bigbrowncow said:
Steve,
I'm not one of the geniuses, having built but one pair of speakers (bipole MLTQWT FR125s) so far. But, I won't let that stop me from commenting.
Looking at your graphs, I would say that you should go for a smaller diameter (say 50mm) and corresponding shorter length port. This should squelch those mid frequency ripples and make for a nicer impulse response. You may also want to play around with driver and port locatons on the line to deal with upper resonances. I don't see the port location on your worksheet, it looks like a different worksheet than I used ("ported box").
If you look around the forums, you will read that people, myself included, have been finding that BR/MLTL implementations of the FR125s result in a bass-heavy response (despite the simulations) and stuffing the ports is the solution. I kept this in mind when designing mine and aimed for a more gradual roll-off (without port stuffing). I still found them bass-heavy. I think that the factory T/S parameters, which I used, may be off a bit.
See:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=70255&highlight=
and
http://www.diyaudio.com/forums/showthread.php?s=&threadid=64557&highlight=
I started with 2" dia x 2.5" long ports and tried foam stuffing in these. I then slipped some 1.5"x3.5" ports inside the 2" ones (these ports sim great with the hobby hifi specs) and have settled (for now) on these with a pair of wool socks rolled up and stuffed inside.
If you want to see what mine look like, they're here:
http://www.diyaudio.com/forums/showthread.php?postid=824319#post824319
Dave (Planet10) has heard them (And many other FR125s implementations) and a bunch of other folks at RAWfest (albeit without socks in the ports) so you can solicit some other oppinions.
I feel that they could be more open in the midrange and I'll play with removing some stuffing eventually. I also think that they perform better in a smallish room and found them straining somewhat to fill the large room at Al (RAW) Wooly's place.
Max
I think GM has given the big clue here. A few posts back.
Be careful with what the simulations tell you. The Xmax that allows the low bass is also responsible for the problem of not being able to suspend at higher power levels.
I would seriously consider the bipolar bipoles if I were building with this driver again or tuning the cabinet to closer to Fs.
Good luck
Be careful with what the simulations tell you. The Xmax that allows the low bass is also responsible for the problem of not being able to suspend at higher power levels.
I would seriously consider the bipolar bipoles if I were building with this driver again or tuning the cabinet to closer to Fs.
Good luck
Comments and Additional Modeling of Bipolar MLTL
Much has been written in this thread and I wanted to summarize some recent work that I have done in an attempt to clear any confusion on previous posts. As you recall, I based my bipolar MLTL design on earlier postings by Greg Monfort (GM). Last summer I constructed a set of speakers per his basic design. My speakers and others built per my interpretation of his original design both sound and measure very good. But later questions were raised by Planet 10 (Dave) about the original simulations by GM in Posts #7 and 9 of this thread. The issue revolves around whether GM’s simulations modeled a single driver or two drivers in the enclosure. My goal herein is to report on my findings and modeling of this design with Martin King’s latest spreadsheets. The shortcut takeaway from this recent work is that the design that I built is correct and results in an outstanding speaker for those who wish to build it.
A few words of comments are necessary about modeling bipolar speakers. Martin has not yet released a bipolar worksheet, so some interpretation of bipolar speakers is needed to properly comprehend their resultant performance. Martin’s tools are adequate to predict the performance of a bipolar configuration if the results from a single driver box are appropriately extended to the bipolar configuration. For example, one can build a bipolar speaker with the assumption of two separate single driver speakers enclosures placed back to back. You can extend this assumption by eliminating the back walls of the two single driver boxes so that you have one enclosure which is twice the volume of the single driver box. You also can evolve to one port in this double volume box by using a port which has twice the area as each port in the single driver boxes. Finally, you parallel connect the two drivers, the resultant impedance is half of the original driver’s impedance. Thus the impedance sensitivity change increases the bipolar speaker sensitivity (SPL) by 3 dB vs. the single driver.
As an aside, there is another way to model a bipolar speaker with Martin’s worksheets. This second method is to change the T/S parameters for the driver to create an equivalent driver that has twice the Vas and Sd but half the Re of the single driver. You arrive at the same results with this second methodology once you account for the fact that the bipolar power response SPL increases only 3 dB (the impedance sensitivity change only). Bipoles don’t have the 6 dB SPL increase you would expect from paralleled drivers as their radiation patterns do not overlap—only the 3 dB impedance sensitivity increase results.
For my simulations I used the single driver in the box methodology. All simulations used the same enclosure with changes as noted. The specified FR125S T/S parameters from the CSS website (www.creativesound.ca) were used for this analysis. The enclosure modeled has inside dimensions as follows: line length of 39.5”, the driver located 14” from the top, the port placed at 37” from the top, and a cross sectional area (no taper) of 43.5 sq. in. The port has a radius of 1.5” and a length of 3” while the stuffing density is 0.50 pounds per cubic foot. These dimensions are identical GM’s original design and to the enclosures that I used in the construction of my bipolar MLTL.
For this enclosure the various simulation cases are:
1. Single driver in enclosure volume. This is the design that GM later stated to Planet 10 that he thought he originally modeled. (Again to create a bipolar speaker you would double the cross sectional area (volume of box) and port tube area.) The simulation results for this case yielded a less than ideal SPL response as it exhibited a 5 peak dB (above the baseline SPL) at 50 Hz and 2 dB ripples in the 350-500 Hz area. I have attached a plot (the top plot in the attachment) from Martin’s worksheet to illustrate the SPL vs. frequency performance. Bottom line for this simulation is that the response for one driver in the box does not yield a top notch design. Upon viewing this plot I don’t think these simulation results are what GM really meant to communicate as a suggested design for this MLTL bipolar speaker. Furthermore, for a bipolar version the box volume would need to be doubled to accommodate the second driver.
2. Two drivers in the enclosure volume. To simulate two drivers in the enclosure volume I divided the cross sectional area by half (from 43.5 to 21.75 sq. in. (this halves the enclosure volume) and reduce the port tube area by half (the tube radius reduces to 1.06” from 1.5” to achieve this area reduction) while all of the other parameters remain the same as for the first case. For these values the on axis calculated SPL response is shown in the attachment as the bottom plot. Excellent performance is demonstrated with flatness within +/- 1.5 dB across the band. The low end 3 dB down point is below 45 Hz. The port response shows a dip just below 50 Hz.
The measured data (see my earlier postings in this thread for thumbnails of the measured on-axis response plus near field speaker and port data) match the simulations for the two drivers in the enclosure case. Clearly, the issues raised by Planet 10 were not valid and GM’s recollections of the original design also weren’t correct either. The original design from GM that I used was clearly for two drivers in the enclosure box case and my measurements support this conclusion. Bottom line is that my bipolar design as given on the Planet 10 website as:
http://homepage.mac.com/tlinespeakers/FAL/downloads/BipolarMLTLDesignPak.pdf
This URL depicts the correct bipolar MLTL design and will yield good results for those who build it.
Another recent data point comes from Fred Thompson who is the prolific amateur speaker builder from Houston. He heard my speakers at the Great Plains Audio Fest in early May and liked their sound. Fred recently built a set of bipolar MLTL from my plans. His comments and impressions are incorporated in this thread on the Audio Round Table forum at:
http://audioroundtable.com/SingleDriverSpeakers/messages/1933.html
I suggest that those interested in this design to read this thread and see Fred’s comments and follow-up posts as he reports on his copies of the bipolar design.
Finally, others have reported that the CSS FR125S and WR125S sound bass heavy for some implementations. My listening and measurements with these bipolar MLTL speakers have convinced me that this particular design does not suffer from any imbalance of bass. They are very neutral and correct sounding.
In conclusion, Dave (Planet 10) owes me big time for incorrectly characterizing this design as flawed in Post #7.
Jim
Bipolar MLTL Speaker Plans Fred Thompson's Comments
Much has been written in this thread and I wanted to summarize some recent work that I have done in an attempt to clear any confusion on previous posts. As you recall, I based my bipolar MLTL design on earlier postings by Greg Monfort (GM). Last summer I constructed a set of speakers per his basic design. My speakers and others built per my interpretation of his original design both sound and measure very good. But later questions were raised by Planet 10 (Dave) about the original simulations by GM in Posts #7 and 9 of this thread. The issue revolves around whether GM’s simulations modeled a single driver or two drivers in the enclosure. My goal herein is to report on my findings and modeling of this design with Martin King’s latest spreadsheets. The shortcut takeaway from this recent work is that the design that I built is correct and results in an outstanding speaker for those who wish to build it.
A few words of comments are necessary about modeling bipolar speakers. Martin has not yet released a bipolar worksheet, so some interpretation of bipolar speakers is needed to properly comprehend their resultant performance. Martin’s tools are adequate to predict the performance of a bipolar configuration if the results from a single driver box are appropriately extended to the bipolar configuration. For example, one can build a bipolar speaker with the assumption of two separate single driver speakers enclosures placed back to back. You can extend this assumption by eliminating the back walls of the two single driver boxes so that you have one enclosure which is twice the volume of the single driver box. You also can evolve to one port in this double volume box by using a port which has twice the area as each port in the single driver boxes. Finally, you parallel connect the two drivers, the resultant impedance is half of the original driver’s impedance. Thus the impedance sensitivity change increases the bipolar speaker sensitivity (SPL) by 3 dB vs. the single driver.
As an aside, there is another way to model a bipolar speaker with Martin’s worksheets. This second method is to change the T/S parameters for the driver to create an equivalent driver that has twice the Vas and Sd but half the Re of the single driver. You arrive at the same results with this second methodology once you account for the fact that the bipolar power response SPL increases only 3 dB (the impedance sensitivity change only). Bipoles don’t have the 6 dB SPL increase you would expect from paralleled drivers as their radiation patterns do not overlap—only the 3 dB impedance sensitivity increase results.
For my simulations I used the single driver in the box methodology. All simulations used the same enclosure with changes as noted. The specified FR125S T/S parameters from the CSS website (www.creativesound.ca) were used for this analysis. The enclosure modeled has inside dimensions as follows: line length of 39.5”, the driver located 14” from the top, the port placed at 37” from the top, and a cross sectional area (no taper) of 43.5 sq. in. The port has a radius of 1.5” and a length of 3” while the stuffing density is 0.50 pounds per cubic foot. These dimensions are identical GM’s original design and to the enclosures that I used in the construction of my bipolar MLTL.
For this enclosure the various simulation cases are:
1. Single driver in enclosure volume. This is the design that GM later stated to Planet 10 that he thought he originally modeled. (Again to create a bipolar speaker you would double the cross sectional area (volume of box) and port tube area.) The simulation results for this case yielded a less than ideal SPL response as it exhibited a 5 peak dB (above the baseline SPL) at 50 Hz and 2 dB ripples in the 350-500 Hz area. I have attached a plot (the top plot in the attachment) from Martin’s worksheet to illustrate the SPL vs. frequency performance. Bottom line for this simulation is that the response for one driver in the box does not yield a top notch design. Upon viewing this plot I don’t think these simulation results are what GM really meant to communicate as a suggested design for this MLTL bipolar speaker. Furthermore, for a bipolar version the box volume would need to be doubled to accommodate the second driver.
2. Two drivers in the enclosure volume. To simulate two drivers in the enclosure volume I divided the cross sectional area by half (from 43.5 to 21.75 sq. in. (this halves the enclosure volume) and reduce the port tube area by half (the tube radius reduces to 1.06” from 1.5” to achieve this area reduction) while all of the other parameters remain the same as for the first case. For these values the on axis calculated SPL response is shown in the attachment as the bottom plot. Excellent performance is demonstrated with flatness within +/- 1.5 dB across the band. The low end 3 dB down point is below 45 Hz. The port response shows a dip just below 50 Hz.
The measured data (see my earlier postings in this thread for thumbnails of the measured on-axis response plus near field speaker and port data) match the simulations for the two drivers in the enclosure case. Clearly, the issues raised by Planet 10 were not valid and GM’s recollections of the original design also weren’t correct either. The original design from GM that I used was clearly for two drivers in the enclosure box case and my measurements support this conclusion. Bottom line is that my bipolar design as given on the Planet 10 website as:
http://homepage.mac.com/tlinespeakers/FAL/downloads/BipolarMLTLDesignPak.pdf
This URL depicts the correct bipolar MLTL design and will yield good results for those who build it.
Another recent data point comes from Fred Thompson who is the prolific amateur speaker builder from Houston. He heard my speakers at the Great Plains Audio Fest in early May and liked their sound. Fred recently built a set of bipolar MLTL from my plans. His comments and impressions are incorporated in this thread on the Audio Round Table forum at:
http://audioroundtable.com/SingleDriverSpeakers/messages/1933.html
I suggest that those interested in this design to read this thread and see Fred’s comments and follow-up posts as he reports on his copies of the bipolar design.
Finally, others have reported that the CSS FR125S and WR125S sound bass heavy for some implementations. My listening and measurements with these bipolar MLTL speakers have convinced me that this particular design does not suffer from any imbalance of bass. They are very neutral and correct sounding.
In conclusion, Dave (Planet 10) owes me big time for incorrectly characterizing this design as flawed in Post #7.
Jim
Bipolar MLTL Speaker Plans Fred Thompson's Comments
Attachments
Re: Comments and Additional Modeling of Bipolar MLTL
All i did was bring up the point after confirming it with GM... what you built with 2 drivers was his 1st approximation for a single driver, and he expressed some dissapointment that you never contacted him so that he could finish it. That you got lucky & it worked is just a big bonus to the diy community. I feel i have nothing to apologize for -- i was just the messenger -- that you would even take affront i find mysterious.
GM posts hundreds of designs and it is important that people know how to interpret them -- you misinterpreted an unfinished design and that confuses things.
dave
Jim Griffin said:
All i did was bring up the point after confirming it with GM... what you built with 2 drivers was his 1st approximation for a single driver, and he expressed some dissapointment that you never contacted him so that he could finish it. That you got lucky & it worked is just a big bonus to the diy community. I feel i have nothing to apologize for -- i was just the messenger -- that you would even take affront i find mysterious.
GM posts hundreds of designs and it is important that people know how to interpret them -- you misinterpreted an unfinished design and that confuses things.
dave
Dave,
I had earlier contact with GM on several different versions of the original design. Viewing his early work it was clear to me that the version that I built was for two drivers in the box based on the CSA of the speaker. I also did a 'resonableness' check by comparing the original design to other MLTL boxes with these same drivers. I don't think luck had anything to do with making the right decisions on this design.
Often people raise a point of contention without valid reasons (modeling, measurements, etc.). That is just idle speculation and not a service to anyone.
I did learn from this experience that it is better to independently verify the design eventhough it may appear that all factors are specified. Both novice and expert builders should measure twice before cutting wood. Bottom line is that this MLTL design is valid and it squeezes the most out of these drivers.
But I look at some of the BIB and other designs that pop-up every day on the web and I'm concerned that all of these designs have been optimized and have had the BS test applied to them. My advice to anyone thinking of building someone else's design is to do your homework before starting.
Jim
I had earlier contact with GM on several different versions of the original design. Viewing his early work it was clear to me that the version that I built was for two drivers in the box based on the CSA of the speaker. I also did a 'resonableness' check by comparing the original design to other MLTL boxes with these same drivers. I don't think luck had anything to do with making the right decisions on this design.
Often people raise a point of contention without valid reasons (modeling, measurements, etc.). That is just idle speculation and not a service to anyone.
I did learn from this experience that it is better to independently verify the design eventhough it may appear that all factors are specified. Both novice and expert builders should measure twice before cutting wood. Bottom line is that this MLTL design is valid and it squeezes the most out of these drivers.
But I look at some of the BIB and other designs that pop-up every day on the web and I'm concerned that all of these designs have been optimized and have had the BS test applied to them. My advice to anyone thinking of building someone else's design is to do your homework before starting.
Jim
Hi guys; let's just celebrate the fact that Jim's bipole design works. I like. It's easy to get wires crossed though with 'net designs -it's happened to me, so I know the feeling from all perspectives!
In defense of the BIB, it's not a new idea, and I'm sure none of us would suggest it to be the match of an MLTL or similar. That said, they do usually measure or perform better than the sim. Given the fact that this particular enclosure is extremely hevily influenced by / dependant upon the room, there's little point in attempting ultimate refinements in the design stages. The dimensions GM and I suggest should allow builders to avoid the major pitfalls, but aren't going to be perfect for every room. Nature of the beast.
Best regards to all
Scott
In defense of the BIB, it's not a new idea, and I'm sure none of us would suggest it to be the match of an MLTL or similar. That said, they do usually measure or perform better than the sim. Given the fact that this particular enclosure is extremely hevily influenced by / dependant upon the room, there's little point in attempting ultimate refinements in the design stages. The dimensions GM and I suggest should allow builders to avoid the major pitfalls, but aren't going to be perfect for every room. Nature of the beast.
Best regards to all
Scott
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.