Offset Bipolar MLTL with CSS EL70 Drivers--Part 1

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Three years ago I designed a bipolar mass loaded transmission line (MLTL) speaker that used the Creative Sound Solutions (CSS) full range FR125S (front) and extended range WR125S (rear) drivers. The drivers in each box were connected in parallel which created a 4 ohms load and both the front and rear drivers were located at the same height. Given the 4.5” diameter size of these drivers, various listeners were amazed how such small drivers could produce generate the bass level that they did. The upper frequency dispersion did suffer a bit because of the directivity of the full range drivers, the overall performance and low cost yielded a very satisfying speaker. While I did notice that there was a dip of 4 dB in the on-axis response in the 400 to 500 Hz range because of the wraparound of energy from the rear driver to the front driver when the front to rear spacing between the drivers was a half wavelength, I could not hear an appreciable degradation to the overall performance of these speakers.

Ordinary free standing monopolar speakers radiate into 4 pi space (completely around the box) at lower frequencies but transition to 2 pi space (only forward radiation from the front baffle) as frequency increases. Hence, this transition in radiation characteristic is commonly called ‘baffle step loss’ as the listener hears 6 dB difference between the 2 pi radiation frequencies vs. the 4 pi radiation portion of the band. If the speaker is intended to be used away from the boundaries of the listening room, the designer typically will compensate for the difference between the 4 pi and 2 pi regions.

Bipolar speaker enclosures have drivers on their front and rear baffles with the drivers are wired in phase so they produce appropriately near omni-directional sound within the room as they radiate into 4 pi space across their entire frequency band. Thus no baffle step compensation is needed with bipolar speakers to equalize their sound across frequency. Hence, bipolar speakers address baffle step compensation with an acoustical answer and no additional equalization is usually needed to balance their frequency response. Finally, because of their rearward radiation, bipolar speakers do require room placement at least three to four feet away from the rear wall boundary so that the rear firing wave does not produce too much reflected sound which raises the reverberant energy within the room.

Recently, I started to consider how to design a bipolar speaker that improved on my earlier efforts. Fortunately, my friend Duke LeJeune of AudioKinesis had developed an outstanding audiophile quality bipolar speaker. Duke’s Dream Maker two-way speakers use premium 10 inch pro audio woofers combined with compression drivers situated in a low-coloration constant directivity waveguide. Duke generously documents his journey to design his speakers in a four-part posting on his Audio Circle forum entitled “The Case for the Controlled-pattern Offset Bipole”. You can read his story at:

http://www.audiocircle.com/index.php?topic=56877.0

After reading Duke’s four-part posting, my takeaway focused on the benefits from the offset bipolar arrangement. First, the offset configuration (difference in vertical spacing between the front and rear drivers) helps mitigate the wraparound cancellation effect between of the front and rear drivers. Furthermore, a wider front baffle increases the front to back driver path length which lowers the frequency region where the wraparound cancellation occurs. Second, if the rear driver is located low on the back side, then the offset driver placement provides boundary enhancement which can smooth the in-room bass response. Furthermore, an offset bipole can be implemented with various bass tunings such as sealed boxes, vented boxes, or transmission line enclosures.

Duke’s efforts prompted me in new directions for an improved full range driver bipolar speaker configuration. To explore the offset bipolar configuration I wanted to use full range drivers and preferably a MLTL as this alignment can extract a pleasing amount of bass from small drivers. What I envisioned was a box that housed two back-to-back MLTLs within each enclosure. The front MLTL would have the driver near ear listening level with the port near the bottom of the box while the rear MLTL would be mirror imaged with the driver low on the rear baffle and the port near the top of the box. For simplicity the internal back wall of each MLTL would be shared within the enclosure. Although Duke’s work to control the directivity of his speakers via a waveguide mounted tweeter can not be implicitly applied be to a full range driver, his notes on speaker toe-in will improve in-room sound with any speakers.

For my offset bipolar design I was fortunate to have access to a prototype set of Creative Sound Solutions (www.creativesound.ca) EL70 four inches diameter paper cone drivers. The CSS EL70s were designed, developed, and manufactured for CSS by Mark Fenlon of Mark Audio. The price for these drivers is $45 each so they should be an excellent value.

To design the enclosure for the offset bipolar speaker I used Martin J. King’s exceptional worksheets. Martin’s worksheets will not explicitly model an offset bipolar speaker but they will model the individual MLTLs within each box. The series connected combination of the two MLTLs in each enclosure will not increase the overall sound pressure level (although the radiating surface is doubled with two drivers the impedance is doubled so the net SPL remains unchanged). Iterations using the MLTL worksheet yielded a design that is a straight pipe with the following internal dimensions: 39 inches height, driver located 14 inches below the top (the driver is located about one third down the line to mitigate response ripples), port located 37 inches from the top, a cross-sectional area of 27 square inches, and stuffing density of 0.54 pounds per cubic feet. The port diameter is 2.0625 inches with a length of 3.5 inches. For my prototypes I chose to use an internal cross-section for each individual MLTL of 3 inches depth and 9 inches width. With 0.75 inch thick material, the external dimensions of the overall bipolar enclosure (two sandwiched MLTLs) is 40.5” high, 8.75” deep, and of 10.5” wide.

The predicted performance of the individual MLTL is shown in attached figure. The SPL response has a F3 point of 40 Hz with a port null at 45 Hz. The graphs are plotted for an input power level of one watt. Note that the deflection is well within maximum deflection (Xmax) limits and the impedance of the individual MLTL is typical for MLTL speakers. When connected in series the overall impedance of the bipolar speaker will be nominally 8 ohms.

I'll attach a photo of the enclosures that were constructed for this design. Walnut panels were used for the sides and top of the box with MDF employed for the front and rear baffles, the shared internal panel and bottom of the box. The terminal plate (I used the Parts Express #260-309) was located on the bottom of the box. Hardwood rails were used to stabilize the enclosure and raise it for access to the terminal plate. The port tubes are Parts Express #260-322 cut to the correct length.

For each pipe I stuffed the volume from just below the front driver to the top of the box with Acousta Stuf™. For the rear driver the volume from just above the driver to the bottom of the box was stuffed. No additional stuffing was used within the boxes. An alternative to Acousta Stuf is to use a 9 x 16 inches sheet of 1 ½ inch thick convoluted acoustical foam behind each driver. Again the front and rear drivers are connected in series within the enclosure and no equalization circuitry is in the signal path.

Early listening tests reveal a relatively well balanced response with a spacious sound within my listening room. Again you have to be careful when positioning bipolar speakers as too much reverberant sound can overpowers the basic presentation of the source material. The speakers need to be placed at least three to four feet away from the wall behind them for best results. Over the bass portion of the band the sound pressure level is surprising for the size of the speakers but that result was expected. While you don’t expect 4 inch drivers to produce blow-you-over subwoofer kind of bass, you do hear an integrated, balanced musical statement. They have excellent sound for most types of music and I was particularly impressed with their performance with classical orchestral arrangements. They exhibit the typical magic of single driver speakers that we have come to expect.

Initially, my speakers imparted a very slight nasal sound to vocals but they seem to produce more pleasing sound as break-in progressed. Another observation with these speakers is that the amount of stuffing material can be varied to better adapt their sound to your listening room. For example, my home listening room has hardwood floor and sheet rock walls with minimal sound damping within the room. For that environment more stuffing was needed to enable the speakers to sound normal. But when placed in a larger room with carpeting and more damping, their sound was dull and tubby. When some of the stuffing was removed, the sound was brighter and more pleasant. My point is that with DIY speakers you have the ability to adjust your speakers to your taste. With more experience with these speakers, we will better understand whether any frequency contouring or other correction will be needed.

Overall I will say that for less than $200 worth of drivers you can yield a relatively small, floor-standing pair of speakers that will satisfy listeners who have small to medium sized listening areas. Acoustical answers to acoustical problems can be very pleasing.
 

Attachments

  • el70mltlresponses.jpg
    el70mltlresponses.jpg
    89.9 KB · Views: 2,227
Jim,

I have the MK worksheet for Bipole, if I remember correctly it does not have an offset calculation for the front and back drivers.

How do you come to the conclusion of setting the offset so that there is no or minimal wrap around at certain freq.? Is there a formula?

I have my drivers ready to design a bipole a guideline for designing the bipole offset would be helpful.

Thanks.
 
You are correct in that Martin's worksheets do not correctly model an offset bipolar speaker. Offsetting the rear driver does increase the wraparound length (the wraparound frequency area is the frequency at which the drivers are separated by a half wavelength) which would lower the frequency range wherein the wraparound would occur. It also tends to disperse the effect. But offsetting the rear driver will not totally eliminate the effect--it does move it downward in frequency so that it is less apparent.

So far my measurements on my prototype have been limited to in-room testing which does not yield a very good measurement of the wraparound effect for the lower frequencies where it occurs. And near field testing does not show the effect so you must test in the far field. In the future I plan to take some outdoor measurements which will better isolate the effect.

If I were you (and if you plan to try an offset bipole) I would use the half wavelength criterion as a guideline to get an idea of the area of concern.
 
ttan98 asked: "What wavelength would I use for 1st approximation?"

Perhaps this suggestion will get you going:
Take a string (or use the dimensions of your design) and measure from center to center of the two drivers. Compute the frequency at which that distance is a half wavelength. Thus you have the value of where the wrap around effect would be centered.
 
Jim, I'm honored that you would see fit to incorporate some of the ideas I talked about into your latest design. And I thank you for giving me credit... though I think you'd already figured out most if not all of it on your own.

Ttan98, the wraparound dip is usually not symmetrical, and the greater the vertical offset between the front woofer and rear woofer (or fullrange driver) the less symmetrical it will be. The dip will have a shallow slope on the low frequency side, and a steeper slope on the high frequency side. This is because the wraparound itself is spread out a bit. The shortest wraparound path length for the rear driver will usually be horizontally around the sides of the enclosure, but there will also be wraparound energy along diagonals as well as over-the-top. Since these other path lengths are longer, they sort of spread out the dip on the low-frequency side.

I think the deepest part of the dip will be close to (perhaps a little below) the frequency where 1/2 wavelength is equal to the depth of the enclosure plus 1/2 of the width. So if the enclosure is 12 inches wide and 8 inches deep, that would give a path length difference of 14 inches, corresponding to 1/2 wavelength at about 480 Hz.

Now a 12 inch wide baffle won't start to baffle-step (which is a form of wrap-around) until about 560 Hz, so we won't have a whole lot of wrap-around energy at 480 Hz, and therefore the dip won't be very deep. The wider-than-deep geometry that I use isn't really necessary; in practice the enclosure can even be a little deeper than it is wide, as the wraparound notch is probably less audible than the floor-bounce notch that virtually all speakers have - but the offset bipole has less floor-bounce notch than most.

You see, the rear-facing driver's proximity to the floor will greatly reduce the depth of the floor-bounce notch because it will not be notching at the same frequency as the higher up front-facing driver. So the floor-bounce notch is significantly filled in by the rear woofer's contribution, assuming it's fairly close to the floor.
 
Last edited:
frugal-phile™
Joined 2001
Paid Member
The wider-than-deep geometry

Duke, good to see you here.

Attached is a sim that Svante did with an 800x400mm baffle. this is actually a box mounted on a wall, which is topologically the same as a push-push bipole with a box 2x as deep. As the depth decreases the width to depth ratio increases. As can be seen this ratio can have a substantial affect on the magnitude of the bipole dip.

dave
 

Attachments

  • 3-bipole_03.gif
    3-bipole_03.gif
    15.5 KB · Views: 2,341
Hi Dave,

Thanks! It's good to be here.

That's a pretty clever way of simulating a bipole with a program that wasn't designed to do bipoles. It can't simulate an offset bipole, but it's still quite useful for investigating different combinations of enclosure width and depth. Comparing the blue trace with the green trace illustrates the advantage of going with a box that is much wider than it is deep, which is educational even if such a confiuration isn't always practical. Nice job!

Duke
 
Last edited:
Hi Jim,
Really like this design. Would it be possible to use this speaker alignment in a one box dipole, such as your CSS FR125S/WR125S design, or is the inside wall necessary? Thanks, Larry

Larry,

I don't think that a single box (without a partition) that includes the equivalent of both MLTLs will work for an offset bipolar which would be implemented with MLTLs. You could do a single box for a design that used say a sealed box or even a ported box. But the MLTLs in this offset design takes into account takes the length between the port and driver and such so I can not envision how a single box could accommodate such an arrangement. The partitioned box--especially for a straight MLTL--is relatively easy to build so you can be successful with this design.

Jim
 
Jim, I've built several different prototype enclosures and think I've just about got it figured out, but have to spend a few days on other projects now (ones that I get paid for). Unfortunately what I'm doing will still require subs; I'm only trying to get down to 80 Hz ballpark. I have a more truly fullrange model on the drawing board, but the enclosure size is pretty daunting (about six cubic feet internal volume).
 
Last edited:
Larry,

I don't think that a single box (without a partition) that includes the equivalent of both MLTLs will work for an offset bipolar which would be implemented with MLTLs. You could do a single box for a design that used say a sealed box or even a ported box. But the MLTLs in this offset design takes into account takes the length between the port and driver and such so I can not envision how a single box could accommodate such an arrangement. The partitioned box--especially for a straight MLTL--is relatively easy to build so you can be successful with this design.

Jim

Thanks Jim. I was hoping to use the driver offset on a project I had already cut wood for. I guess I could add a partition in the middle, but then would have to change the single port to two different sized ports.
 
Jim. I am in the process of building your offset Bipolar enclosures with the CSSEL70 drivers. The plan drawn by "did" mentions that the stuffing hangs from the top of the box to just below the driver. That makes sense for the front section where the speaker is at the top of the box, but surely for the rear facing section where the speaker position is reversed to the bottom of the box, the stuffing would be from the bottom of the box to just above the driver? Did you put the stuffing in place before you closed it all up or did you make the speaker baffles removeable, so that you could alter the amount of stuffing if it was necessary. Thanks.
 
frugal-phile™
Joined 2001
Paid Member
The plan drawn by "did" mentions that the stuffing hangs from the top of the box to just below the driver. That makes sense for the front section where the speaker is at the top of the box, but surely for the rear facing section where the speaker position is reversed to the bottom of the box, the stuffing would be from the bottom of the box to just above the driver?

Since It is essentially 2 identical TLs, with the back one upsidedown, i would consider the top of the back one to be the bottom (wrt gravity). In other words i agree with you. I'll add some comments to the drawing.

dld.
 
Leeoh,

Yes, the stuffing goes from the top of the box to just below the driver for the front driver (right side up MLTL) and from the bottom of the enclosure to the just above the rear driver (upside down MLTL).

I place the stuffing inside the box via the driver hole cutouts and you can vary the amount by this manner. You can vary the amount of stuffing to adjust the bass performance a bit from over stuffed to lightly stuffed to best match your listening room dynamics.

Good luck with your project.

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