on Justin's toes, I'd love to see a good description myself...

To me it makes sense - the front chamber pretty much working like a traditional vented box - but also stimulating a column of air in the coupling vent which in turn works like a subwoofer stimulating that second chamber (again, maybe easier to imagine a passive radiator), which then has it's own port to the outside world.

But yes, the tuning - why "equal port lengths" and why the result? That's definitely worth diving into, and even though the whole daisy-chain of sub-box-vent-box-vent to me makes intuitive sense "why big group delay?", the tuning aspect is more of a mystery.

I would presume (but never dove into it) that if you designed an ABC box using the "three equal ports" as a starting point - that'll at least get you to a known place.
Then - let's say you take some measurements in-car, and decided that... I don't know, you have a very small car, so you are perfectly happy with your SPL around 60hz, but you are still seeing a slight rise from there to 40hz... and you'd rather flatten it and extend your response to 30hz. So, maybe you take your existing design, and lengthen the port coming out of the larger chamber - I'd think you could do that. Then they wouldn't be equal, and I think you'd have the tweak you want.
...what I don't know is, would that impact the other chamber? Part of me thinks "yes", because that would fundamentally alter the pressure dynamics inside that chamber. Part of me thinks "no", because that second vent will still function as a vent, and that sub is still the driving pressure actuator in that chamber.

So what if you had the OTHER problem? Maybe you like your low-end response, but you want to change the upper response. This seems very tweaky, like a bandpass box, but whatever - let's say you want to raise or lower your tuning.
I'd be interested in the effects of both the coupling port and the output port, on that secondary chamber.
I'd speculate that you could most easily tune the output port, on the secondary chamber, by shortening/lengthening it.

Changing the length on the coupling port really makes me wonder - especially from my comments on group delay - could you tune the coupling vent higher (shorter length) and then tune the output vent lower (longer) to get the same tune? Could there be a benefit (group delay) or is the combined effect of the shorter-one-place, longer-the-other just going to "net nothing" for you?
Is it as simple as "if you tune the coupling vent higher, without lengthening the output vent, you'll effectively tune that chamber higher?"

I'm definitely curious. Always had luck with that design, but definitely fall short of an expert, despite experience.

2)
Then in the ABC design, with a port between the larger and smaller chambers... most of those call for the same port dimensions in all three sections. How does the lower frequency work? and the port that connects the two chambers seems to not be tuned to 30 Hz?
it seems like if that is for the 30 Hz side, then by the time the 30 Hz tunes come out of the other two ports that they would be 180 out of phase?
And why isn't each port tuned to a specific frequency, along with the two chamber volumes differing?

I used to build a lot of these myself. I always used the "same port dimensions" for all three, but I have speculated that you could probably substitute an effectively same-tune port of different dimensions, if for some reason (maybe depth) you really wanted slot vents to the outside, but needed to use maybe two round ports in the corners, for the coupling port. I never did experiment with that, will be interesting to see what Justin thinks.

I will comment that I always did like this design, specifically because the default tuning of both chambers really seems to line up incredibly well against the typical car interior's cabin gain... So unlike a high-tuned vented box or some poorly designed bandpass boxes that become one-note wonders, you essentially have SPL boost between your two tuning frequencies... it gives you more boost towards the lower frequency, and that slope between the lower and upper tuning frequency lines up really well to pretty much flatten out inside a car.
...which also explains why these things would be total failures in a larger, more open environment like a listening room or a bar or club. The room gains wouldn't be complimentary, like they are in a car.

There's probably too much group delay to call this an "SQL dream", but for someone who likes it loud, and doesn't quite reach the IASCA-SQ-competitor level of audiophile, it can really be a great fit.

Also, Justin-
Could you maybe voice your opinion on what I've always thought was a misnomer, "Aperiodic Bi-Chamber" just doesn't fit - it's all vents. Never seen one done aperiodic. I feel like someone plugged in a word there, for marketing reasons... "Eh, sounds fancy. They'll never know."
That being said, the other acronym "Dual chamber tri-tuned" or DCTT I always felt sort of misnomer-y on also. Yes, dual chamber. Tri-tuned? Ehhhh... I guess kinda sorta but... really there's two vents, and yeah a vent in the middle, but...
Well, maybe that is a good link back to my other questions on the tuning what-ifs.

[Holmz]

My further confusion includes:
If the port(s) create nodes that also limit cone excursion then:

1) How does the sub stay cool in a bandpass if there is limited excursion at/around each control node?

2) What causes the impedence to rise at the places of low excursion?
(the inductance and resistance is the same, so is it back EMF?)

3) If ported arraignments have a higher impedence at the frequencies where there is low excursion, then how are they able to sound louder? (as the delivered wattage must be going down proportional to the impedence rise)

...
There's probably too much group delay to call this an "SQL dream", but for someone who likes it loud, and doesn't quite reach the IASCA-SQ-competitor level of audiophile, it can really be a great fit.
...

If one could remove the group delay, then it seems like the main limiter to an "SQ dream" is transient response?

It is possible that one could believe that a bandpass limits distortion, and if that were part of SQ, then could it be worthwhile to consider a bandpass box without any desire for SPL?

[geolemon]

If one could remove the group delay, then it seems like the main limiter to an "SQ dream" is transient response?

It is possible that one could believe that a bandpass limits distortion, and if that were part of SQ, then could it be worthwhile to consider a bandpass box without any desire for SPL?

If you are asking "Could you DSP the group delay out?" then I suppose technically that might be possible... I've never heard of anyone doing that. It's not a simple delay, it's an increase by frequency (so, like vented) and in the case of a 6th order bandpass doubly complex - so you'd have to have some pretty trick DSP software with a way to implement variable delays by frequency... I'm sure it's theoretically possible, but I don't know if anyone has made software with those features.

You can design a bandpass box without the SPL boost - I'd say "just give it a wide passband" but of course I"m sure you've played with modeling software, and that's not always that simple - but you get the idea. Right driver, 4th or 6th order, tuned conservatively on the low end, then widen out the upper tune - you can shape that curve. But there's the old question of "why bother?"

...of course you do mention a good "why bother" - the fact that the port does act like an acoustic filter, at a slope like a crossover, higher frequencies don't escape. So sure - it hides some distortion. The downside is, if you are beating the snot out of the sub you won't hear that bottoming-out death knell - it'll just die. Then you'll know.

[Holmz]

If you are asking "Could you DSP the group delay out?" then I suppose technically that might be possible... I've never heard of anyone doing that. It's not a simple delay, it's an increase by frequency (so, like vented) and in the case of a 6th order bandpass doubly complex - so you'd have to have some pretty trick DSP software with a way to implement variable delays by frequency... I'm sure it's theoretically possible, but I don't know if anyone has made software with those features.
...

Most poeple ether do ^all that^ with a FIR dsp, or using MiniDSP with DIRAC to do that sort of thing.
i suppose it seems possible to do it with IIR?

...
You can design a bandpass box without the SPL boost - I'd say "just give it a wide passband" but of course I"m sure you've played with modeling software, and that's not always that simple - but you get the idea. Right driver, 4th or 6th order, tuned conservatively on the low end, then widen out the upper tune - you can shape that curve. But there's the old question of "why bother?"
...

What modelling S/W?
i have not been able to find anything.

...
...of course you do mention a good "why bother" - the fact that the port does act like an acoustic filter, at a slope like a crossover, higher frequencies don't escape. So sure - it hides some distortion. The downside is, if you are beating the snot out of the sub you won't hear that bottoming-out death knell - it'll just die. Then you'll know.

Does "beating the snot out of it" consist of sending it DC, or over excursion?
With a 300W amp and the control points I do not think I will be running into XMax, nor thermally cooking it.

Now if one wanted to make up for a lack of Midbass, then one could cross the sub over higher. But then to avoid localising it, the distortion needs to be somewhat lower.

Then there is also a double whammy as the lower excursion, around the control points, also lowers distortion... so one starts out low, and then any distortion gets further filtered within the box.


But it I am not sure which bandpass we are taking about? The ABC? or the one with a chamber on top of a dual bass reflex.
i am still unclear on the which chambers are responsible which bands?

[geolemon]

Most poeple ether do ^all that^ with a FIR dsp, or using MiniDSP with DIRAC to do that sort of thing.
i suppose it seems possible to do it with IIR?

I'm not talking about simple delay - like those slick auto-EQ systems sometimes adjust both EQ and distance for time arrival...
I'm talking about group delay, which is a variable by frequency, imposed by the vent - which comes on progressively - and that's the simple description of a normal vented box.
Once you get into a bandpass box, you have two vents, two tuning frequencies, a much more complex group delay plot.
...then an ABC box? I can only imagine the group delay - plus in that case, my belief is you have more group delay from the high-tune chamber, yet you have the direct output of the sub. In that case, I would believe it impossible to DSP out, no matter how fancy your DSP. It's only the bandpass box that I can even fathom it as possible, and I don't think people are doing that often, if ever.

For reference, here's a group delay plot for a simple vented box:

What modelling S/W?
i have not been able to find anything.

For an ABC box? I built a spreadsheet like 15 years ago, I'd have to dig that out of the archives...
But it started with a basic vented box tune, and then calculated the response based on the individual chambers, combined.

If you want to build one, you basically model up a standard vented box (like I mentioned earlier - the bigger the better, as "compact" really kills your benefits). I used to use round ports - use a reasonable sized round port. You don't want chuffing, but you also don't want to make your box absolutely HUGE because of all the port displacement inside). Your software will give you your total box volume and the dimensions of ONE port. You'll actually make three ports that size.
And I don't even recall now where the tuning target was... I think you want to try to tune it reasonably low, like 30hz...
I'll have to dig out that workbook, now I don't even recall. I believe that both chambers will end up with an effective tune higher than your base tune, but it wasn't too extreme. You'd end up with a rising response between those two tuning frequencies, and that combines nicely with cabin gain - but like I said, just an odd FR plot for anything else.

Now it's time to build. You build a single box, two chambers:
Chamber 1: 2/3 of your total volume. You also need to include your sub displacement and the displacement of 1.5 vents, in this chamber.
Chamber 2: 1/3 of your total volume. You also need to include the displacment of the other 1.5 vents.

Effectively, the larger chamber gives you your lower tune, as it does for a bandpass box. And the smaller chamber will give you your higher tune - it doesn't care that it's a moving mass of air pressurizing it rather than a cone. I'm sure you could build one with three passive radiators rather than vents. That would be interesting.

Does "beating the snot out of it" consist of sending it DC, or over excursion?
With a 300W amp and the control points I do not think I will be running into XMax, nor thermally cooking it.

Now if one wanted to make up for a lack of Midbass, then one could cross the sub over higher. But then to avoid localising it, the distortion needs to be somewhat lower.

Then there is also a double whammy as the lower excursion, around the control points, also lowers distortion... so one starts out low, and then any distortion gets further filtered within the box.


But it I am not sure which bandpass we are taking about? The ABC? or the one with a chamber on top of a dual bass reflex.
i am still unclear on the which chambers are responsible which bands?

On both the ABC and the bandpass - and for any box, for that matter - if you keep port dimensions the same, the larger you make your chamber, the lower you make the tune. So for both bandpass and ABC, your larger chamber is your lower tune, controlling the lower part of the response plot. And your smaller chamber is your higher tune, controlling the shape of the upper part of the response plot - as well as whether you have a dip between peaks.

Only the bandpass will mask distortion, because the subwoofer is 100% inside the box - all that is escaping the box is the pressurized air fluctuations, the indirect air that's being stimulated by the sub, in the box.
The ABC box, you also have the subwoofer cone itself outside the box, working pretty much like a normal vented box - so you'll still hear all the high frequency content of distortion, abuse, bottoming out, lead-slap, you'll hear it.

Sure - on a 300w amp, you aren't TOO likely to encounter that. But you could - get a cheap sub, and/or build a huge box, and/or tune it pretty high then play a 20hz tone, etc, etc.

For the strategy of raising the crossover point for make up for poor midbass - yes, it makes sense conceptually... but also it's not as simple as just raising your crossover frequency. I take that back - for an ABC box - sure.
But for a bandpass box - again, that upper tuning frequency acts as a low pas filter. So if that's cutting off your response at 70hz... you could totally disable that crossover and you still aren't getting 100hz out of that box.

[Justin Zazzi]

Hi Justi,
A couple of questions in 6th and 8th order bandpass boxes...

There are many variations of 6th and 8th order enclosures, which to kinda answer geolemon's question ... I doubt there is a super scientific method to how the various configurations were named which makes it harder to identify which is which by just using words. To make sure we're talking about the same ones, please share a picture of the one you're interested in. For example, one of these:
https://www.the12volt.com/caraudio/s...ofer-boxes.asp

My further confusion includes:
If the port(s) create nodes that also limit cone excursion then:

1) How does the sub stay cool in a bandpass if there is limited excursion at/around each control node?

Great question. My only thought is the ported nature allows fresh air exchange with the outside world, allowing the internal box temperature to be cooler than a sealed box. This might be a substantial effect or a tiny one depending on the port non-linearities. Also, if you know the woofer needs to be small excursion then you could optimize the design with a higher BL^2/Re ratio which might make it more energy efficient (1w/1m) to generate less heat in the first place.

2) What causes the impedence to rise at the places of low excursion?
(the inductance and resistance is the same, so is it back EMF?)

I have not studied bandpass and higher-order enclosures as much, but the impedance peaks I have seen are due to back emf when the cone has the greatest velocity (near greatest excursion). There is an impedance minimum in a normal ported enclosure at the box tuning frequency there the velocity is minimum (and excursion is minimum), for example.

3) If ported arraignments have a higher impedence at the frequencies where there is low excursion, then how are they able to sound louder? (as the delivered wattage must be going down proportional to the impedence rise)

I think it's the other way around: low impedance at low velocity (low excursion). See question 2 above.

If one could remove the group delay, then it seems like the main limiter to an "SQ dream" is transient response?

It is possible that one could believe that a bandpass limits distortion, and if that were part of SQ, then could it be worthwhile to consider a bandpass box without any desire for SPL?

If you can remove group delay with a FIR filter like you suggest, then latency will suffer so it becomes harder to sync with video such as watching a movie. You would still need to tackle transient response, yes, and also directivity and timbre and all those other things too but purely in terms of a subwoofer enclosure I think that would be most of the problems yes.

[Justin Zazzi]

I'm not talking about simple delay - like those slick auto-EQ systems sometimes adjust both EQ and distance for time arrival...
I'm talking about group delay, which is a variable by frequency, imposed by the vent - which comes on progressively - and that's the simple description of a normal vented box.
Once you get into a bandpass box, you have two vents, two tuning frequencies, a much more complex group delay plot.
...then an ABC box? I can only imagine the group delay - plus in that case, my belief is you have more group delay from the high-tune chamber, yet you have the direct output of the sub. In that case, I would believe it impossible to DSP out, no matter how fancy your DSP. It's only the bandpass box that I can even fathom it as possible, and I don't think people are doing that often, if ever.

For reference, here's a group delay plot for a simple vented box:

Seems like you're both talking about the same thing. A couple super-fancy dsp systems like Dirac and the APL1 use a combination of IIR an FIR filters to correct frequency and phase response so that the impulse response becomes more ideal, which results in the group delay being more ideal (flat) too. The same concepts can be used to correct group delay in a high-order subwoofer enclosure at the expense of latency which I mention in my previous post. I've always wanted to try this.

[dgage]

If you want to learn more about higher order subwoofer enclosures, about the most knowledgeable I’ve seen are at data-bass.com though those enclosures are more for high output outdoor venues or high performance home theater. But the concepts scale.