David Weems' Stuffed, Taperd-Pipe...

[rogoll]

Need some feedback from someone familiar with tapered-pipes (BassBox Lite doesn't address them). In his book, David Weems (Designing, Building, and Testing Your Own Speaker System, pages 19-20) gives the plans for one with external dimensions of 8" wide by 9" deep by 39-3/8" tall. This yields an internal width of 6-1/2" wide by 8" deep (apparently using 3/4 material for the sides, and 1/2" material for the front and back). There's a rectangular opening (port) at the bottom front measuring 6-1/2 wide by 3-1/4", and an internal partition that is 30" long attached to the interior beginning from just above the port and slanting towards the center at the top, forming the tapered-pipe. The entire box is filled with damping material (he suggests 6 ounces). He does not specify any particular driver...just one that's 6-1/2 wide. (See this page for a drawing.)

I'm planning to use the Audax AP170Z0 6-1/2" Shielded Aerogel Midbass. My major question is, what will happen if I change the external depth from 9" to 10" using 3/4" material, which by my calculations would make the enclosure 1/2" deeper?

Also, is 6 ounces of fill (polyester or AcoustaStuff) sound like it would be enough to fill this much internal space?

Thank you,
Leonard

[Pete Schumacher ®]

Tapered TL understanding has come a long way since then. That pipe is actually reversed from what it should be to get the flattest extension into the deepest bass regions. Tapered TLs start at the closed end with the largest area and end at the open terminus with a much smaller area.

http://www.quarter-wave.com

[ajinfla]

Tapered TL understanding has come a long way since then. That pipe is actually reversed from what it should be to get the flattest extension into the deepest bass regions. Tapered TLs start at the closed end with the largest area and end at the open terminus with a much smaller area.

You mean like Fig 2-7 ?

[Pete Schumacher ®]

You mean like Fig 2-7 ?

But even that one is hardly optimal. Tapered TLs with the smoothest response have the driver mounted about 1/5th the way from the closed end, not ON the closed end.

For a subwoofer or a woofer crossed very low, mounting on the closed end is certainly an option, but not if the woofer is to be used as the bottom of a two way crossed above 1000Hz.

[HareBrained]

I've seen the driver placed anywhere from 10%-50% along the line and still sound great.

There is a spreadsheet on the MJ King's website that offers a starting point. Martin use to offer his MathCad implementation for free use but the license was abused. Regardless of where you start, that is just a starting point. They require acoustic tuning to extract the most from the enclosure.

There is supposed to be a relationship between Qts and suitability but I don't remember the exact range of values. I know the Morel MW164 fell within the range at 0.68 because that's what I used to build mine (some years ago). But think think that was a RoT and not hard-and-fast.

You may want to research mass-loaded TLs (MLTL). They're typically a straight (no taper) with a port as the "open end". They seem to be favored for lower-Qts drivers.

[Pete Schumacher ®]

I've seen the driver placed anywhere from 10%-50% along the line and still sound great.

There is a spreadsheet on the MJ King's website that offers a starting point. Martin use to offer his MathCad implementation for free use but the license was abused. Regardless of where you start, that is just a starting point. They require acoustic tuning to extract the most from the enclosure.

There is supposed to be a relationship between Qts and suitability but I don't remember the exact range of values. I know the Morel MW164 fell within the range at 0.68 because that's what I used to build mine (some years ago). But think think that was a RoT and not hard-and-fast.

You may want to research mass-loaded TLs (MLTL). They're typically a straight (no taper) with a port as the "open end". They seem to be favored for lower-Qts drivers.

The lowest amount of passband ripple will occur with driver 20% of the way down the line (derived using Martin's worksheets). Adequate damping material can help other positions come close. But the smoothest response with the least damping material comes with the 20% mounting.

As with vented boxes, drivers with higher Qts will typically need larger volumes to optimize performance. When you get to values as high as .68, it becomes tougher to keep response from peaking as you approach tuning.

I'm not sure that MLTL is any better for lower Qts drivers than a tapered TL. Maybe Paul K could run a couple sims to see what happens?

[rogoll]

You mean like Fig 2-7 ?

Yes, according to his book these are two different "animals". Fig. 2-7 is a "Transmission Line Enclosure", while Figure 2-9 is a "Tapered-Pipe Enclosure". I am not about to begin to say that I understand the differences, nor the prevailing theory behind enclosures of this nature. From the text it seems that the mass of the damping material, and the "leakproofness" of the enclosure (obviously, other than at the port) have some of the most significant impact on the sound quality.

At this point in time (and in my woodworking/speaker building skills) the tapered-pipe of Fig. 2-9 would make a better project for me.

[Paul K.]

The reverse-tapered TL in Weems' book seems to be an attempt at a "generic" ML-TQWT or, perhaps, a sort of Voight (sp?) pipe, where the driver is located at or near the halfway point in the line's length. It's almost impossible to design a generic TL of any type that you can expect to have reasonable success with just any driver. If you have a specific driver in mind (with its T/S values) with an idea of how big the cabinet can be and its approximate shape, I'd be happy to model TLs for you. Some drivers just don't work at all well in any kind of TL, while others will work well in one type, and even others will work well in several types. Basically I've found that drivers with Qts below 0.3 generally don't work well in any kind of TL, but that may say more about my (in)ability to handle those types of drivers!? Too, I probably have different criteria on what constitutes a "good" result than others.
Paul

Yes, according to his book these are two different "animals". Fig. 2-7 is a "Transmission Line Enclosure", while Figure 2-9 is a "Tapered-Pipe Enclosure". I am not about to begin to say that I understand the differences, nor the prevailing theory behind enclosures of this nature. From the text it seems that the mass of the damping material, and the "leakproofness" of the enclosure (obviously, other than at the port) have some of the most significant impact on the sound quality.

At this point in time (and in my woodworking/speaker building skills) the tapered-pipe of Fig. 2-9 would make a better project for me.

[rogoll]

If you have a specific driver in mind (with its T/S values) with an idea of how big the cabinet can be and its approximate shape, I'd be happy to model TLs for you. Some drivers just don't work at all well in any kind of TL, while others will work well in one type, and even others will work well in several types. Basically I've found that drivers with Qts below 0.3 generally don't work well in any kind of TL, but that may say more about my (in)ability to handle those types of drivers!? Too, I probably have different criteria on what constitutes a "good" result than others.
Paul

Paul, does the information in my first post (very top of page), including the link to the Parts Express page give you enough info?

Thank you,
Leonard

[Paul K.]

Yep, it does. I've printed the T/S values published by Audax and those from PE's measurements; these two sets aren't the same but are pretty close. I'll see what is the best type of TL for this driver contained in a cabinet size/shape like proposed by Weems and with your suggested modifications. "Best" TL to me implies as smooth an overall response as possible with the lowest F3 for the enclosed volume of the cabinet in mind. I should be able to let you know something today or tomorrow.
Paul

Paul, does the information in my first post (very top of page), including the link to the Parts Express page give you enough info?

Thank you,
Leonard

[Paul K.]

I modeled your selected Audax driver in Weems' proposed enclosure and have attached the system response graph (the solid red line). I assume no one will argue that the response is downright ugly, which really isn't unexpected when one "designs" an enclosure apparently using the driver's physical size as the sole parameter (a 6-1/2" driver). I'll now proceed to come up with something I expect to be a bunch better.
Paul

[Paul K.]

Okay, I modeled an ML-TL, which means it looks like a tall, vented cabinet. The internal cabinet dimensions I used are 6.5"W x 8.5"D x 39"H. I flipped the two drivers, placing the tweeter at the top with the Audax below it, making the center of the Audax be 8" below the top (note: all dimensions and dimensional references are internal). There is a mass-loading port whose center is located 3" above the bottom; this port has a diameter of 2.5" and a length of 3". I used a stuffing density of 1 lb/cu.ft. which will require 9-10 ounces uniformly distributed by density in only the top 19" of the cabinet. You will, of course, need bracing and I'd suggest the "window frame" style and at least 3: one below the woofer's cutout and the other two spaced unevenly below the first one. I've attached its system response, and F3 is about 38-39 Hz. Keep in mind this driver has an Xmax of only 3 mm and could easily be over-driven. Now, I have also modeled an ML-TQWT that's much better than Weems' TL and looks very similar in cabinet structure to his TL. The response doesn't reach as deep (F3 about 45 Hz) and isn't as flat overall as this ML-TL. If you're interested, I can describe and post its response, but this ML-TL is the easiest to build and gives the theoretically lowest bass response in this cabinet size.
Paul

Edit: I forgot to mention that when modeling both Weems' TL and this ML-TL, I used T/S values as measured and published by PE. If I plug in the T/S values from Audax (I presume), I'd shorten the port's length in this ML-TL to 2.5" for a "perfect" match, but not doing so would not likely be audibly noticeable.

[rogoll]

Wow! Thanks for all the work and assistance!

[rogoll]

Okay, I modeled an ML-TL, which means it looks like a tall, vented cabinet. The internal cabinet dimensions I used are 6.5"W x 8.5"D x 39"H. I flipped the two drivers, placing the tweeter at the top with the Audax below it, making the center of the Audax be 8" below the top (note: all dimensions and dimensional references are internal). There is a mass-loading port whose center is located 3" above the bottom; this port has a diameter of 2.5" and a length of 3". I used a stuffing density of 1 lb/cu.ft. which will require 9-10 ounces uniformly distributed by density in only the top 19" of the cabinet. You will, of course, need bracing and I'd suggest the "window frame" style and at least 3: one below the woofer's cutout and the other two spaced unevenly below the first one. I've attached its system response, and F3 is about 38-39 Hz. Keep in mind this driver has an Xmax of only 3 mm and could easily be over-driven. Now, I have also modeled an ML-TQWT that's much better than Weems' TL and looks very similar in cabinet structure to his TL. The response doesn't reach as deep (F3 about 45 Hz) and isn't as flat overall as this ML-TL. If you're interested, I can describe and post its response, but this ML-TL is the easiest to build and gives the theoretically lowest bass response in this cabinet size.
Paul

Edit: I forgot to mention that when modeling both Weems' TL and this ML-TL, I used T/S values as measured and published by PE. If I plug in the T/S values from Audax (I presume), I'd shorten the port's length in this ML-TL to 2.5" for a "perfect" match, but not doing so would not likely be audibly noticeable.

Let me get clarification: Am I correct that there is no internal partition?
Thanks,
Leonard

[Paul K.]

Yes, you're absolutely correct; there is no internal, angled divider. As I suggested you will need to add braces in some form that allow the whole internal height of the cabinet to act as the line's length. The cabinet's height will have a quarter-wavelength resonant frequency which, when augmented by the mass-loading port's dimensions, forms the combined system tuning frequency (about 41 Hz in this case). Don't forget to mount the tweeter above the woofer and make the woofer's center be 8" from the inside of the top.
Paul

Let me get clarification: Am I correct that there is no internal partition?
Thanks,
Leonard

[rogoll]

It certainly is a small world. Paul K.: I just noticed your Location: Hilliard, OH. That's where my parents were living when my father finally retired. I then moved them to the "Sunny South" to keep an eye on them.

[Paul K.]

Yes it is a small world. We lived in Sunny California from 1988 until we retired in 2003 and moved to Hilliard where both our daughters live. One more thing on the ML-TL: you can mount the port on either the baffle or rear panel. If the cabinet would be placed very close to the wall behind, the baffle would be a better location for the port.
Paul

It certainly is a small world. Paul K.: I just noticed your Location: Hilliard, OH. That's where my parents were living when my father finally retired. I then moved them to the "Sunny South" to keep an eye on them.

[HareBrained]

The predicted response of the MLTL for this driver looks mighty fine. Great extension for a 6.5" driver. Not to mention the position appears perfect for driver relationship to the listener. And the drivers are only $30-32 each.

Leonard, what do you plan to use with the 170Z? The flat response up to 1kHz looks very seductive but the rising response above 1kHz looks daunting.

Paul, this driver is very sensitive and could match a number of high-frequency drivers in a 2.5-way configuration. Could you drop a second driver into the sim and post how it changes the box? TIA.

[Paul K.]

Paul, this driver is very sensitive and could match a number of high-frequency drivers in a 2.5-way configuration. Could you drop a second driver into the sim and post how it changes the box? TIA.

I can certainly do that but if you want the same F3, the box volume will essentially have to be doubled. You could double the depth, or increase both depth and width appropriately. If you're willing to give up some of F3, doubling the volume wouldn't be necessary, of course. Enlarging the box will require change the port's dimensions, too. And, with two woofers, the design center would have to end up a bit lower on the baffle at the junction of the two woofers. So, what do you want to aim for on F3?
Paul

[HareBrained]

I can certainly do that but if you want the same F3, the box volume will essentially have to be doubled. You could double the depth, or increase both depth and width appropriately. If you're willing to give up some of F3, doubling the volume wouldn't be necessary, of course. Enlarging the box will require change the port's dimensions, too. And, with two woofers, the design center would have to end up a bit lower on the baffle at the junction of the two woofers. So, what do you want to aim for on F3?
Paul

You bring up a good point about the drivers having to shift. It would push the tweeter position up ~3.5 inches, which bisects the top of the line, and means it needs to be even higher, requiring the Fc be even lower. Not optimal. If the drivers center were to shift down 3.5", I don't think it would be optimum. But would it be acceptable? And I don't like the idea of the MTM-arrangement.

I think I'll drop the 2.5 idea and wait to see what Leonard comes up with. Thanks for the offer.