Patrick Bateman said:
just a guy said:
well you got me. tomato tomatoe here.
i really don't see a significant difference between this one and the AWC, although your point about the chamber is noted.
so if an AWC has a "bulged" area behind the woofer, does it stop being an AWC? this is pretty grey and i'm not trying to go "rebel" on you here, just to consider an alternative cabinet to the sonotube.
and the sonotube probably does lend itself better to a ported system like the original HSU SW.
you could always do a TH in a rectangular box with a single divider between the diagonals as pictured in another thread. this would be simple if you could get the 45 degree angles cut on the plywood just right. of course the chambers would be equal.
nuconz said:
You're correct. Tapped horn, acoustic wave cannon, dual reflex bandpass, front loaded horn, transmission line, single reflex bandpass are all very similar. The overall efficiency is largely a product of the box volume. Chambers and ports can be used to shape the response, trading off efficiency in one octave for another.
Patrick Bateman said:
ok, i'll take the bait here since i don't know and haven't perused all the threads.
which of the above has "greatest" efficiency, and by how much, and is the "most" linear, and by how much?
and of course, how "difficult" is it to build?
probably need to add something about aesthetics also since not all of us want to sleep on the couch or in the car!
and probably need to add something about "cost".
so we now have efficiency, linearity, difficulty, aesthetics, and cost as parameters. are any others required?
thanks.
None of those things matter much at all if the sub doesn't fit your goals. Start with -
How many subs do you want/ can you get away with?
What is the absolute max size per cab?
What is the exact frequency range you need to produce?
What is the spl level you would like to achieve?
Do you have any preference of the subjective presentation of the bass? (lean and dry or loud and boomy)
Everybody has their own opinion of what is best, everyone has different goals in mind. ANY of the designs you listed have the potential to be great, given a list of goals that suit them.
How many subs do you want/ can you get away with?
What is the absolute max size per cab?
What is the exact frequency range you need to produce?
What is the spl level you would like to achieve?
Do you have any preference of the subjective presentation of the bass? (lean and dry or loud and boomy)
Everybody has their own opinion of what is best, everyone has different goals in mind. ANY of the designs you listed have the potential to be great, given a list of goals that suit them.
nuconz said:
For my money, there are only two truly competitive subwoofer options. Single reflex bandpass is easy to build, filters out 2nd and 3rd harmonic distortion, and offers good efficiency.
Tapped horns offer more output than single reflex bandpass, but comparable efficiency when cabinet size is factored in (in other words, a TH can give you more SPL *but* requires a bigger cabinet than would be needed by the same woofer in a single-reflex bandpass.) A TH isn't as easy to build as a single reflex bandpass. A TH has the same distortion cancelling attributes as a bandpass. In a lot of ways, it's like a dual reflex bandpass. Like a DR BP, it generally has wider bandwidth than a SR Bandpass.
Sealed subs are the easiest to build, but don't have the distortion cancelling attributes of a TH or a single reflex bandpass.
Vented subs suffer from the same flaw as sealed subs.
Dual reflex bandpass is a lot of work to get right. In most situations, I'd take a TH over one, though they DO have their merits.
Patrick Bateman said:
ok, based on what i've seen of the TH in these posts, you get about a 2 octave BW. so a 20 hz horn needs a crossover at 80 hz or so. so a dual TH setup might be required to get "up to" mid-bass (300 hz or so).
i'm wondering if there is a SW design out there that is NOT a direct radiator that offers a BW of 20 hz - 300 hz or so?
i do like the idea of using the front and rear radiation of the woofer as in the TH and the DRBP. to me the important thing with anything is serviceability, especially with a SW in which the woofers are "internal".
in what way is the DRBP "a lot of work to get right"?
thanks.
nuconz said:
Why would you *want* a sub to cover four octaves? There's a millions reasons this is a bad idea.
#1 - you should be using at least three subs. The use of multiple subs smooths response. (see Geddes, Toole, et al)
#2 - If you have 3 or more subs which reach up to 300hz, you're going to have a significant amount of midrange information emanating from your subs. What a great way to ruin the imaging of your mains!
#3 - Subs have waaaaaay more distortion than any other speaker, due to high excursion, high power, etc. You can reduce it with a bandpass or tapped horn design, but if you're trying to squeeze four octaves out of a sub, you'll have tons of 2nd harmonic distortion due to your design choice.
'Gargle' is a function of the square of the bandwidth (in octaves).
If a woofer has one decade of response (30hz~300hz) and 5.5mm of excursion, it will have about 3% frequency modulation distortion.
3% FMD is nasty sounding, it makes things sound 'congested', and female vocals sound like they are 'gargling'.
If a woofer has one decade of response (30hz~300hz) and 5.5mm of excursion, it will have about 3% frequency modulation distortion.
3% FMD is nasty sounding, it makes things sound 'congested', and female vocals sound like they are 'gargling'.
There are also advantages of using (multiple) subs up to maybe 200hz. Using stereo speakers in the modal range begs for uneven frequency response. Using several ways to cover this range is obviously the better choice, but i would rather use one way for 20-200 and accept some FM than have the upper modal range at +-10db due to room modes.
MaVo said:
I agree with you. I thought about this when I replied a day ago, but didn't want the post to get too long. I have a hunch that the best solution would be to use multiple subs for two octaves, say 50-200hz. Then use a single sub for the subterranean bass. Do you remember what the formula is to determine when you can STOP using multiple subs? IIRC, it depends on the room dimensions, and it's the frequency when the subs are simply pressurizing the room.
Patrick, i completely agree with this, actually this is what i want to try next. One should be able to get a single sub working up to maybe 40hz, cancelling the first modes with proper placement, so that the multiple subs can cancel the 2nd order and above modes. I hope this will work in the transition area between single and multiple subs, since i dont know how the room behaves with 5 sources playing.
Toole wrote that there is no sharp frequency, but an area where the room characteristic changes around the frequencies of 200-300hz. The white paper section on harman.com is interesting here.
Toole wrote that there is no sharp frequency, but an area where the room characteristic changes around the frequencies of 200-300hz. The white paper section on harman.com is interesting here.
bend width effect and frequency?
Where can I find formulas on this effect?
Does it tend to be linear? Actually, I am not even sure which direction it goes - for example, tighter bends make a tube (TL, horn, whatever) 'shorter' as frequency goes higher?
I was under the impression that we are stuck with bends as a compromise and that the general rule was always "less bends are better and gentler bends are better" without realizing that 'this may be another case where a benefit may possibly be derived from a function that had been previously as having only adverse effects.
Is there a decent writeup somewhere on exactly what happens when the folds become more and less acute, as a function of frequency?
Mark Seaton said:
Where can I find formulas on this effect?
Does it tend to be linear? Actually, I am not even sure which direction it goes - for example, tighter bends make a tube (TL, horn, whatever) 'shorter' as frequency goes higher?
I was under the impression that we are stuck with bends as a compromise and that the general rule was always "less bends are better and gentler bends are better" without realizing that 'this may be another case where a benefit may possibly be derived from a function that had been previously as having only adverse effects.
Is there a decent writeup somewhere on exactly what happens when the folds become more and less acute, as a function of frequency?
I think there are some AVI movies of this on John Sheerin's site.
(Edit: it's Sheerin without a "g" at the end, and I found the link:
http://ldsg.snippets.org/HORNS/waves.html )
(Edit: it's Sheerin without a "g" at the end, and I found the link:
http://ldsg.snippets.org/HORNS/waves.html )
Re: bend width effect and frequency?
If I'm not mistaken, Akabak is the only program which can model the effect. Programs like horn response have no ability to simulate folds in the line.
Having said that, is it worth the trouble?
Whenever I've personally built subs, I've found that my time is best spent optimizing the FINISHED sub, not the computer model. For example, in this thread I detailed how the subwoofer wasn't performing properly because an end cap was resonating. Stiffening that end cap made a dramatic and audible improvement. I'm not convinced that tweaking folding schemes in Akabak for hours on end would yield worthwhile improvements.
Of course, these opinions are based on the fact that I'm building one or two subs. For a guy like Danley, who's building hundreds and selling them, the effort is worthwhile.
neededandwanted said:
If I'm not mistaken, Akabak is the only program which can model the effect. Programs like horn response have no ability to simulate folds in the line.
Having said that, is it worth the trouble?
Whenever I've personally built subs, I've found that my time is best spent optimizing the FINISHED sub, not the computer model. For example, in this thread I detailed how the subwoofer wasn't performing properly because an end cap was resonating. Stiffening that end cap made a dramatic and audible improvement. I'm not convinced that tweaking folding schemes in Akabak for hours on end would yield worthwhile improvements.
Of course, these opinions are based on the fact that I'm building one or two subs. For a guy like Danley, who's building hundreds and selling them, the effort is worthwhile.
And as a sidenote to Akabak's modeling of bends (aka the "AcouMass" element), I must say that it smoothes the upper freq ripples, but more important: it causes the predicted low freq output to fall off!
This is almost exactly the opposite of what is to be expected, and when I modeled a mid-freq TH the same happened, so it is not in a specific freq band but a modeling problem. Therefore I would simply advise to not model bends in Akabak or if you find the problem, post the solution here.
Unfortunately I'll leave for 1,5 month of trekking in Chile and Peru within a few hours, so I won't be here for a while, but when I return I'll check in on this again.
cheers!
Joris
This is almost exactly the opposite of what is to be expected, and when I modeled a mid-freq TH the same happened, so it is not in a specific freq band but a modeling problem. Therefore I would simply advise to not model bends in Akabak or if you find the problem, post the solution here.
Unfortunately I'll leave for 1,5 month of trekking in Chile and Peru within a few hours, so I won't be here for a while, but when I return I'll check in on this again.
cheers!
Joris
I think the accoumass element assumes that the wavelength is much bigger than the acoustic size of the element, so it only works good at very low frequencies. it doesnt simulate things like high frequencies being reflected by the walls and finding their way out of the horn. it works a little like a lowpass filter.
simulating bends with akabak is something i have no clue how to do correctly for the higher frequencies, because akabaks node structure tells little about how the thing actually looks like (for example, how big the bend angles are).
simulating bends with akabak is something i have no clue how to do correctly for the higher frequencies, because akabaks node structure tells little about how the thing actually looks like (for example, how big the bend angles are).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.