Transmission Line with 8" Bass Driver?

H

hulminator

Enthusiast
I'm trying to build my own speakers using off the shelf drivers, and after hearing several transmission line speakers have decided that is the direction I absolutely want to pursue. I'm fairly competent at electrical and mechanical calculations, but I have to admit the math (and the lack of quality literature) is pushing my abilities.

What I would appreciate is a simple gut check; whereas most TL designs are two way with 5.5-6.5 inch midbass drivers, I was hoping to use an 8" driver in the TL with a mid dome I already have my hands on. My concern would be focusing too much reinforcement on low bass (30-50Hz) at the expense of the midbass. I like bass, but I don't need a system that sounds like it's all subwoofer.

Basically, if I tune the TL to a quarter wavelength of the driver (34Hz) would this focus too much energy too low?

Additionally, someone posted fairly nice measurements of a TL and it showed frequencies of up to 400Hz coming out of the line (fully stuffed) at less than -3dB. At that point you've gone around multiple 360 degree circles of phase, wouldn't you get fairly awful cancellations/peaks if this is the case?
 
Swerd

Swerd

Audioholic Warlord
There are two people who regularly hang out here, TLS Guy and GranteedEV who might be useful for this.

In addition, you might look for help from Paul Kittinger. He can be found if you post your question at the Parts Express Tech Talk forum.

Paul designed the bass cabinets for the Philharmonic Audio floor standing speakers. They use an 8" woofer in a TL design:

Floorstanding~Philharmonic Audio

Bass Response

What 8" woofer are you thinking about using?
 
GranteedEV

GranteedEV

Audioholic Ninja
Basically, if I tune the TL to a quarter wavelength of the driver (34Hz) would this focus too much energy too low?
It's complicated. IT depends on factors such as the driver's own parameters and the loading of the bass into the room, as well as the size of the box. For some people, it also depends on the amplfier's output impedance. Sometimes you can design a line, or even a ported box, to emphasize maximum output, and then equalize down the knee of its curve to match the room.

Either way, there's no one right solution in my opinion, as no factor can behave independant of other factors.

I will model up the seas driver for you though later. With qts = 0.30, it's probably better suited for for an ML-TL than a Tapered TL, but I will run either, or, depending on what you want.
 
H

hulminator

Enthusiast
If you could enlighten me on the benefits/disadvantages of both, that would be incredible. As much as I've had trouble figuring out transmission lines, I've seen even less on mass loaded lines.
 
GranteedEV

GranteedEV

Audioholic Ninja
Let's start by considering what we're trying to make. A classical transmission line would not even have an open end, and is pretty pointless for our purposes. So what we're really modelling, are Quarter Wave Tubes, which resonate at a frequency defined by its length, since they are a mass of air interacting with the acoustic impedance of the larger mass of air on the other side of the open end.

A loudspeaker QWT will have a lot of upper band resonances too (since it will be excited by frequencies that are a multiple of its) and be, pretty much, very, very long (necessary for augmenting bass frequencies in loudspeakers). A tapered QWT will reduce the length necessary and also minimize that ripple. What that means is that the closed end is larger than the open end (and you can play around with taper ratios too, but generally you're pretty satisfactory using a 10:1 taper ratio). What we've further discovered is that you don't want the driver at very closed end of the line - you want to offset it by a given amount.

So what's an ML-TL?

Well, you take a QWT - either straight or tapered, and where an opening would be, you place a mass-loaded resonator - IE a passive radiator or a tuned port. Now, in addition to any augmentation the line would have provided, you're getting further augmentation from the vent.

In essense, an ML-TQWT is a vented speaker and a QWT. Not every vented speaker is an ML-TQWT - because they may not be long enough to be relevant at the frequencies near the vent's range and not every QWT is an ML-TQWT - because you need the exit to be mass-loaded. Many long format vented speakers are actually unintentionally ML-TQWTs - which means that the QWT portion has not been factored into the overall bass tuning - which is a worst case scenario.

What are the pros and cons of a typical TQWT and an ML-TQWT? My opinion:

1) Both are resonators. This means that they will augment the bass but below their cutoff frequency, lose control of the driver and result in a 24db/oct rolloff. For some QWTs this frequency is below the range where we care, so it can appear as a 2nd order slope. By stuffing a QWT to reduce harmonics, you are also reducing the Q of the line resonance, which will flatten the impedance curve to more closely resemble a sealed box (good for high output impedance tube amp users).

2) A TQWT will have a lower Q - or wider bandwidth - to its resonance. This can help higher Q drivers by giving them a smooth rolloff. But that also means that for lower Q drivers, the augmentation is limited and you don't get much output. An ML-TQWT will have a higher Q - or a narrower bandwidth for its resonance. You can get a lot of solid bass augmentation for this, but it's a bad mix with a higher Q driver as you will get peaky bass. Still, tuned right, and using the correct driver (such as the lower-qts CA22) you can have a smoothly sloping frequency response or a flat bass frequency response. However since the bandwidth is small, you just need to be careful with various factors.

3) Because of the mass loading, an ML-TQWT will generally not need to be as large as a TQWT to achieve a similar tuning frequencies.

4) In an ML-TL, you will need to toy with not only the offsetting of the driver, but the offsetting of the port, in order to achieve a smooth response.

So what's better? Some say that the stuffed TQWT has a smoother measured impulse response at the cost of its size. It's well-suited for a driver with qtc of around .36 to .38. But some who have compared them (such as the aformentioned Paul Kittinger) have said that they couldn't hear the difference if properly designed, and it really boils down to application.
 
Last edited:
H

hulminator

Enthusiast
Whew, okay one step at a time.

What initially sparked my interest in TL designs was listening to PMC speakers. They utilize a tapered transmission line (but at a much lower ratio, something more like 1.5:1)((Additionally, they simply place acoustic foam on certain faces of the TL to absorb the standing waves/high freq instead of packing)). What I heard was some of the most agile, solid, tuneful bass I've ever heard out of a speaker. Hence my interest.

The basic operation then, is to take the backwaves of the driver, and then emit them into the room roughly in phase with the front waves. The first standing wave in the tube counteracts the resonant freq of the driver(?) and then we try to minimize its harmonics.

If I'm correct so far, the part I have trouble with is conceptualizing the interaction of the port/radiator with the QWT. I understand how the QWT interacts with the speaker, and i understand how ports/radiators interact with the speaker, but to have the port/radiator act on the speaker through the QWT confuses me.

If I wanted to stick to a simple tapered QWT, I should perhaps find a driver with a higher Q, correct?
 
GranteedEV

GranteedEV

Audioholic Ninja
I think you've got the gist of it. The volt radial 3853 and 3143 have Q = 0.39 and 0.35 respectively, and i'm pretty sure those are some of the drivers which PMC uses in their large studio monitors.

I'm not saying you have to sell drivers and find others to work in a line though.

Here's a driver that would work well in a 3-way below ~600hz:

http://www.sbacoustics.com/index.php/products/woofers/8-sb23nacs45-8/
 
Last edited by a moderator:
Swerd

Swerd

Audioholic Warlord
Seas says the Qts of the CA22RNX is 0.41.

For grins, I listed a few others:

8" WoofersFs (Hz)QtsVas (L)Sens (dB)Price
Seas CA22RNX290.419789.5$102.00
Seas U22REX/P-SL290.2910390.5$148.20
SB Acoustics
SB23NRXS45-8
270.389488.5$99.05
Scan Speak
22W/8851T-00
290.2512688$337.20

<tbody>
</tbody>

The last two are the woofers Paul Kittinger used in the Philharmonic 1 & 2 (SB Acoustics) and Philharmonic 3 (Scan Speak), which both feature ML-TL cabinets.

I'd agree that "agile, solid, tuneful bass" is a good description of the ML-TL speakers I've heard.
 
H

hulminator

Enthusiast
I was planning on using scrap MDF to mock up different enclosures, I may create a ported insert that fits into the rectangular mouth of the QWT. How exactly would I do the calculations on the port diameter and length? Same as a regular ported box with the volume of the QWT equaling the volume of the box, or is it just something I have to play with?
 
GranteedEV

GranteedEV

Audioholic Ninja
The woofer mentioned earlier was the CA22RNY. I agree, the CA22RNX seems more suited.

I was planning on using scrap MDF to mock up different enclosures, I may create a ported insert that fits into the rectangular mouth of the QWT. How exactly would I do the calculations on the port diameter and length? Same as a regular ported box with the volume of the QWT equaling the volume of the box, or is it just something I have to play with?
You really shouldn't be calculating these sort of things yourself; you should be modelling them in CAD to view the consequences. On that note, an ML-TL with the port at the open end of a QWT will be mis-tuned. What you'd need to do is model the ML-TL entirely separate from a TQWT. I recommend using Martin J King's MathCAD worksheets as they will handle both of these scenarios. The worksheets have been shown for classic TLs to give the same results as Augsperger's TLwrx 3.0 software, but are more versatile.

For an ML-TL, the approach I would use is to open up a basic box design software, such as WinISD or Jeff Bagby's Woofer Box Designer to design a ported box with the desired tuning - the volume and port length/diameter of the ML-TL will be the same. I then transfer that over to the ML-TL Mathcad worksheets (they use port radius btw), and make adjustments accordingly to line length, taper, etc. You can also have slot ports.

likewise for a crossover - you will want quality CAD such as LSPCad, Passive Speaker Designer Lite, Active Crossover Designer, or Passive Crossover Designer. There are too many factors which Calculations alone won't cover.
 
Swerd

Swerd

Audioholic Warlord
I was planning on using scrap MDF to mock up different enclosures, I may create a ported insert that fits into the rectangular mouth of the QWT. How exactly would I do the calculations on the port diameter and length? Same as a regular ported box with the volume of the QWT equaling the volume of the box, or is it just something I have to play with?
GranteedEV and Paul Kittinger use specialized MathCad worksheet calculators developed by Martin J. King:

Quarter Wavelength Loudspeaker Design.

I don't understand all the details, but if you enter the T/S parameters of a woofer, this software simulates the predicted bass responses as you try the various bass TL alignments, in a trial and error fashion. It allows you to quickly arrive at an optimum design (with all dimensions including port and stuffing density) without building any test enclosures. From the designs Paul has done for Salk and Philharmonic, this seems to work very well.

For an example, see the 2nd post in this DIY thread
http://forums.audioholics.com/forums/diy-corner-tips-techniques/68531-new-diy-mtm-towers-designed-dennis-murphy-paul-kittinger.html
 
TLS Guy

TLS Guy

Seriously, I have no life.
I'm trying to build my own speakers using off the shelf drivers, and after hearing several transmission line speakers have decided that is the direction I absolutely want to pursue. I'm fairly competent at electrical and mechanical calculations, but I have to admit the math (and the lack of quality literature) is pushing my abilities.

What I would appreciate is a simple gut check; whereas most TL designs are two way with 5.5-6.5 inch midbass drivers, I was hoping to use an 8" driver in the TL with a mid dome I already have my hands on. My concern would be focusing too much reinforcement on low bass (30-50Hz) at the expense of the midbass. I like bass, but I don't need a system that sounds like it's all subwoofer.

Basically, if I tune the TL to a quarter wavelength of the driver (34Hz) would this focus too much energy too low?

Additionally, someone posted fairly nice measurements of a TL and it showed frequencies of up to 400Hz coming out of the line (fully stuffed) at less than -3dB. At that point you've gone around multiple 360 degree circles of phase, wouldn't you get fairly awful cancellations/peaks if this is the case?
I think you have some misconceptions.

First you need to understand there are a whole variety of deigns under the TL banner.

However I'm pretty hair shirt as to what I regard as a TL in the traditional sense.

So here I'm going to discuss the reverse tapered aperiodically damped line. This is what PCM build, although in third part test they appear to be under damped, as there are still two peaks of impedance.

The next issue is that you need to pay attention to the all the Thiel/Small parameters. The whole object in my view is to build an essentially non resonant system, with bass augmentation, ideally over about 1.5 octaves or so. A line with output at the port at 400 Hz only -3 db would not be a properly designed line. The object is to have system Qt to be in the 0.5 range.

Now the total system Q can never be less than the driver Qts in any speaker design. The bass cut of frequency will never be below the free air resonance of the driver, Fs. I think a driver with a Qts of around 0.3 to 0.35 is the sweet spot for TLs.

Now after the work of George L. Augspurger, we know that we have to also calculate line length and volume. You can't pull the pipe parameters and especially the pipe tuning frequency out of the air, just like you can't with a vented reflex system.

The dimensions of your pipe are going to be determined by the T/S parameters of the driver.

So you need to calculate, pipe length, Lp. pipe frequency Fp, and pipe volume Vp.

Having got to that point, you need to determine taper. I find a 3:1 taper works well. You do need to fold the pipe properly, or you will have high output of mid range frequencies. At the turns you need to keep the volume linear, with corner pieces.

You place the driver/ drivers about 1/3 the distance from the closed end to suppress the odd harmonics. (A closed pipe only generates odd harmonics).

Without damping the line will have two peaks of impedance, just like a reflex box. You add damping uniformly throughout the line, but staying a little short of the open end, until you only see one peak of impedance on measurement. Do not damp past that point.

I would recommend you read the three papers by George Augspurger before designing your line. He also has a disc for modelling the line for purchase.

You can see my TL systems via my signature. This will give you an idea of typical line geometry.



I hope this helps.
 
M

MJK

Audioholic Intern
If you are going to use a classic tapered TL with an offset driver, a high taper ratio has some advantages (> 3:1 maybe even approaching 10:1). The length will be shorter for the same 1/4 wave fundamental tuning frequency while at the same time the harmonics will be pushed higher in frequency where the fiber damping is more effective. Offsetting the driver will help reduce the ripple created by the higher harmonics. Placing the driver at 1/3 of the length will suppress the 3/4, 7/4, 11/4, ... wavelength standing waves but keep the 5/4, 9/4, ... wavelength standing waves. Placing the driver at 1/5 of the length will not suppress any of the higher harminics but will reduce the excitation and response of all these standing waves, this is a compromise that many people prefer.

If you offset the driver in a tapered TL to mitigate the 3/4 wave then for a three way design the next excited standing wave may be above the low pass crossover frequency. If done carefully, a tapered TL with an offset driver in a three way speaker can provide the TL bass you like with minimal ripple generated by higher harmonics.

If you look at the alignment tables on my site you can scope out the length of the TL for a gived tuning frequency as a function of the taper or expansion rate.

http://www.quarter-wave.com/TLs/Alignment_Tables.pdf

The alignment tables should give you a potential starting geometry for your driver in a matter of minutes. Also in the alignment tables are a number of studies that trade-off TL design choices, this might provide a feel for the options available in a TL design.
 
Last edited by a moderator:
jinjuku

jinjuku

Moderator
Thanks Martin. Your site and writings have been one of my biblical references to learning about TL and ML-TL.
 
Last edited:
TheWarrior

TheWarrior

Audioholic Ninja
I've bookmarked that site now, didn't read its entirety yet, but those are the equations I've been looking for!
 
H

hulminator

Enthusiast
Thanks Martin, I've been using your site extensively and am planning on purchasing the mathcad worksheets anyway, just because past the speakers I'm building, it fascinates me to play with all the parameters. Why do you say that the 1/5th placement is a compromise people prefer, over the 1/3rd length?
 
M

MJK

Audioholic Intern
I typed to fast this morning without the proper thinking. Offsetting the driver can help mitigate the higher modes in a TL. The logic is provided below while referencing the graphic at the top of my home page Quarter Wavelength Loudspeaker Design showing the pressure (top plot) and velocity (bottom plot) profiles.

The standing waves in a TL are the fundamental 1/4 wave and the harmonics 3/4, 5/4, 7/4, 9/4, ..... waves. These modes receive the maximum excitation if the driver is placed at the closed end of the TL. In the plot on my homepage this is at x = 0.

If the driver is placed at at x = L / 3, then the 3/4, 7/4, 11/4 ... modes will not be excited. The driver is at the location of maximum velocity for these standing waves, see the blue curve in the lower of the two plots.

If the driver is placed at at x = L / 5, then the 5/4, 9/4, 13/4 ... modes will not be excited. The driver is at the location of maximum velocity for these standing waves, see the brown curve in the lower of the two plots.

So placing the driver at the maximum velocity location eliminates the standing wave. A compromise is to place the driver between L/3 and L/5 so it is close to the maximum velocity position for all of the standing wave harmonics. The fundamental will still be excited to produce the bass while the harmonics will be almost not excited to greatly reduce the ripple.

Hope that clears thing up, my apologies for any confusion.
 
H

hulminator

Enthusiast
TLS Guy, I was just about set on using SEAS CA22RNX as my bass unit, then I saw that the distributor for SEAS here also carries Volt, which is what PMC uses in their larger monitors. Can you offer any advice on this front? I've played around with the idea of using one of the 8-10" units as my bass driver, possibly using the 8" midbass in a two way, or even going crazy and using the ribbed 12". They seem to have a very good reputation, but so does SEAS.
 
GranteedEV

GranteedEV

Audioholic Ninja
I'm sure the Volt drivers are good, but they're not known, tested quantities.

If I were picking a 10" driver, the usher 1001B would be my choice.
 

Latest posts

newsletter

  • RBHsound.com
  • BlueJeansCable.com
  • SVS Sound Subwoofers
  • Experience the Martin Logan Montis
Top