Crossover frequency for transmission line

Hi there
I am wanting to build a 3 way speaker with a transmission line as bass maybe from about 30 hz tuning frequency using a Morel Ultimate UW1058 (25hz to 800hz)

with a conventional sealed box for a midrange - using a Morel Supreme SCM634 ( 50Hz -7kHz)

and a tweeter ( not concerned with that yet)
My question is 2 fold
1) Where would I look to cross the bass to midrange , there is no info out there as to how wide (octaves) the bass band can be with a TL speaker

2) if I cross to the midrange at say 150 hz do I tune the midrange box to the 50 hz resonant frequency as claimed by the speaker manufacturer or do I tune the box closer to the crossover frequency I want?

thanks in advance
Clive
 
  • Like
Reactions: 1 user
My experience from recent simulations has been that there's fairly continuous behavior between vented/ported/bass reflex configurations and TLs where you can trade off size (vented -, TL +) for port air speed (vented +, TL -).

You could start by simming a recommended vented box configuration for that woofer, see what happens when you extend the port to a quarter wave length (286cm by my calc), and then adjust the port and box dimensions to where the response (which will be very wonky) starts to nudge towards what you would want it to be.

Another possible approach: start with the most naive possible transmission line as a simulation and add stuffing to get the resonances somewhat tamed. This approach has the advantage of giving you a more direct picture of what kind of bandwidth you can expect. From the graph attached, the first null generated by pipe resonance centers around 90Hz.

So far in my sims and reading I've run into four main ways you might improve that, in rough order of how often they're talked about here: mass loading (MLTL), tapering, driver offset, and a coupling chamber. Look those up if it looks like getting rid of that first null might be getting you in the ballpark of the response you're looking for.
 

Attachments

  • morel-uw-1058-striaght-pipe.png
    morel-uw-1058-striaght-pipe.png
    13 KB · Views: 301
  • morel-uw-1058-striaght-pipe-2.png
    morel-uw-1058-striaght-pipe-2.png
    6.1 KB · Views: 290
  • morel-uw-1058-striaght-pipe-4.png
    morel-uw-1058-striaght-pipe-4.png
    15.4 KB · Views: 291
Last edited:
  • Like
Reactions: 1 user
To answer your questions more directly:

1) You could cross as high as let's say 500Hz or higher if you wanted, if that would serve other design goals like reducing midrange excursion. At 500Hz crossover and 30Hz tuning it looks like you'd be exposing the listener to 4 nulls. Those nulls get progressively shallower as you move up in frequency, although group delay around them will get worse. The depth and the group delay can be addressed by more stuffing, at the cost of reduced bottom end efficiency.

2) You could probably tune the midrange box higher than the recommended 50Hz and get a slight efficiency boost, at the possible cost of phase and group delay issues (higher Q/boominess).
 
If the TL is designed properly, optimized for the intended woofer and built as designed the TL itself will not limit where you can cross the woofer to a midrange anymore than it would be limited if the woofer were in a sealed or typical vented box. Of course coming up with a proper and optimized design can only be accomplished if you have accurate software to use for modeling the design, like Martin King's software.
Paul
 
Hi Altie

looking at your diagram 3 those null frequencies start at around 80 hz about one and a half octaves above the tuned frequency ( these definitely do not look good on sound quality)
so maybe one needs to look at a 4 way setup and cross over at the 90 hz mark as per
Speakers: Principles

Paul and Altie will a sealed bass show these same nulls ( I am sure not)

thanks again
 
Of course a sealed box will not have nulls like this but that's not the point. If you correctly design an optimum TL, most likely a tapered one to minimize its length along with appropriate stuffing, the nulls will be optimally smoothed out and all but negligible, allowing crossing as desired. With the Morel woofer the OP wants to use and its fS and Qts, the optimum tuning frequency (1/4-wavelength resonant frequency from the line's length and taper ratio) is going to be at 18-20 Hz, BTW.
Paul

Hi Altie

looking at your diagram 3 those null frequencies start at around 80 hz about one and a half octaves above the tuned frequency ( these definitely do not look good on sound quality)
so maybe one needs to look at a 4 way setup and cross over at the 90 hz mark as per
Speakers: Principles

Paul and Altie will a sealed bass show these same nulls ( I am sure not)

thanks again
 
Pete
excuse my ignorance here , just recently bought the loudspeaker cookbook :)
but what is the baffle step frequency and how do I get this from the specs ? https://www.morelhifi.com/wp-content/uploads/2013/11/UW1058.pdf

""tune the midrange box" - so it gives the flattest response in the passband, combined with the rolloff you want below the LF crossover frequency" so you are saying make my midrange box to suit an Fs of the crossover frequency and not of the midrange speaker which is lower than the crossover

Paul are you saying with a tapered line that is well padded those null areas will not be apparent? what frequency would you cross over at?
 
Thank you. That 18-20Hz tuning sounds plausible to me from the sims I've been doing lately, but for my own education, how'd you arrive at it?

@Cliveinuk, like pkitt is saying, yes those nulls would destroy midrange qualities, no you don't have to tolerate them. Look up some details on mass loading, tapering, driver offset, coupling chambers and stuffing, and try running sims so you can see how those approaches would help. The numbers and graph I posted are for the most basic possible box you could call something like a transmission line for the driver you posted - *not* one that's optimized like pkitt is talking about, or one you would actually want to use. It's a starting point.
 
Baffle step is the frequency where the sound starts radiating into a hemisphere rather than a flat plane, it depends on your baffle width (Fbsc =115/baffle width in metres). You don't have to crossover at this frequency, but it considerably simplifies the crossover design.

re:' you are saying make my midrange box to suit an Fs of the crossover frequency and not of the midrange speaker which is lower than the crossover" - yes (although I wouldn't put it quite like that, Fs pertains to a driver only, not a box or crossover)
 
It also depends on the eventual speaker placement. Baffle Step Correction is only useful if the speakers are a fair distance from any walls.

If the speakers end up close to a wall behind them BSC does more damage than good ie bloated ugly bass.

If they end up close to a corner you will need to do the opposite of BSC: reduce the relative bass level, not increase it.

Due to circumstances mine are in corners and I had to reduce the bass by nearly 8dB for a flat response at my listening position.
 
My speakers sit on a floor-level fireplace "hearth" with their backs 5 to 9 inches from the wall, depending on cabinets' depths (with almost all of the cabinets I build having the terminus or ML-port on the rear panel) and I've found that incorporating ~ 3 dB of BSC works just fine with well-defined and not bloated or ugly bass. See the attached photo of my setup.
Paul


It also depends on the eventual speaker placement. Baffle Step Correction is only useful if the speakers are a fair distance from any walls.

If the speakers end up close to a wall behind them BSC does more damage than good ie bloated ugly bass.

If they end up close to a corner you will need to do the opposite of BSC: reduce the relative bass level, not increase it.

Due to circumstances mine are in corners and I had to reduce the bass by nearly 8dB for a flat response at my listening position.
 

Attachments

  • 100_0200.JPG
    100_0200.JPG
    189.2 KB · Views: 141
When Qts is higher than 0.4, the optimum tuning frequency will be lower than fS, which is specified as 26 Hz. So I guessed at 20 Hz, did some quick modeling and found a little bit lower would be even better.
Paul

Thank you. That 18-20Hz tuning sounds plausible to me from the sims I've been doing lately, but for my own education, how'd you arrive at it?

@Cliveinuk, like pkitt is saying, yes those nulls would destroy midrange qualities, no you don't have to tolerate them. Look up some details on mass loading, tapering, driver offset, coupling chambers and stuffing, and try running sims so you can see how those approaches would help. The numbers and graph I posted are for the most basic possible box you could call something like a transmission line for the driver you posted - *not* one that's optimized like pkitt is talking about, or one you would actually want to use. It's a starting point.
 
10" sub and a 6" midrange per side.... Looking at the response and impedance plots, cross somewhere around 200Hz. The subwoofer impedance rises and the response gets a little bumpier higher up.

Baffle Step or Diffraction loss describes what happens around the frequency where the wavelength of sound waves approaches that of the baffle size. Freq below that tend to wrap around the speaker and radiate in all directions. Freq above it tend to project off the baffle forwards. The response is adjusted to attenuate the higher frequencies therefore apparently increasing the lower freqs.

"The Edge" program can be useful in identifying and correcting.
Tolvan Data.

J.
 
Modeling results

While there are likely other equally usable solutions, I modeled a line 80" long and tapered at 24:1 having a starting area of 180 in2 and an ending area (as well as the terminus) of 7.5 in2, with the the center of the UM1058 located 20" from the beginning of the line. The first 54" of the line is stuffed with polyester fiber at a density of 1 lb/ft3. Without the stuffing, the line is tuned to 21-22 Hz. After adding stuffing the tuning frequency becomes ~18 Hz. I've attached the modeling graph showing the predicted anechoic system bass response (red line) for an input of 150w/1m, which shows an f3 of ~21 Hz and an output SPL of 104 dB above 30 Hz with a very smooth response. At this input power, the woofer reaches Xmax at ~25 Hz, then exceeds it at ~17 Hz and all lower frequencies. Also with this input power, the peak air velocity in the end of the line and terminus is ~4.5% of the speed of sound, or ~15 m/s, and occurs at ~17 Hz.
Paul
 

Attachments

  • UW1058.gif
    UW1058.gif
    6.5 KB · Views: 134
Paul, when you say the woofer centre is 16 inches from the beiginning of the line do you mean there is a box that is 16 inches long and does not count as part of the line length

here is a question ? does the baffle step diffraction get influenced by the material used in the baffle
So for instance 12mm mdf vs 25 mm concrete as extreme examples Will the concrete exibit this to a much lower extent?

Speaker design is not as simple as ABC then

thanks again to everyone for the effort in replying

Clive