I am planning to build a TL subwoofer for my home cinema. The driver I have is 12", but with a very low compliance. Fs is 27Hz.
From the literature I understand stuffing is extremely difficult to model. So it will be hard to get it right, or assess that it is right.
I know stuffing is used to slow down sound speed and make the TL appear longer due to the slower speed. Which also means the correct tuning of the TL is dependent on exactly the right amount of stuffing.
Space is not an issue. I will design the TL as a single folded tapered line. The maximum dimension is 2.5 meters. So I could achieve a 4.5 meters line length. That is more than I need.
If I use no stuffing at all, I can calculate the line length exactly for the desired resonance frequency. But if it is slightly too long, nothing can be done.
If I plan on using stuffing, I must construct a shorter line and adjust the tuning my stuffing. Which can be complicated.
A third option could be to construct the line slightly too short and fine tune it by a minimum amount of stuffing.
I will put some damping material (BAF wedding) on the inside to avoid reflections.
What is the opinion of the experts: would it be right to try to avoid stuffing at all, at most a minimum amount. Or calculate the line based on a generous amount of stuffing as per most recommendations?
From the literature I understand stuffing is extremely difficult to model. So it will be hard to get it right, or assess that it is right.
I know stuffing is used to slow down sound speed and make the TL appear longer due to the slower speed. Which also means the correct tuning of the TL is dependent on exactly the right amount of stuffing.
Space is not an issue. I will design the TL as a single folded tapered line. The maximum dimension is 2.5 meters. So I could achieve a 4.5 meters line length. That is more than I need.
If I use no stuffing at all, I can calculate the line length exactly for the desired resonance frequency. But if it is slightly too long, nothing can be done.
If I plan on using stuffing, I must construct a shorter line and adjust the tuning my stuffing. Which can be complicated.
A third option could be to construct the line slightly too short and fine tune it by a minimum amount of stuffing.
I will put some damping material (BAF wedding) on the inside to avoid reflections.
What is the opinion of the experts: would it be right to try to avoid stuffing at all, at most a minimum amount. Or calculate the line based on a generous amount of stuffing as per most recommendations?
One of the members here goes by pkitt and knows TL design more than most others. He'll probably see your post and reply.
But to my very limited knowledge of TL, older TL theory speaks of stuffing slowing things down. However, newer and more relevant theory suggests that it reduces very little.
But to my very limited knowledge of TL, older TL theory speaks of stuffing slowing things down. However, newer and more relevant theory suggests that it reduces very little.
Strictly for sub duty, no stuffing for any properly designed sealed alignment beyond internal reflections to keep it from sounding 'hollow' works for me.
Well.....not really true for the TL design routines currently in vogue, but between no stuffing and especially since using a low compliance [horn] driver.......take a 'page' from Bose's Wave Cannon design and use a 2:1 CR [pipe area = Sd/2], which does allow a simple ~34400/4 line length when the driver offset is at 0.333 of closed end.
What with room gain and an acoustically damped pipe and assuming the driver has decent power handling down low, consider tuning a half octave below Fs to get down to the THX 20 Hz reference and use a bit of EQ if more mid-bass 'punch' is desired.
GM
edit: FWIW, Hornresp 'says' that using a 1:1 CR only shifts it lower a Hz or two, but adds enough extra gain in the 1st octave to justify the extra bulk.
Last edited:
If you're building a TL for subwoofer duty, aim to keep stuffing to a minimum, though the amount of stuffing required will really depend on the characteristics of your design. Hornresp is pretty decent at emulating the impact of stuffing. I would suggest that keep an eye on the simulated Fb in Horresp for the build (including stuffing), then stuff to meet that target Fb.
More notes about using Hornresp to design a TL can be seen at my site in my signature.
Just in case you'd like to see some data....
17 foot pipe sub 12-230 Hz ±5dB
See post #2, the second and third charts. With light stuffing as in the charts, that labyrinth has some pretty noxious bumps and the heavier stuffing reduces them a lot.
The driver impedance plots are quite dramatic too:
Long pipe to sequester rear wave
See post #14.
Some TL builders mount the driver downstream in the pipe to upset pipe resonances. I'd like to see the plots although hard to imagine anybody having trustworthy before-and-after evidence.
B.
17 foot pipe sub 12-230 Hz ±5dB
See post #2, the second and third charts. With light stuffing as in the charts, that labyrinth has some pretty noxious bumps and the heavier stuffing reduces them a lot.
The driver impedance plots are quite dramatic too:
Long pipe to sequester rear wave
See post #14.
Some TL builders mount the driver downstream in the pipe to upset pipe resonances. I'd like to see the plots although hard to imagine anybody having trustworthy before-and-after evidence.
B.
Last edited:
IMO it's not very wise to commit time and resources to take on a TL build without first at least simulating what its response might be liket, and the sims say that offset driver mounting and reduce or eliminate the first cancellation notch, and some stuffing can be used to smooth out the passband if necessary. Hornresp sims are pretty accurate (once you build what you sim, of course). It's been basically bang on for my POC1 to POC5 builds. Two of those builds were modeled and built as offset TLs.
Well then, it should be really easy to find and post REW measurements of the success of the offset strategy, when we find somebody who has done before-and-after plots, like the kind I posted with stuffing.*
OK to endorse Hornresp for TLs. But let's see the tests.
B.
*possibly somebody with a little bit of scientific curiosity temporarily blocked or filled the offset section (might be one-third of the pipe length) and ran the tests.
Blocking or filling the offset will change the length and volume and therefore the tuning and overall response of the TL. It is therefore would not be a valid way of doing the comparison.
IMO all that's important is to show that the measurements prove that the offset approach eliminates the notch, which we already know exists with end-loaded TLs. There is really no need to do a before/after test, which would require building a box that would allow significant repositioning of the woofer along the path.
IMO all that's important is to show that the measurements prove that the offset approach eliminates the notch, which we already know exists with end-loaded TLs. There is really no need to do a before/after test, which would require building a box that would allow significant repositioning of the woofer along the path.
Thanks all for your replies. As it seems, there is a slight bias for not stuffing.
Yes, you are right, I should not start building without modelling first. Thing is, I totally neglected the existence of modern simulating software. I have been out of TL building since the early 1980-ies. Back then I interacted with computers with punch cards.
So I'll take a look at Hornresp and hope I can avoid installing a Windows machine. Linux-only guy here.
Yes, you are right, I should not start building without modelling first. Thing is, I totally neglected the existence of modern simulating software. I have been out of TL building since the early 1980-ies. Back then I interacted with computers with punch cards.
So I'll take a look at Hornresp and hope I can avoid installing a Windows machine. Linux-only guy here.
Please let us know if you manage to run Hornresp on Wine. Like you, I'm Linux-only, and don't really want to pull out my old Win XP CDs and install Windows just to tinker with one or two rarely-used bits of software.
-Gnobuddy
@Gnobuddy:
There was some delay in the project, but I just finished installing Wine to run Hornresp.
Since you are Linux only I trust you know how to install Wine. My first attempt to install from WineHQ failed on Debian Jessie 9.4
However, installing from the Debian repos was not a problem. I installed wine and wine32.
There was just an error with installing libuuid. Which failed because I did not have the libuuid user and group. And the libuuid installer did not install it automatically. However I think this is related with using nis as I do. I added libuuid manually in both /etc/passwd and /etc/group and that solved the error.
Then Wine needs the courier font. Install it by ttf-mscorefonts-installer.
Then Hornresp.exe just works.
jlinkels
There was some delay in the project, but I just finished installing Wine to run Hornresp.
Since you are Linux only I trust you know how to install Wine. My first attempt to install from WineHQ failed on Debian Jessie 9.4
However, installing from the Debian repos was not a problem. I installed wine and wine32.
There was just an error with installing libuuid. Which failed because I did not have the libuuid user and group. And the libuuid installer did not install it automatically. However I think this is related with using nis as I do. I added libuuid manually in both /etc/passwd and /etc/group and that solved the error.
Then Wine needs the courier font. Install it by ttf-mscorefonts-installer.
Then Hornresp.exe just works.
jlinkels
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.