Transmission line subwoofer question

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I've decided to build a transmission line subwoofer, but I have a few questions about driver selection and pipe design before I start ordering parts.

I liked the idea of using a sonotube to build a TL subwoofer however after further reasurch I decided that the pipe should have a tapered design. My ceiling is just under 8 feet so I'm assuming I don't want to build the tube much longer than 7 feet, or so. I'm planning on building it from MDF with 8-16 sides and finishing it with veneer. The math on the building dimensions is pretty simple and I think I have the carpentry skills and tools to cut the pieces acuratly (16 sides will be tricky).

I have been using Martin King's Mathcad worksheets for modeling driver and tube combinations.

Ok, onto the barrage of questions. (I think some of my confusion stems from differences between TL fullrange and subwoofer design.)

I have been using Martin King's MathCAD worksheets for modeling driver and tube combinations.

Ok, onto the barrage of questions. (I think some of my confusion stems from differences between TL full range and subwoofer design.)

How large of a taper ratio (cross-section area at driver to cross-section area at opening) is acceptable? I have seen suggestions at 2:1 and 3:1. Are there any negatives to increasing this to say 8:1 as an example?

I have seen recommendations that the cross-section area at the driver be as close to Sd as possible. Are there negative aspects to increasing this dimension, and if so what are they? Also, is there any rules of thumb for what the cross-section area at the opening should be and what are the consequences of deviating form this?

Should tube volume be taken directly into consideration during design or just indirectly through line length and cross-section areas at the driver and the opening?

My thought on suitable driver parameters are high Qts, low VAS, and fs suitable to the line length accounting for line taper and stuffing density. Any objections/corrections to this?

I'm still not sure if I want to use a single woofer, probably a 10", or go with a push push configuration. It will be in a 14"x12"x8" room so one woofer would probably be enough, but the push push definitely has its drawing features.

Well, I think that’s it for now. Thanks in advance for the help.

-Street
 
Some notes:

1) It's possible to make a tapered design with nested Sonotubes. See last month's AudioXpress.

2) For a TL, you want a low Qts, no more than .4 or so.

3) For the taper, Augspurger recommends a ratio up to 1:4. Martin King's charts to to 10.

4) In that small room, I would think one 10" woofer in a small box would be sufficient to rattle the walls.

5) Ten days ago, I didn't know my Vas from a hole in the ground. You can find an enormous amount of info on the web -- and here. Google and search.

6) Doubt everything I say, and most of what I write.
 
Ok I think I have found answers to many of my questions and have come up with a few new ones too.

I think I will likely use a high taper ratio of 7:1 to 10:1. After doing some modeling in mathcad I don’t see any problems with such ratios.

X-section area has no correlation to Sd, but is based on the Vas. What is the relationship between this? Or should I just tweak x-sections in the mathcad worksheet until I’m happy with the result (what I have been doing so far)?

I think I’m pretty close with desired driver parameters with 0.5<Qts<1.0, low to medium Vas and appropriate fs. Dave I think I recall reading somewhere that a Qts no more than 0.4 is more a guidline for full range TLs and that subwoofers do beter with higher Qts. I am far from certain on this though. Any input would be appreciated.

Martin advises that the first acoustic impedance peak should be at or below the driver resonance ( http://www.diyaudio.com/forums/showthread.php?postid=155469#post155469 ). How much below the driver resonance is acceptable? Also, what phase characteristics should I be paying attention to when playing with the mathcad worksheets?

I have modeled a few Lambda drivers which look very appealing. How large of an effect does the faraday ring have on the sound quality of the driver? I was looking to spend about what the LE motor versions (no faraday ring) cost however if the improvement with a single faraday ring is substantial I might be able to scrape together the cash.

-Street
 
frugal-phile™
Joined 2001
Paid Member
Street said:
I think I’m pretty close with desired driver parameters with 0.5<Qts<1.0, low to medium Vas and appropriate fs. Dave I think I recall reading somewhere that a Qts no more than 0.4 is more a guidline for full range TLs and that subwoofers do beter with higher Qts. I am far from certain on this though.

I don't know where the Qts <0.4 comes from ... i'm always on the lookout for drivers with Q higher than this for TLs.

How large of an effect does the faraday ring have on the sound quality of the driver?

Quite a bit.

dave
 
This is from memory, so you need to model it to confirm:

Peerless 850146
So=4*Sd
Sm=0.5*Sd
L=120"

I don't have a feel for Martin's tables, since I've used the worksheets enough to plung right in. Augspurger's table work better with larger drivers, since his F3 will always be higher than Fs. You don't have to have Fp>Fs. I never do with small (8" and less) drivers.

Bob
 
This is from memory, so you need to model it to confirm:

Peerless 850146
So=4*Sd
Sm=0.5*Sd
L=120"

I modeled this driver using your dimensions. The frequency response looks decent, however, the first peak in the acoustic impedance is around 12 Hz or so. Should this not be matched to the free air resonance of the driver? In this case 22.6 Hz from the peerless spec sheet. Is this difference acceptable? I have been trying to keep these matched as closely as possible, normally within one or two Hz. If a larger difference does not pose any problems it would give me more freedom when modeling.

-Street
 

GM

Member
Joined 2003
>How large of a taper ratio (cross-section area at driver to cross-section area at opening) is acceptable? I have seen suggestions at 2:1 and 3:1. Are there any negatives to increasing this to say 8:1 as an example?

====

Taper ratio is a function of how long the pipe can be and how much boost you want. I'm not much of a TL fan, but some drivers just won't work in an otherwise vented alignment, so when all else fails.......

====
>I have seen recommendations that the cross-section area at the driver be as close to Sd as possible. Are there negative aspects to increasing this dimension, and if so what are they? Also, is there any rules of thumb for what the cross-section area at the opening should be and what are the consequences of deviating form this?

>Should tube volume be taken directly into consideration during design or just indirectly through line length and cross-section areas at the driver and the opening?

>My thought on suitable driver parameters are high Qts, low VAS, and fs suitable to the line length accounting for line taper and stuffing density. Any objections/corrections to this?

====

Drivers with low Vas for the intended BW don't do well in vented designs once you move away from a golden ratio cab due to insufficient area around the driver, causing peaking at Fb that no reasonable amount of stuffing will damp.

A driver wants to 'feel' a certain amount of acoustic mass to perform properly, so Vb for the intended BW is where I start. I look at what the T/S max flat is for a starting point, with 0.707 to 1.414x Fs being the BW of max performance. This pretty much dictates a Qt around 0.4 as the ideal, just like a reflex or ML-TL design. IMO, low Q drivers are better off in Daline type pipes, while high Q drivers need larger, heavily stuffed pipes to acoustically lower their apparent Qt.

Once the Vb/Fp is chosen, the taper increases with decreasing pipe length. A pipe generally has the smoothest response when the driver is at some point down its side, and the greater the taper, the further down/up the pipe it must be towards the big end. If it's on the end, then the taper should terminate in a very small terminus. This is a good way to get virtually IB response in a smaller package.

Anyway, unless you're dead set on a TL, for the 850146, a minimally stuffed ML-TL looks pretty good and isn't large (by my standards anyway). Punch these numbers into the MLTQWT worksheet to view it:

L = 50.63"

S0 = Sd*2.0326

SL = Sd*2.0326

density = 0.25lbs/ft^3

rport = 1.88" (minimum)

Lport = 8" (of course this is adjustable, I just chose an alignment that normally works well in a typical room)

X0 = L*0.4367

Of course for pounding LF you'll either need a true subwoofer driver or two of these as part of the mains, or double S0/SL/vents and do a bipole sub.

HTH,

GM
 
Street,

Don't look at the impeadance peak, look at the FR peak at cut-off. Now, a problem with the strawman I gave you is that it doesn't take room lift into account. But, it does show you wat is possible, if not exactly practical. You have to wipe your mind clean of all of the rules-of-thumb and old-wives tales about TL's. Reread GM's post. The short, fat pipes IS the way to go.

Bob
 
The designing is comming along pretty good thanks to the information you all have given me. I was wondering if anyone knows of a good resource on room gain. I have found plenty of information on calculating room modes but very little on translating this to room gain. Due to the relativly small size of my room (12'x14'x8') room gain will likely have a significant impact on FR which needs to be compensated for.

-Street
 
I think I'm pretty much decided on my design and I'm looking for a little feedback on what I've come up with. I'm going to use the Lambda SB10 driver with the single faraday ring motor, which can be seen at http://www.lambdacoustics.com/drivers/SB10.html.

My transmission line will have a quasi-cone shape with x-sectional area of 2800 cm^2 at the driver end and 150 cm^2 at the open end. I have decided on using 8 sides made of MDF. The driver will be mounted on the bottom facing downwards and slightly off center. The line length will be 78 inches. I would have made the line length a bit longer but ceilings have a bad habit of getting in the way. This will be stuffed to taste although probably close to 0.7 lbs/ft^3 from modeling.

I'm planning on finishing it with two different types of veneer, which will alternate around the 8 sides. I will likely use walnut and kewazinga for this so although it will be large it should be an attractive piece of furniture.

Oh, I also need to buy an amp to power this. Any suggestions on an amp or diy kit that would match well with this driver?

Let me know what you think.

-Street
 
>I was wondering if anyone knows of a good resource on room gain. I have found plenty of information on calculating room modes but very little on translating this to room gain. Due to the relativly small size of my room (12'x14'x8') room gain will likely have a significant impact on FR which needs to be compensated for.

====

In a perfectly sealed room that's built like a vault, room gain begins at ~the average of all the first axial modes, rising at 12dB/octave with decreasing frequency, so the ideal speaker would be a sealed cab with a complementary curve to yield a flat amplitude response.

So theoretically, your room gain would begin ~at:

~(1130/2)/(12*14*8)^0.33 = ~52.44Hz and ramp up to +12dB/~26.22Hz, etc..

So much for theory. In reality, most closed in HIFI/HT rooms have a choppy looking response that averages 6-9dB down to around 1WL/pi of the longest dim (~30Hz in your room) and rolling off dependent on how lossy the room's construction is. If there's an open doorway, then the room becomes part of a tuned multichamber resonant ckt, making for some interesting room gain measurements. Toss in furniture, etc., and little wonder you couldn't find much info on the subject. Still, it's best to design a room around theory whenever possible.

FWIW, I prefer to design around very little gain since distortion wise it's better to EQ it out than boost it in.

====

>I think I'm pretty much decided on my design and I'm looking for a little feedback on what I've come up with. I'm going to use the Lambda SB10 driver with the single faraday ring motor, which can be seen at http://www.lambdacoustics.com/drivers/SB10.html.

My transmission line will have a quasi-cone shape with x-sectional area of 2800 cm^2 at the driver end and 150 cm^2 at the open end.

> I have decided on using 8 sides made of MDF. The driver will be mounted on the bottom facing downwards and slightly off center. The line length will be 78 inches. I would have made the line length a bit longer but ceilings have a bad habit of getting in the way. This will be stuffed to taste although probably close to 0.7 lbs/ft^3 from modeling.


====

Hmm, a taper ratio of ~18.67:1 in such a long pipe will get you a severely underdamped (ringing) response (poor transient response). 8-10:1 is usually the limit and the pipe's Vb needs to be ~ = or < the T/S max flat alignment for good performance.

====


>Oh, I also need to buy an amp to power this. Any suggestions on an amp or diy kit that would match well with this driver?

Let me know what you think.

====

With such a low efficiency driver, the amp needs to be rated at least its peak transient rating, or 1kW/8ohms for this one. Any of the better pro sound amps from Mackie, Crown, etc., will be fine since HF BW/quality isn't an issue. The lower its damping factor and cutoff frequency the better IMO. My bang/buck fave was the QSC 'USA' series, but I don't keep up with this stuff anymore, so don't know if they're still available.

HTH,

GM
 
You're welcome! Messrs Thiele and Small (T/S) used electrical filter theory to arrive at a way to design cabs based on their electro-mechanical properties to a much greater degree than previously. They calculate the aneochoically maximally flat response, i.e. F3 and Fb is the same frequency. Any other alignment, whether over/underdamped, F3 will be >Fb. I was referring to the Vb required to make this alignment, and the one that most speaker programs default to. With published specs, the SB10 vented defaults to 5.7ft^3/17hz Fb in BoxPlot.

Different programs yield slightly different numbers due to whether they used the original formulas or which one of the lumped constants versions.

GM
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.