Announcement

Collapse
No announcement yet.

help with buyout TL MTM idea?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • help with buyout TL MTM idea?

    after the success of the RS150 MT TL project, i've been considering an MTM TL with some of the Jamo buyouts i had sitting around. I'm not expecting real deep bass response from a driver with an Fs in the mid 60s, but i want to try to do better than what i've modeled in a vented alignment. while i've seemingly figured out how to work the ML-TL worksheets, i'm not sure i'm on the right track with this one yet... maybe they're just terrible drivers for a TL, but i seem to get solutions that take me toward ports of very very small lengths (usually a clue not to have a port, i'd guess)... if anyone could get me into the ballpark of an enclosure, i'd really appreciate it. size-wise, there's no particular constraint, but i was envisioning a floorstander similar in size to the tritrix, maybe a little shallower.

    the two drivers would be the Jamo fiberglass 5.25" woofers in series. measured specs for individual driver are:

    Re - 3.37
    Le - .81 mH
    Vas - 5.7 L
    Fs - 64 hz
    Qts - .42
    Qes - .50
    Qms - 2.85

    Sd estimated at ~100 cm2 (half-surround diameter is a little under 4.75")
    BL calculated at 6.6 using martin's worksheets

    http://www.parts-express.com/pe/show...number=299-935

    if these are just really poorly suited for this application, that's understandable, i'd just like to know why (other than the 65 hz Fs). i still think i'm probably modeling something wrong, though. for the two drivers in series, i doubled Vas, Re, Le, and BL, and Sd, but left the others the same as a single driver.
    thanks
    Last edited by patrickm; 11-11-2010, 11:36 PM.
    TM (RS125/TB25-302), TM RS150S / Vifa BC25SC06, RS150-4 / Vifa ML-TL, 3CR-AL Ultra Budget: Electric Blue TM, TMM, MTM, Dragonflies , Mounties: Mini HT Satellite,sub for minions

  • #2
    Re: help with buyout TL MTM idea?

    With the Qts being very close to 0.4 (it will be a bit higher than 0.42 if you model with any added resistance for Radd), an optimum tuning frequency for flatest response down to just above the knee in the curve will likely be close to 60 Hz. So, using that as the 1/4-wavelength resonant frequency for a traditional, non-tapered line (and not an ML-TL), the required line length would be ~57". If you tapered it at 10:1, or if you made it an ML-TL, the line length needed would be ~34". Once you have the system tuning frequency optimized, you can increase the volume within the line to lower F3, but at some point with an ML-TL, the port length may indeed be nothing more than the wall thickness. It would help if you showed us what kind of response you're currently getting, however.
    Paul

    Originally posted by patrickm View Post
    after the success of the RS150 MT TL project, i've been considering an MTM TL with some of the Jamo buyouts i had sitting around. I'm not expecting real deep bass response from a driver with an Fs in the mid 60s, but i want to try to do better than what i've modeled in a vented alignment. while i've seemingly figured out how to work the ML-TL worksheets, i'm not sure i'm on the right track with this one yet... maybe they're just terrible drivers for a TL, but i seem to get solutions that take me toward ports of very very small lengths (usually a clue not to have a port, i'd guess)... if anyone could get me into the ballpark of an enclosure, i'd really appreciate it. size-wise, there's no particular constraint, but i was envisioning a floorstander similar in size to the tritrix, maybe a little shallower.

    the two drivers would be the Jamo fiberglass 5.25" woofers in series. measured specs for individual driver are:

    Re - 3.37
    Le - .81 mH
    Vas - 5.7 L
    Fs - 64 hz
    Qts - .42
    Qes - .50
    Qms - 2.85

    Sd estimated at ~100 cm2 (half-surround diameter is a little under 4.75")
    BL calculated at 6.6 using martin's worksheets

    http://www.parts-express.com/pe/show...number=299-935

    if these are just really poorly suited for this application, that's understandable, i'd just like to know why (other than the 65 hz Fs). i still think i'm probably modeling something wrong, though. for the two drivers in series, i doubled Vas, Re, Le, and BL, and Sd, but left the others the same as a single driver.
    thanks

    Comment


    • #3
      Re: Here's what I quickly modeled

      It's an ML-TL in a cabinet having internal dimensions of 6"W x 6"D x 36"H. The design center is 8" from the top (the center of the tweeter sandwiched between the two Jamo drivers). The port's center is located 2" above the bottom, having a length of 1.5" and a diameter of 2.5". I used a stuffing density of 1 lb/cu.ft., needing ~6 ounces total of stuffing in the top 18" of the cabinet. The system tuning frequency ended up at 58-59 Hz and I've attached the system response graph. You could set this same model up, then make changes to see if some improvements can be wrought.
      Paul
      Attached Files

      Comment


      • #4
        Re: help with buyout TL MTM idea?

        Finally got around to working on these. Not sure how much help they'll be to everyone else since both drivers are NLA (though a lot of the 5.5" Jamo woofers had similar t/s parameters), but I was fairly impressed for a pair of speakers that cost $51 in parts, not including cabinet (about another $45 in lumber).

        I'm still playing with the tuning, but right now have a pair of 1.5" ports at about 3.7" in them, which puts it in the mid 50s, based on impedance sweeps. i'll adjust as they break in and eventually post a final result, but it's decent as-is.

        The design is an MTM TL, with the 4-ohm woofers wired in series. The woofers are the Jamo 20435, the tweeters are the Jamo 20138, both available for around $5-6 each during buyout season. Sticking with the super-budget theme, they use a minimalist crossover that has 6 parts and costs $8 each.

        I will add a sketch, but the woofer is a 1st order electrical w/ a Zobel, and the tweeter is a 1st order electrical with a series padding resistor.

        woofer: .40 mH inductor in series, and (12 uF and 8 ohm Zobel across the woofer terminals)
        tweeter: 10 uF and 2.7 ohm in series with + lead.

        The cabinet is 7.5" w x 8.5" d x 36" tall, using .656" stock. (lousy home depot cheaping out on 3/4"....) the taper is about 18:1, with a 1/2" MDF divider running from 1/2" in front the back wall to a point equidistant from the top/front/back (i forget the precise length). These are a gift, and will likely be against or at least near a wall, so i made them front-ported, with a pair of the 1.5" buyout ports trimmed to 3.7" long.

        overall, they don't seem to have a lot of "air" in the top end, but it's definitely listenable, and you have to expect some flaws at this price point.

        my measurement system is currently in my home theater room and suffering a lof of reflections, but i'll try to post some measurements this weekend, though they won't be pretty.
        Attached Files
        Last edited by patrickm; 12-31-2010, 12:10 AM.
        TM (RS125/TB25-302), TM RS150S / Vifa BC25SC06, RS150-4 / Vifa ML-TL, 3CR-AL Ultra Budget: Electric Blue TM, TMM, MTM, Dragonflies , Mounties: Mini HT Satellite,sub for minions

        Comment


        • #5
          Re: help with buyout TL MTM idea?

          Looks good. I also built a MTM using those woofers, but a differant Jamo tweet. For what they are and how much I paid, I am happy with them. Decent detail and not too fatigueing. I need to pad the tweeter down a little more though.

          Your build looks nice. Alot better then how mine are sitting right now :D
          What's life without a little boom...

          Comment


          • #6
            Re: help with buyout TL MTM idea?

            Originally posted by Paul K. View Post
            With the Qts being very close to 0.4 (it will be a bit higher than 0.42 if you model with any added resistance for Radd), an optimum tuning frequency for flatest response down to just above the knee in the curve will likely be close to 60 Hz. So, using that as the 1/4-wavelength resonant frequency for a traditional, non-tapered line (and not an ML-TL), the required line length would be ~57". If you tapered it at 10:1, or if you made it an ML-TL, the line length needed would be ~34". Once you have the system tuning frequency optimized, you can increase the volume within the line to lower F3, but at some point with an ML-TL, the port length may indeed be nothing more than the wall thickness. It would help if you showed us what kind of response you're currently getting, however.
            Paul
            Don't mean to hijack the thread, but how do you come up with the volume for a TL? And how do you figure the slant and resulting volume and length.
            Thanks,
            Mike
            "We're speaker geeks, not speaker nerds. Nerds make money!" Marty H
            Bismarck, North Dakota
            My Current System: HiFiMe DIY T3 Amp, Kenwood Basic C1 Preamp, and Paul Carmody Sunflowers
            My Garage System TPS3116D2 Amp, DIY PS 95 Speakers, DC 130 Sub

            Comment


            • #7
              Re: help with buyout TL MTM idea?

              Just so it's clear, what I modeled for the OP was a simple ML-TL, not a tapered TL like he built. When I model a TL, whether tapered or mass-loaded, I start out by making the internal width of the cabinet equal to or slightly larger than the driver's flange diameter. Based on the driver's Fs and Qts I can easily calculate the likely optimum system tuning frequency (Fr), plus or minus a few Hz, that will give a flat response from the knee in the response curve on up to 1 kHz. Once I have a starting system tuning frequency I can calculate the required line length in inches whose 1/4-wavelength resonant frequency is equal to Fr by dividing 13600 by 4 times that frequency. This line length applies only to a non-tapered line, which I would never build. Instead I would taper it and use at least a 10:1 taper. A line tapered at 10:1 needs to be only ~60% as long as the equivalent non-tapered line. Arbitrarily I choose a starting line depth to go along with the already-chosen line width, then I model a line that tapers to 1/10 the starting depth, including some stuffing, and see what results. From that first exploration I will change the line length or taper or stuffing or line depth or line width or all of these parameters, until I get acceptable results in F3, response flatness and smoothness, and cabinet size/shape. An equivalent ML-TL can also be modeled by making its line length equal to the same 60% length, but keeping the cross-sectional area constant (no taper), and using a mass-loading port whose area is ~10% of the line's cross-sectional area, then adjusting the port's length and stuffing amount to get acceptable results.
              Paul

              Originally posted by mikejennens View Post
              Don't mean to hijack the thread, but how do you come up with the volume for a TL? And how do you figure the slant and resulting volume and length.
              Thanks,
              Mike

              Comment


              • #8
                Re: help with buyout TL MTM idea?

                Originally posted by Paul K. View Post
                Just so it's clear, what I modeled for the OP was a simple ML-TL, not a tapered TL like he built. When I model a TL, whether tapered or mass-loaded, I start out by making the internal width of the cabinet equal to or slightly larger than the driver's flange diameter. Based on the driver's Fs and Qts I can easily calculate the likely optimum system tuning frequency (Fr), plus or minus a few Hz, that will give a flat response from the knee in the response curve on up to 1 kHz. Once I have a starting system tuning frequency I can calculate the required line length in inches whose 1/4-wavelength resonant frequency is equal to Fr by dividing 13600 by 4 times that frequency. This line length applies only to a non-tapered line, which I would never build. Instead I would taper it and use at least a 10:1 taper. A line tapered at 10:1 needs to be only ~60% as long as the equivalent non-tapered line. Arbitrarily I choose a starting line depth to go along with the already-chosen line width, then I model a line that tapers to 1/10 the starting depth, including some stuffing, and see what results. From that first exploration I will change the line length or taper or stuffing or line depth or line width or all of these parameters, until I get acceptable results in F3, response flatness and smoothness, and cabinet size/shape. An equivalent ML-TL can also be modeled by making its line length equal to the same 60% length, but keeping the cross-sectional area constant (no taper), and using a mass-loading port whose area is ~10% of the line's cross-sectional area, then adjusting the port's length and stuffing amount to get acceptable results.
                Paul
                Thanks Paul!
                "We're speaker geeks, not speaker nerds. Nerds make money!" Marty H
                Bismarck, North Dakota
                My Current System: HiFiMe DIY T3 Amp, Kenwood Basic C1 Preamp, and Paul Carmody Sunflowers
                My Garage System TPS3116D2 Amp, DIY PS 95 Speakers, DC 130 Sub

                Comment


                • #9
                  Re: help with buyout TL MTM idea?

                  Is there a calculator or formula available for determining tuning frequency? The ones I've found require Vb. The way I'm understanding this is that I'm trying to find my Vb. Am I missing something??? I'm trying to figure a TL for a Tang Band W5-704D (Fs=55Hz, Qts=.38).
                  Sorry again, about hijacking. I can start a new post if need be.
                  Thanks,
                  Mike
                  "We're speaker geeks, not speaker nerds. Nerds make money!" Marty H
                  Bismarck, North Dakota
                  My Current System: HiFiMe DIY T3 Amp, Kenwood Basic C1 Preamp, and Paul Carmody Sunflowers
                  My Garage System TPS3116D2 Amp, DIY PS 95 Speakers, DC 130 Sub

                  Comment


                  • #10
                    Re: help with buyout TL MTM idea?

                    When I design TL, no matter its specific configuration, the system tuning frequency is based on Fs and Qts. When Qts is ~0.4, the system tuning frequency for an optimally flat overall response will be pretty close to Fs. If Qts is higher than 0.4, a better tuning frequency will be lower than Fs, and when Qts is lower than 0.4, the tuning frequency will be higher than Fs. For TLs, I prefer to work with drivers whose Qts is usually no lower than 0.3, preferably higher. Let's say we have a driver with Fs of 30 Hz and Qts of 0.4. I'd want the system tuning frequency to be ~30 Hz. If the line were neither tapered nor mass-loaded, its length would need to have a 1/4-wavelength resonant frequency also of ~30 Hz. Multiplying 30 x 4, then dividing that into 13600 gives a line length of 113". If I taper the line at 10:1, the required length will be ~0.6 x 113", or 68". If the driver's Qts were higher than 0.4, I'd make the line longer (and/or its taper larger) so its resonant frequency would be lower than 30 Hz, and if Qts were lower than 0.4, I'd make the line shorter (and/or have less taper) to make its resonant frequency be higher than 30 Hz. The volume contained within the line will ultimately determine how low F3 is, and I get to the F3 I want by fiddling with cross-sectional area changes and stuffing density/length.

                    For your W5-704D, with its Qts of 0.38, you'd likely want the line's resonant frequency (or system tuning frequency for an ML-TL) to be 50-55 Hz. Shooting for a resonant frequency of 52.5 Hz, a non-tapered line would need to be ~65" long. If tapered at 10:1, its length becomes ~39", and an equivalent ML-TL would also be ~39" long with the area of its mass-loading port equal to ~10% of the line's cross-sectional area. All of this is best accomplished by modeling, and I highly recommend using Martin King's MathCad-based modeling worksheets. I've told you how I model and what I normally try to achieve, but others have different criteria based on their personal goals that are equally valid.
                    Paul

                    Originally posted by mikejennens View Post
                    Is there a calculator or formula available for determining tuning frequency? The ones I've found require Vb. The way I'm understanding this is that I'm trying to find my Vb. Am I missing something??? I'm trying to figure a TL for a Tang Band W5-704D (Fs=55Hz, Qts=.38).
                    Sorry again, about hijacking. I can start a new post if need be.
                    Thanks,
                    Mike

                    Comment


                    • #11
                      Re: help with buyout TL MTM idea?

                      Originally posted by mikejennens View Post
                      Don't mean to hijack the thread, but how do you come up with the volume for a TL? And how do you figure the slant and resulting volume and length.
                      Thanks,
                      Mike
                      there are a lot of ways to figure these things out in different sequences, but to estimate the volume, it's just the cabinet volume like a normal cabinet, minus the diagonal piece's volume. if you set the top of the divider halfway back in the cabinet, the taper ratio should be the same as the area on either side of it at the base (i.e., 1 square inch to 10 square inches). the length is really measured along the center of the volume, so it's usually a bit above the length up and down the divider. to minimize disruptions in the changing cross section as you would go from the small end to the large end, i try to keep the top of the divider the same distance from the front baffle, back baffle, and top of the cabinet. without little corner braces (like the tritrix cabinet), it's not perfect, but it's not terrible in my experience.

                      i found paul's initial model (and his other posts since) very helpful in explaining his design process (and if i haven't said so before, thanks again). I did not build exactly what he suggested originally, but used it as a solid starting point for a project i was using to learn a few things on my own. i had the old version of martin's worksheets, and went from there. a lot of it was just watching changes in the curves as i modified different variables, trying to get a sense of how/why/how much some of them affected the modeled response.

                      also, no worries about hijacking the thread. since these are low-stock, buyout drivers, i figured this was as much to document some of the outcome of the previous assistance as anything else. (though some of the other Jamo's probably would fill in with some crossover mods).
                      TM (RS125/TB25-302), TM RS150S / Vifa BC25SC06, RS150-4 / Vifa ML-TL, 3CR-AL Ultra Budget: Electric Blue TM, TMM, MTM, Dragonflies , Mounties: Mini HT Satellite,sub for minions

                      Comment


                      • #12
                        Re: help with buyout TL MTM idea?

                        I'd sure like to see your modeling worksheets just out of interest. I would make one suggestion, though, for future tapered TLs you might build. It's always been common practice to place the angled divider's end just as you described, spaced equally from the baffle, back panel and top (or bottom) of the cabinet (and, as you noted, there is absolutely no reason to place angled deflectors at the two corners, either). But when you place the divider's end this way, you obviously create a section in the middle of the line's length that has no taper. From the extensive modeling I've done, it's actually better to offset the divider's end this way: make the distance from the baffle to the front of the divider's end be larger than the distance from the end of the divider to the top (or bottom) of the cabinet, and make the distance between the back of the divider and the back panel of the cabinet even less. This way you'll end up with three shorter straight section that will effectively be "tapered" due to the successively smaller cross-section areas they'll have. This will not be a large difference but its effect is visible and beneficial.

                        I'm glad I was able to be of help.
                        Paul

                        Originally posted by patrickm View Post
                        there are a lot of ways to figure these things out in different sequences, but to estimate the volume, it's just the cabinet volume like a normal cabinet, minus the diagonal piece's volume. if you set the top of the divider halfway back in the cabinet, the taper ratio should be the same as the area on either side of it at the base (i.e., 1 square inch to 10 square inches). the length is really measured along the center of the volume, so it's usually a bit above the length up and down the divider. to minimize disruptions in the changing cross section as you would go from the small end to the large end, i try to keep the top of the divider the same distance from the front baffle, back baffle, and top of the cabinet. without little corner braces (like the tritrix cabinet), it's not perfect, but it's not terrible in my experience.

                        i found paul's initial model (and his other posts since) very helpful in explaining his design process (and if i haven't said so before, thanks again). I did not build exactly what he suggested originally, but used it as a solid starting point for a project i was using to learn a few things on my own. i had the old version of martin's worksheets, and went from there. a lot of it was just watching changes in the curves as i modified different variables, trying to get a sense of how/why/how much some of them affected the modeled response.

                        also, no worries about hijacking the thread. since these are low-stock, buyout drivers, i figured this was as much to document some of the outcome of the previous assistance as anything else. (though some of the other Jamo's probably would fill in with some crossover mods).

                        Comment

                        Working...
                        X