Thank you, Art, it is good to be reminded of this during all the simulating. I think it is clear that the BandPass design has a number of potential disadvantages; IMO by the time the required response is achieved it isn't necessarily going to be the smallest, or loudest, or least complicated, or greatest in the time domain. My interest in it remains mostly in any sonic benefits of the acoustic filter, especially with respect to attenuating spurious above-band noises.
I think only testing will tell me for sure; unfortunately that has meant a fairly large learning curve on my part and would now need a great deal of measurement, tinkering and refinement (to make sure I tested a decent example). So.. now that I understand (through simulation) some of the trade-offs, I might only go to the effort of a physical BP prototype if there is an actual need. Which will depend on whether I find sealed or Bass Reflex designs (with my drivers) to be lacking, particularly with respect to being audibly locatable. Those will be much easier and quicker for me to test in the first instance; if they aren't broken then I won't need to fix anything.
So then the question might just become sealed or BR. I'm in a much better position to make that call, it'll probably just come down to the cost (size and money) vs output required. I usually prefer sealed and suspect this might also have less going on to make unwanted frequencies; I could even double up the drivers (if necessary for SPL). But am not against BR so if (as I suspect) it is a significantly more practical answer then I'd go with it.
In a way 'all' these options are a slightly frustrating compromise to me. Several years ago I had my own detached house so could contemplate large TLs and horns or infinate baffle designs, but now I'm reduced to small rented rooms so can't really have what I'd like. I'm hoping the drivers and amplifiers will last into the future though, and be flexible enough for other more interesting designs if circumstances should ever permit. Perhaps that would be another reason to not get 'too' invested in complicated BP designs, at this stage; even if I really liked them and wanted to stay with them, I'd probably want to re-do them if/when size was less of an issue.
I think only testing will tell me for sure; unfortunately that has meant a fairly large learning curve on my part and would now need a great deal of measurement, tinkering and refinement (to make sure I tested a decent example). So.. now that I understand (through simulation) some of the trade-offs, I might only go to the effort of a physical BP prototype if there is an actual need. Which will depend on whether I find sealed or Bass Reflex designs (with my drivers) to be lacking, particularly with respect to being audibly locatable. Those will be much easier and quicker for me to test in the first instance; if they aren't broken then I won't need to fix anything.
So then the question might just become sealed or BR. I'm in a much better position to make that call, it'll probably just come down to the cost (size and money) vs output required. I usually prefer sealed and suspect this might also have less going on to make unwanted frequencies; I could even double up the drivers (if necessary for SPL). But am not against BR so if (as I suspect) it is a significantly more practical answer then I'd go with it.
In a way 'all' these options are a slightly frustrating compromise to me. Several years ago I had my own detached house so could contemplate large TLs and horns or infinate baffle designs, but now I'm reduced to small rented rooms so can't really have what I'd like. I'm hoping the drivers and amplifiers will last into the future though, and be flexible enough for other more interesting designs if circumstances should ever permit. Perhaps that would be another reason to not get 'too' invested in complicated BP designs, at this stage; even if I really liked them and wanted to stay with them, I'd probably want to re-do them if/when size was less of an issue.
Hi Kev
One interesting point about BP4 boxes is Geddes used to recommend them for his multiple sub setup system. At some point it changed to sealed subs / any old subs you have lying around / multiple different subs is OK.
In an attached UK house I think the neighbours will be kicking your front door in before the subtle nuances of sealed vs BP4 become apparent.
However, I will say that quality of drivers may become an issue. My 4x10" sealed scanspeak driver setup (as stereo bass units) absolutely killed the 3x tempest sealed 15's (3x 15's in 660L) at lower volumes in sound quality.
Rob.
One interesting point about BP4 boxes is Geddes used to recommend them for his multiple sub setup system. At some point it changed to sealed subs / any old subs you have lying around / multiple different subs is OK.
In an attached UK house I think the neighbours will be kicking your front door in before the subtle nuances of sealed vs BP4 become apparent.
However, I will say that quality of drivers may become an issue. My 4x10" sealed scanspeak driver setup (as stereo bass units) absolutely killed the 3x tempest sealed 15's (3x 15's in 660L) at lower volumes in sound quality.
Rob.
Did you ever measure the impedance curve of that build, Art?
I ran into a similar response issue with my "Enigma" build, where there was a "bump" at the lower end of the passband, and it turned out that the build was tuned too low because using the usual vent calcs resulted in vents that were too long (the calculation gets more inaccurate as the volume of the vent gets larger than about 10% of next box volume), Proximity of the vent to the rear wall also affected Fb. I ended up having to trim a good >20% off the vents to flatten the passband.
Your experience with the THD (compared to the vented box) is also interesting. The THD curve for my Enigma build reflected the reduction in THD caused by the acoustic filtering of the vented chamber. FWIW, I also used a bit of stuffing in the vented chamber as I found that subjectively that significantly improved the sound (measurements show that it also significantly reduced the higher out of band vent resonances).
If you're going to try a BP4 build, I suggest going with an offset-driver design, like the "Enigma" build in my page. Why? The shape of the box and the location of the driver in the box can be used to reduce or completely null out the 1st harmonic resonance (you can actually see this effect in the FR posted by Art for his build). I've got a spreadsheet on my website that can be used to generate an equivalent Hornresp model for such a design.
Whatever you do, I also suggest designing your build in such a way that you can easily change vent lengths. Even a few cm change in length can make a pretty big difference, not only to the shape of the passband, but to the amount of out of band noise as well. Finally, including some stuffing in the vented section, near to the panel that's farthest away from the opening of the vent(s). This should reduce some of the out of band noise. I lucked out with the Enigma - the out of band noise in the final version was so low that I could use it without a LP filter.
No, but since the response is kind of inverted from the Hornresp sim, I'd expect the impedance curve would be too.
I found after building the Bowtie sub my original 2012 Hornresp model was done completely wrong, even though it resulted in a similar response curve as the finished cabinet
.I had figured the tapered inlet and exit would need some compensation, after adjusting the port width (lots of clamping and cursing) got it to do 40-100Hz. For rentals, I figured the "bump" at the lower end was a bonus.
The Hornresp sim above was done long after I sold it.
I found the transient response lousy when A/B tested compared to the LAB12 in BR, FLH, or TH designs.
I didn't find the THD a sonic issue even though the upper pass band distortion measured 3 to 4 times higher than the LAB12 in BR, FLH, or TH designs tested at 49v, while the Bowtie low end distortion was similar.
Maybe the LAB12 just does not like moving +/-7mm 120 times a second
Art
Hi Rob, thanks again for your thoughts. Yes it was Earl Geddes using bandpass boxes that got me wondering about using them myself. I recall a post saying that he used them to reduce localisation in his distributed subwoofer setup, but also that he would otherwise have preferred sealed. So I think if my drivers/boxes don't give their location away, I might stay with sealed (or BR) for practical reasons. So it is very helpful to hear your thoughts on the comparison, thank you!
Yes I'm very interested in sound quality, and more so at normal living-room volumes than at the highest ones; I do watch occasional films with rumbles and explosions etc, but it is only really music where I want top sound quality. The Lab12 has a fairly good reputation, though they are a reasonably old and modestly priced design, optimised (IIRC) for use in big horns, and mine are not the true/original Eminence version either. So we shall have to see how they perform in practice; if they make spurious noises or need BP to hide out-of-band deficiencies then that could make the decision for me.
Thanks,
Kev
Thank you for the tips, Brian. The offset design makes sense, now that you mention it. I'd noticed that the hornresp simulations do show interaction of the box and port together on what I've simply been calling the port resonance; it isn't an open pipe in isolation.
If I build anything with a port I'll make at least one prototype to enable large adjustment and testing before settling on the final thing. But yes it might still be useful to have the final version somewhat adjustable too; these will see different rooms and different positions so I might want to tweak them in ways other than DSP. Thank you for the suggestion!
Cheers,
Kev
Another BP tip is that if you look at the JBL TCB design you can go push push. Also the magnet side of the driver is in the ported part of cab which helps with cooling. I read somewhere that it also blocks / reflects higher frequencies away from the port. How true / measurable that would be I have no idea.
Rob
Rob
Thanks again, Rob. It does sound like there are a great many tweaks to try out with Band-Pass designs. There seem so many subtle and not-so-subtle things about their characteristics and configuration, and hoe different drivers work with them, that I begin to understand why people's success is so varied (not even counting car subwoofers).
Which in itself is interesting. For now I probably just need something to work well enough, whilst I concentrate on my main speakers. But I think ultimately I'll have to try bandpass designs too; there is clearly a lot going on with them, and some common ground with the MEH/unity-horn main speakers.
Cheers,
Kev
Which in itself is interesting. For now I probably just need something to work well enough, whilst I concentrate on my main speakers. But I think ultimately I'll have to try bandpass designs too; there is clearly a lot going on with them, and some common ground with the MEH/unity-horn main speakers.
Cheers,
Kev
The problem is the bump is in the wrong location.
The bump needs to be in the 80-100hz range to bring in the midbass.
That's why the Devastator enclosure on avsforum.com is designed the way it is.
Is that a 2pi or 0.5pi HR model?
If that is a 2pi HR model, then you definitely gonna get muddy bass in a car.
Transfer Function is a beast below 50hz.
If that were my 2pi HR model, then I would have that 90hz hump around 125dB so the curve flattens out at 0.5pi with transfer function.
The 2x12" Bowtie was designed to be small enough to be transported in the trunk of a car for DJ rental use, not as a car sub. It was designed around cabinet dimensions common to the rest of my speaker inventory at the time.
The Bowtie Hornresp model is 2pi:
The HR sim does not have the LF bump.
I didn't spend much time on the sim, as the cabinet had already been sold, and I had no intentions to build any more.
Like any bandpass subwoofer with output only through a port, the upper response is indirect, a resonant delayed response from the internal air spring.
Efficient for the enclosed volume, but lacking in transient accuracy.
That type of response is OK for the intended users, but not something I would choose for my own use.
Kev06 plans to use the FAB12 (not the LAB12) for home use.
Art
Adding a 12" LP filter to a direct radiator box also adds delay .
In any case, using DSP to address the TD should improve the "transient accuracy" a bit.
Oh, I found the THD curve for my Enigma (V2) BP. The curve was taken at the driver's rated power level. THD of ~1% from 50 Hz up is not too bad. Above 110 Hz or so it's a bit less relevant as the BP is rolling off pretty quickly above that frequency. Below 50 Hz the acoustic filter is no longer effective, so THD rises quickly.
(forgot to mention, that little blip in the THD above 150 Hz was caused by a buzzing x-over board, which I subsequently fixed).
.In any case, using DSP to address the TD should improve the "transient accuracy" a bit.
Oh, I found the THD curve for my Enigma (V2) BP. The curve was taken at the driver's rated power level. THD of ~1% from 50 Hz up is not too bad. Above 110 Hz or so it's a bit less relevant as the BP is rolling off pretty quickly above that frequency. Below 50 Hz the acoustic filter is no longer effective, so THD rises quickly.
(forgot to mention, that little blip in the THD above 150 Hz was caused by a buzzing x-over board, which I subsequently fixed).
A bit of an apples and oranges comparison.
A bit fairer to compare a BP4 box to a sealed box, and a BP6 box to a vented box ?
Rob.
No, it's a fair comparison. His BR has the same volume as my BP4. My BP4 is more efficient per a given voltage up to 38.97 volts. The BP4 still reaches a higher SPL even though the BR can take 50 volts.
Also, note the BR's port velocity is around 24m/s. The BP4 is below 13m/s.
Original WinISD would go into the red above 15m/s. In another diyaudio thread, we established 17m/s should be the maximum.
Also, note the BR's port velocity is around 24m/s. The BP4 is below 13m/s.
Original WinISD would go into the red above 15m/s. In another diyaudio thread, we established 17m/s should be the maximum.