Hi camplo, could you post some hornresp shots of a tuning example you think is superior in SQ? And what you mean by tuning half below intended cutoff?
Every time I've used hornresp to model a bass-reflex for lowest possible tuning, I end up with big excursion problems, requiring a lower drive voltage to survive, way below PMax.
Or I should say requiring both a lower voltage and also a high pass filter...usually steep... that brings its own group delay.
Personally, I view SPL as an important component of SQ...one that contributes greatly to both realism and enjoyment.
Put another way, I will gladly accept some group delay of unknown audibility, to gain an easy to hear improvement in SPL.
That's a real world SQ improvement imo.
But in the case of a BR, I think of the vents as adding low-end extension vs using the same driver sealed.....not as adding any SPL.
So I can't liken vents to a horn which as others have pointed out, does increase efficiency.
Increased low-end extension also greatly improves SQ imo, even if it comes with its increased group delay ime.
My Danley designed Labhorns, have the greatest SPL and lowest group delay of any subs I've built. They should, being essentially just a sealed with a horn attached, huh?
Too bad the design calls for a block of 6 used together for optimum results, which equates to a round horn dia of 7ft !
I've played with eliminating nearly all group delay on sealed, vented, and the FLH's, via linearizing phase.
In outdoor listening I believe it can make a difference. Indoors, I can't tell beans.
But this is comparing well designed boxes against themselves, linear phase vs not. If the subs were poor designs with lots of group delay, then maybe indoors could be heard too....Dunno...
Camplo,
your main fallacy lies in the comparison of horns that are designed with a lower cutoff (and that definitely WILL have a lower cutoff if done properly) and the tuning of a given driver with a lower fb than usual. You compare two things that are not the same. It is not even apples to oranges. The cutoff frequency is not dependant on fb alone.
But if you go a bit furter and make an apples to apples comparison and build a reflex speaker whose cutoff frequency is an octave lower than the lowest frequency you usually feed it with - then yes you will improve performance at your lowest intended frequency. But that comes at the cost of decreased efficiency and the use of a different driver as well.
There is some small truth to your view however. You can actually use low tunings that fall off less sharply than an ordinary reflex tuning and that can have better transient Response. I for instance use my drivers in a Vb>Vas and tuning with fb<fs. The result is a response that starts to drop early and is around -6dB @ 30 Hz.
Apart from that there are not many situations where one actually tunes fb=fs like you seem to assume as normal. More often one uses Vb<Vas and f>fs because the TSPs of most of the woofers used for reflex tuning dictate this when a flat response is aimed for. And don't forget that the resonance of a typical woofer doesn't ring like a bell because it is usually very well damped due to the low Qts values. And reflex isn't therefore the coupling of two bells either. It is the higher filter order of a bass-reflex compared to a closed box that causes the increased group-delay distortion.
But you may probably like some tuning idea that was once published in EW+WW, where a driver was in a box that was sized as if it was a closed box with a Qtc of 0.707. But then a port was added with a tuning an octave below fc (which is definitley not the same as an octave below fs) and an LTF circuit was used to get the desired response. The advantage was the reduced excursion in the octave below fc.
Regards
Charles
your main fallacy lies in the comparison of horns that are designed with a lower cutoff (and that definitely WILL have a lower cutoff if done properly) and the tuning of a given driver with a lower fb than usual. You compare two things that are not the same. It is not even apples to oranges. The cutoff frequency is not dependant on fb alone.
But if you go a bit furter and make an apples to apples comparison and build a reflex speaker whose cutoff frequency is an octave lower than the lowest frequency you usually feed it with - then yes you will improve performance at your lowest intended frequency. But that comes at the cost of decreased efficiency and the use of a different driver as well.
There is some small truth to your view however. You can actually use low tunings that fall off less sharply than an ordinary reflex tuning and that can have better transient Response. I for instance use my drivers in a Vb>Vas and tuning with fb<fs. The result is a response that starts to drop early and is around -6dB @ 30 Hz.
Apart from that there are not many situations where one actually tunes fb=fs like you seem to assume as normal. More often one uses Vb<Vas and f>fs because the TSPs of most of the woofers used for reflex tuning dictate this when a flat response is aimed for. And don't forget that the resonance of a typical woofer doesn't ring like a bell because it is usually very well damped due to the low Qts values. And reflex isn't therefore the coupling of two bells either. It is the higher filter order of a bass-reflex compared to a closed box that causes the increased group-delay distortion.
But you may probably like some tuning idea that was once published in EW+WW, where a driver was in a box that was sized as if it was a closed box with a Qtc of 0.707. But then a port was added with a tuning an octave below fc (which is definitley not the same as an octave below fs) and an LTF circuit was used to get the desired response. The advantage was the reduced excursion in the octave below fc.
Regards
Charles
Last edited:
Actually some are doing that (for vented) and the alignment is called LLT for large, low tuned. Typically HT subs with high diamenter massive cones and mega excursion in huge boxes tuned to 10-15 hz. This is well below audibility and into the feel region so 30 Hz and up are like sealed but you can still shake the house
~420liter tuned to fs/24hz, 15hz/octave below desired cutoff, Sealed, and 17hz tuning.
Hi camplo, i guess the reason you provided 4 examples is to show the continuum available in the tradeoff between group delay vs freq response.
If i look at your first example (fs/24hz) vs the last (fs/17hz), the tradeoff seems resonable and worth exploring. About a 1dB drop in response at 30Hz for the 17Hz tuning, in exchange for almost cutting group delay in half 30Hz, 7ms vs 13ms...(scales are way different on delay graphs??)
Like i mentioned earlier, when I've tried to use hornresp to model this tuning tradeoff, it always fell apart when looking at displacement and power handling.
Would you post the hornresp Input Parameters and Driver Parameters for the fs/17Hz example ?
So I can see how displacement holds up for it at large signal levels.
big thx
HornResp won't allow you to control scaling...I know. I can't say exactly what my programming was and its a miracle I remembered, the next wake up, what the tunings were, to a graph I created at 6am post 24hrs. its the 2x 18h+ driver, s1 5000, s2 5000, s3 660, s4 660, l12 72.60 l23 .1, l34 109.2...this is about the same box I used for the graphs. Slide the length of the port to raise tuning.
I did scrutinize excursion. At 30hz, excursion is un bother by the tunings below it. 24hz tuning was not influential enough to 30hz, that 15hz made no major difference at 30hz, as well....which is kinda what was catching my eye to begin with. I'm always staring at 30hz because thats my design cutoff.
2.1MM at 30hz/100db with one driver. I'm modelling for 2.
Also, view the -30db timings in those graphs. 50hz-40hz is an interesting area. If you imagined sound as color...imagine a trail of lingering fading colors blurring the true colors of the view. Sounds kinda like mushrooms....the room is worse behaved, but lets not have this tendency be a large part of direct energy.
I did scrutinize excursion. At 30hz, excursion is un bother by the tunings below it. 24hz tuning was not influential enough to 30hz, that 15hz made no major difference at 30hz, as well....which is kinda what was catching my eye to begin with. I'm always staring at 30hz because thats my design cutoff.
2.1MM at 30hz/100db with one driver. I'm modelling for 2.
Also, view the -30db timings in those graphs. 50hz-40hz is an interesting area. If you imagined sound as color...imagine a trail of lingering fading colors blurring the true colors of the view. Sounds kinda like mushrooms....the room is worse behaved, but lets not have this tendency be a large part of direct energy.
Last edited:
Oh ok, I thought you were simming simple bass-reflexes...my bad.
Here's what I do for whatever sub I'm modeling, once small signal (1 watt) looks good.
Raise Eg to PMax and study for problems...displacement, port velocity, loss of overall spl, etc.
If you do that with no problems, sounds like a winner to me !
Regarding looking at around -30dB...I assume you mean the spectrograms?
Not sure you can read that much into them.
But then again, that's my focus...keep it simple and keep my eye on what I can directly change with acoustic design and processing.. ie mag and phase...
Here's what I do for whatever sub I'm modeling, once small signal (1 watt) looks good.
Raise Eg to PMax and study for problems...displacement, port velocity, loss of overall spl, etc.
If you do that with no problems, sounds like a winner to me !
Regarding looking at around -30dB...I assume you mean the spectrograms?
Not sure you can read that much into them.
But then again, that's my focus...keep it simple and keep my eye on what I can directly change with acoustic design and processing.. ie mag and phase...
When you increase the box volume less power will be needed to reach the same spl so the power handling will go down in one sense because you can no longer feed the same power in without exceeding excursion, on the other hand you need less power to get to the same place within limits depending on the parameters of the driver being used.
Here are some simulations of the TD18, red is a fairly small sealed box with a Q of 0.7, orange is an EBS alignment tuned to 33Hz, bigger box than sealed, and yellow is 420 litres tuned to 15Hz. When plotted on the same graph it is easier to see the difference.
The big box can go lower but at reduced maximum level. By changing the volume and tuning the rolloff can be adjusted anywhere in between something like sealed and drop off a cliff vented.
If maximum output to 30Hz is desired then standard EBS gets you there.
Attachments
Here is the group delay graph, I didn't post it before because it seemed pointless. The more you push the venting the more group delay you get.
As for the spectrogram BassBox does not have one just turn the group delay chart around and imagine the colours
The "fingers" are reflections that hornresp calculates, in anything other than a bare box they won't be there anyway. You can also turn it off in hornresp with the mask / unmask reflections control.
The main point was to overlay the responses on the same graph for easy comparison.
As for the spectrogram BassBox does not have one just turn the group delay chart around and imagine the colours
The "fingers" are reflections that hornresp calculates, in anything other than a bare box they won't be there anyway. You can also turn it off in hornresp with the mask / unmask reflections control.
The main point was to overlay the responses on the same graph for easy comparison.
Attachments
'Famous'? Don't know if there's any since there's 6th order assisted and any passive alignment with a bigger box, lower tuning than a T/S max flat alignment qualifies.
GM
Correct.
It's what me and numerous others keep recommending, though I prefer a bit over damped since the room will 'ring it up' some if not designed not to.
Right, historically it was about getting sufficient mid-bass efficiency and not allowing the driver to over ex-curse at rated power.
GM
Just to clarify, is the specified input power of 50 watts shown in the attachment with reference to a nominal 8 ohm load, giving a constant input voltage Eg of 20V?
Attachments
Thanks for the clarification. The 50W and 20V chart results look identical, which confirms that a nominal 8 ohms is being assumed (as distinct from the actual Re value of the driver).
As for the spectrogram BassBox does not have one just turn the group delay chart around and imagine the colours
Of course you are joking because they do not display the same information