I am trying to understand the problems with pushing drivers in the range below their Fs. For example trying to use a bank of small drivers with Fs of say 100 Hz to make bass to 40 Hz. Having many small drivers will give one enough cone area to move enough air to be heard but I have tried it and the bass, while there in adequate SPL is not clean. I did not do any sweep generator or instrument analysis during the one time I tried this experiment so I am going to take an edumedicated guess. Is it because Fs falls at the 2nd harmonic of the bass passband being pushed and as such any bass reproduced at exactly half Fs, or near it, will ring the driver's Fs where it is very efficient and therefore come out of the driver as mostly 2nd harmonic and thus distortion?
If this is the key issue then a really high slope LPF could be used, but this would not be of any use if the bank of drivers is being asked to operate as full range. What may be needed is a circuit that could lower the electrical Q of the driver bank at Fs without throwing away too much electrical efficiency at all other frequencies. Just thinking aloud here.
If this is the key issue then a really high slope LPF could be used, but this would not be of any use if the bank of drivers is being asked to operate as full range. What may be needed is a circuit that could lower the electrical Q of the driver bank at Fs without throwing away too much electrical efficiency at all other frequencies. Just thinking aloud here.
rcavictim said:
The explanation is simpler than that. Small drivers tend to have lower Xmax than large drivers, so your maximum possible output will be lower even if the total cone area is the same. Remember if your swept volume (area * Xmax) is halved, max output is 6 dB down. That means unless you get a veritable wall of these little punks they'll lack the visceral slam of a good large woofer.
Cheers,
Francois.
You have just found out experimentally what Thiele, Small and others have found out before.
The higher the F3 compared to Fs, the cleaner the bass.
The lower the F3 compared to Fs, the sloppier the bass.
The "classic" bass reflex has F3=Fs, Vas=Vb, Qts=0.4
You can get away with Qts=0.55 or so, and F3=0.7 Fs. After that, bass tends to get distinctly distorted.
The higher the F3 compared to Fs, the cleaner the bass.
The lower the F3 compared to Fs, the sloppier the bass.
The "classic" bass reflex has F3=Fs, Vas=Vb, Qts=0.4
You can get away with Qts=0.55 or so, and F3=0.7 Fs. After that, bass tends to get distinctly distorted.
You can do what you want to by using a sealed box ot TL so that the rolloff below Fb is 12dB/octave; you counter that with 12dB/octave EQ boost below Fb. But doing so requires a lot of power- sixteen times the watts for each octave below Fb- and a lot of power handling from the drivers to take that kind of inout. Then when you consider that for each octave drop the driver excursion is increased by a factor of 16 you can see why doing this with small low power/low excursion drivers is not viable.
Below Fs 2 things happens with those little drivers. 1. The SPL decreases rapidly with frequency (12db/oct) and 2. Distortion increases with excursion. You need lots of drivers.
Determine what level of output you want at the bottom and how hard you really want them working at that frequency (excursion). This will tell you how many drivers you need. Then you need a 12db/oct low pass filter set at in your example 40hz, so the net response is flat up to Fs. You'll also need a high pass rumble filter at the same point so they don't go into overexcursion trying to play the lower frequency content. If you are willing to go with the extra drivers, then starting at 25-30hz would be better. If you're going to build a monster, don't go half way.
Determine what level of output you want at the bottom and how hard you really want them working at that frequency (excursion). This will tell you how many drivers you need. Then you need a 12db/oct low pass filter set at in your example 40hz, so the net response is flat up to Fs. You'll also need a high pass rumble filter at the same point so they don't go into overexcursion trying to play the lower frequency content. If you are willing to go with the extra drivers, then starting at 25-30hz would be better. If you're going to build a monster, don't go half way.
richie00boy said:
Rich:
In his original paper, Thiele gave nine setups, (he called them alignments), for speakers to properly conform to. Of course, there are an infinite number of alignments that will work, but in those pre-home computer days, Thiele chose nine examples ranging from those having extreme low Qts/ Vb much less than Vas to those with extremely high Qts/Vb much greater than Vas. He invited the reader to extrapolate for boxes that fell in between the nine alignments-but not outside the extremes of them.
The most extreme alignment of Vb being greater than Vas was the following example:
Qts=0.55
Vb=Vas X 3 (including Thiele's recommended 40% overvolume)
Fb=0.838 Fs
F3 = .6 Fs
Note that the F3 is actually below Fs and Fb in these alignments where Qts is greater than 0.4, and Vb is greater than Vas. That is why commercial units have either been Qts is 0.4 or below, and Vb is equal to or less than Vas.
This is Thiele's most extreme alignment for high Qts and Vb much greater than Vas. anything else yields unacceptable distortion.
The audio writer David Weems, who leans toward DIY and not so much toward high end standards, said that he built a box with a woofer of 0.6 Qts and a correspondingly large Vb/Vas ratio, and he said the distortion was very, very apparent. And Weems is not high end oriented.
An ELF alignment is something different, as I understand it. There, the Qts is 0.4 or below, and the Vb is equal to or less than the Vas. The tuning, however, is just made much lower. That setup is likely to lead to a box with low distortion, as long as your amplifier power is sufficient to drive the setup at Fb to listenable levels. Because at Fb, your sensitivity is 9 dB, 12 dB, or even more down below midpoint, instead of 3 dB down.
The whole concept becomes clear when you realize that in a ported box, the smaller your Vb is comapred to your Vas, the less distortion there will be. At the same time, you must realize that the smaller your Vb is compared to Vas, the higher your F3 is driven compaered to Fs.
Sealed boxes are another brute!
Distortion below Fs is due to Cms distortion.
You can assume that cheap drivers- less quality.But how many cheap drivers..... how much Vd.
Jbl,high quality pieces- Lower distortion!
eg: 'almighty subwoofer' Jbl2226 -ported almost one octave below its Fs still has pretty low distortion(measured)
I would only use Linkwitz transforms if i was sure i could operate the driver in its linear region(not likely at all with recent attempts to extract every millimetre of excursion out of typical moderate quality sub units)
Cheers!
Distortion below Fs is due to Cms distortion.
You can assume that cheap drivers- less quality.But how many cheap drivers..... how much Vd.
Jbl,high quality pieces- Lower distortion!
eg: 'almighty subwoofer' Jbl2226 -ported almost one octave below its Fs still has pretty low distortion(measured)
I would only use Linkwitz transforms if i was sure i could operate the driver in its linear region(not likely at all with recent attempts to extract every millimetre of excursion out of typical moderate quality sub units)
Cheers!
richie00boy said:
Yes. Thank you. I will be Emailing a moderator to have the text changed right away.
richie00boy said:
I am not certain the original post in this thread dealt with sealed boxes exclusively. So many people have taken to setups where the port is tuned well below optimum and then the whole setup is equalized that I just sort of assumed that is what the original author was doing.
I have seen the term "ELF" applied to ported boxes with lower than optimum tuning so often, I assumed that is what you meant. I don't know about URPS
kelticwizard said:
I started this thread and I guess I wasn't clear. My curiosity is actually in large open baffle with a number of smallish, high Qts drivers allowed to run FR and deal with the low bass as well but the way you all have jumped in and expanded the scope of the discussion has been educational.
One concept is just take a sheet of 3'x6' to 4'x8' and poke about 60 or 80+ holes in it and use it to store drivers. The pattern, spiral, concentric circles, or grid, and power or frequency tapering along the surface (aperture distribution) being separate issues.
Please, what is Cms distortion? Cannot 'Member Stuff?
Ditto for ELF/URPS?
RCA,
A friend of mine did something similar with 64 4's. His sealed version uses 32. Don't worry about big baffles because the rolloff below Fs is twice as steep as the OB rolloff. I'd set the baffle width to start the roll dipole rolloff just below Fs and start 18db/oct high pass filter down at the bottom.
Using EQ boost isn't going to do anything because your baffle will already be pretty small and the drivers are the limiting factor. I think I'd use the U baffle to get 6db more output. I wouldn't let them go full range because the operation will muck up the upper frequency operation.
Something I've been thinking about on your line array design is that those wings swept forward add almost nothing to the effective baffle width because it adds every little extra distance for the rear wave to travel to get to your ears. Sweeping them back at the same angle would increase the distance by 5-6 times over what it currently is. Turn it around and see how the bass changes.
A friend of mine did something similar with 64 4's. His sealed version uses 32. Don't worry about big baffles because the rolloff below Fs is twice as steep as the OB rolloff. I'd set the baffle width to start the roll dipole rolloff just below Fs and start 18db/oct high pass filter down at the bottom.
Using EQ boost isn't going to do anything because your baffle will already be pretty small and the drivers are the limiting factor. I think I'd use the U baffle to get 6db more output. I wouldn't let them go full range because the operation will muck up the upper frequency operation.
Something I've been thinking about on your line array design is that those wings swept forward add almost nothing to the effective baffle width because it adds every little extra distance for the rear wave to travel to get to your ears. Sweeping them back at the same angle would increase the distance by 5-6 times over what it currently is. Turn it around and see how the bass changes.
Cms distortion is the Spider-which does most of the restoring force action,nonlinearity
Occurs below resonance. BL distortion near resonance.
www.klippel.de
Cheers!
Occurs below resonance. BL distortion near resonance.
www.klippel.de
Cheers!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.