How do you kill midrange spikes on transmission lines? Is it done with stuffing? Taper? I am considering a fullrange speaker and wondering what route to take. I know quarter-wave pipes have spikes at 1/4wl, 3/4wl, 5/4wl, 7/4wl, etc. so I wonder what you do to prevent midrange smear. All comments welcome.

Doug

Use the proper Qts driver.
ron

Quarter-wave resonant cabinets have acquired a whole litany of terms, none of which make any sense. First of all, what we call transmission line speakers bare no resemblance to electrical transmission lines or audio wave guides, my understanding of which can be written on the head of a pin. The TL speaker works on the basis of standing waves of odd-order harmonics -- a stopped pipe. The classic TL speaker "fixes" the standing waves by stuffing the bejesus out of the pipe, leaving for all practical purposes an IB. BTW, this is not all bad, since a true IB will give you some of the best bass you will ever hear if you can stand the real estate required.

Martin King invented the term Mass Loaded Transmission Line -- MLTL -- a few years ago, which addresses the physics of the problem, but leaves the mathematically challenged bewildered as to the connection between TL's and a 40" pipe that supports an Fp of 40 Hz.This has generated a long, loud and misinformed argument that MLTL's are simply BR's in disguise. They are not -- the physics are entirely different. If you are wondering why your new tower BR tunes way too low and has lumpy mid-bass, I have the answer for you.

OK, now to your question. The tuning frequency of a quarter-wave pipe is dependent on the length, cross-sectional area, taper and port dimensions. The greater the taper ratio, the shorter the pipe. TQWT's (Voigt pipes, pointy at the top) are the longest, conventional TL's the shortest. Tuning frequency can be lowered by reducing the size of the port and increasing its length. If you can't run Martin's model and choose to cut-and-try, you chances of getting it right are minimal.

A quarter-wave pipe will have a series of harmonic in the pipe output. An end-loaded conventional TL is the worse, followed closely the TQWT. The 3rd or 5th harmonic can be easily suppressed by by placing the driver at the harmonic node. Normal design procedure is to place the driver at the 3rd harmonic in a TQWT and at the 5th harmonic in everything else. The physical dimensions of the pipe make this convenient. With one of these harmonics suppressed, the bottom end of the frequency response smooths dramatically. Further smoothing of the mid-bass is accomplished by moving the port up the pipe a little. Again, this must be modeled. There is no chance of guessing at it. If you have done the design right, you will need very little stuffing. My designs use no stuffing at all. I rely on a fiberglass lining to bring the frequency response to better than +/- 3 dB throughout the bass and mid-range.

Bob

Why not use ported reflex instead? As I understand it, the helmholtz property makes no harmonics. It is like a spring or a pendulum with only one primary and no harmonics. What do you think?

Doug

Thanks for your answer. I guess its a delicate balance getting the right amount of stuffing to remove harmonics without taking life from the midrange. Sounds iffy.

Doug

The current trend in speaker aesthetics is the tower speaker, driver at ear level, baffle as narrow as possible and still hold the driver. So.... We use one of the calc programs, WinISD, Unibox, whatever, design a BR, then adjust the dimensions to fit the form factor. Voila, a tower BR. But wait! The tuning is all wrong. Got to shorten the port by half. And the mid-bass is lumpy. Got to add stuffing.

Mother Nature is a b...., and she makes the rules. Once the ratio of cross-sectional area to length reaches a certain number, and I don't know what that is, the Helmholtz resonance breaks down and quarter-wave modes set in. I do know that the normal range of dimensions for tower BR's guarantees that they are not BR's.

Here's my Fostex FE167E BR:

24 liters, tuned to 60 Hz.

Here's my Fostex FE167E quarter-wave pipe:

53 liters, tuned to 40 Hz. Note that both speakers have nearly the same cross-section and the baffle width is identical

That's probably about right for an EBS BR, but to have the box operate as a helmhotz resonator, it would have to have something
like golden ratio dimensions. This would make it pretty chunky in decor terms, and it would have to be stand mounted to get it to ear level. Not real good WAF.

Final point. MLTL's seem to control the driver better below cut-off. I don't know why that happens, but it does. This means a lot less doppler distortion under heavy bass. I believe that this is why the bass of MLTL's sounds so good. At least with pipes the size and cut-off of my stuff.

Bob

There is nothing magical about either method. However, pipe tuning requires standing wave modes, so equally spaced harmonics are assured. You can position things to get around them, but they are a necessary consequence of the design.

I also want to point out that cabinets that are physically tall aren't necessarily configured as transmission lines. There are standing waves modes inside any chamber, but if the primary tuning feature is Helmholtz resonance, then the speaker is configured as a bass-reflex system. If the primary tuning feature is length related, then it is a tuned pipe. Unwanted standing wave modes should be suppressed in any loudspeaker cabinet, and the same techniques that are used to suppress harmonics in a tuned pipe can be used to supress standing waves in bass-reflex cabinet, or any other, for that matter.

What do you mean by midrange smear? Not sure I follow what you are trying to describe.

If you design the enclosure geometry and distribute the stuffing correctly, then I don't think you will have spikes in the midrange response. You may end up with a +/- 2 dB ripple but this is small compared to all of the other artifacts from the room and front baffle on the SPL response.

Probably the safest and smallest classic TL design is a tapered pipe. This enclosure will push the 3/4, 5/4, 7/4 ... standing waves higher in frequency where they should not be a problem.

Martin

The MathCad worksheets can calculate the response of either a Helmholtz or a 1D pipe resonance. The lumped parameter freeware programs most people use only calculate Helmholtz resonance so they never predict standing waves. Classic bass reflex enclosures will have standing waves that effect the SPL response, see Figure 6 in my ML TQWT article.

People assume that if the SPL plot shows a flat response it must be flat, unfortunately this is not the case. One of the benefits of using the current versions of the MathCad worksheets is that you also model standing waves in one of the three enclosure directions. If you run the worksheets three times, you can get a hint about the impact of the standing waves in a classic bass reflex design. I have a 3D version of the worksheet but it is not completely debugged and is sitting on the far back burner.

One of the other great myths of bass reflex design is that you only line the walls and place nothing in the center of the enclosure. This will produce almost the worst case standing wave situation, the worst is totally a empty box. None of the freeware lumped parameter models will give you a hint at the potential for standing waves in the bass reflex enclosure.

If you are designing a tall bass reflex enclosure then I recommend using my Ported Box worksheet which will include any standing waves in the long direction and allow you to optimize the placement of both the driver and the port along the tall dimension. Using the freeware lumped parameter models will only tell part of the story, I hate surprise endings.

Martin