I have a few questions about designing transmission line or daline enclosures. I hope someone will be kind enough to help.
I am investigating using the Jordan JX92S in a daline enclosure. The specs for the driver are here: http://www.ejjordan.co.uk/jx92.html. The company publishes plans for a box that looks essentially like a daline, although they call it a transmission line. The specs are on this page (scroll down some): http://www.ejjordan.co.uk/enclosures.html.
I might just use those plans. However, I am a little concerned about the fact that the design allows for conversion between a transmission line (or daline) configuration and a ported configuration. I would like to verify that the dimensions are indeed correct for the TL configuration. I am also considering a different daline configuration entirely.
My question first question is, beginning with little knowledge of transmission line design principles, and without easy (free or cheap) access to MathCad, do I stand a chance? (By the way, I tried using the demo version of MathCad 8 with Martin King's free software, but on my computer [MS Windows 2000 Pro] the MathCad demo wouldn't do much except crash.)
I've read the original daline article: http://tech-diy.com/LS_Articles.htm. The author says to tune the box to about 1.5 times the tuning frequency of the transmission line, but he gives no instruction on how to do that. The box is not "ported" because the line is long -- equivalent to Fs/4. But it's not exactly sealed either. Perhaps it's "aperiodic." (I told you I don't know much about this stuff.) Can anyone tell me how to size the box? And should the line be tuned somewhat below Fs? And how to tune the line?
Best regards,
Dave
I am investigating using the Jordan JX92S in a daline enclosure. The specs for the driver are here: http://www.ejjordan.co.uk/jx92.html. The company publishes plans for a box that looks essentially like a daline, although they call it a transmission line. The specs are on this page (scroll down some): http://www.ejjordan.co.uk/enclosures.html.
I might just use those plans. However, I am a little concerned about the fact that the design allows for conversion between a transmission line (or daline) configuration and a ported configuration. I would like to verify that the dimensions are indeed correct for the TL configuration. I am also considering a different daline configuration entirely.
My question first question is, beginning with little knowledge of transmission line design principles, and without easy (free or cheap) access to MathCad, do I stand a chance? (By the way, I tried using the demo version of MathCad 8 with Martin King's free software, but on my computer [MS Windows 2000 Pro] the MathCad demo wouldn't do much except crash.)
I've read the original daline article: http://tech-diy.com/LS_Articles.htm. The author says to tune the box to about 1.5 times the tuning frequency of the transmission line, but he gives no instruction on how to do that. The box is not "ported" because the line is long -- equivalent to Fs/4. But it's not exactly sealed either. Perhaps it's "aperiodic." (I told you I don't know much about this stuff.) Can anyone tell me how to size the box? And should the line be tuned somewhat below Fs? And how to tune the line?
Best regards,
Dave
I just downloaded that demo because I am trying to analyze a Daline using Focal drivers.
My suggestion would be to post your MathCAD problem on the Everything Else forum, where more people can see it and they might come up with something.
There has been talk of starting a SPICE/Simulation Program forum by people in the Everything Else section, so those people should be thinking right along with you.
Re: Re: design of dalines (transmission lines)
If you'll run it on the Jordan JX92S specs, I'll help you move a... well, I'll help you move.
I'll give it a try, but my guess is that the program I have will only run on an old system, like Win 95 or something.
If some folks here want to do a project, I have some optimization software that's better than anything you can find on the net for freebees. It's in C++.
kelticwizard said:
If you'll run it on the Jordan JX92S specs, I'll help you move a... well, I'll help you move.
I'll give it a try, but my guess is that the program I have will only run on an old system, like Win 95 or something.
If some folks here want to do a project, I have some optimization software that's better than anything you can find on the net for freebees. It's in C++.
Haven't read the article and have too slow a connection to bother with downloading it, but though I'd never heard of Dalines until I got on-line, I've been building them off n' on for several decades as a way to extend the LF response of low Q/low Fs drivers, ergo my understanding of the concept wouldn't work well with higher Q/Fs drivers since with increasing Q/Fs, the basic body of the speaker (Vb) increases, with both combining to shorten the TL portion (vent).
Anyway, I find the max flat alignment, then increase it 1.4141x to find the net Vb, then add a long vent ~Sd in area and long enough to tune the cab/TL to Fs. Since the vent is in series with the cab, it's the sum of the lengths minus the end correction (radius*0.613) that yields the right tuning. If you add a vent that's a 1/4WL long, it will be tuned way too low. Obviously, the most accurate results will be when you use actual T/S measurements Vs published.
FWIW, I designed a JX92S ML-TL (aka MLTQWT) for Jay Fisher over on the fullrange forum and he's very pleased, saying that the pipe provides enough gain to not feel the need for any baffle step compensation (BFC), a big plus with such a low efficiency driver IMO.
WRT the mathcad problem, if you have Norton AV 2003 or other comprehensive AV program, try turning off 'scripting' to load the spreadsheets. Of course don't do this while on-line or get back on without resetting it. This is the only way I can load his later spreadsheets.
GM
Anyway, I find the max flat alignment, then increase it 1.4141x to find the net Vb, then add a long vent ~Sd in area and long enough to tune the cab/TL to Fs. Since the vent is in series with the cab, it's the sum of the lengths minus the end correction (radius*0.613) that yields the right tuning. If you add a vent that's a 1/4WL long, it will be tuned way too low. Obviously, the most accurate results will be when you use actual T/S measurements Vs published.
FWIW, I designed a JX92S ML-TL (aka MLTQWT) for Jay Fisher over on the fullrange forum and he's very pleased, saying that the pipe provides enough gain to not feel the need for any baffle step compensation (BFC), a big plus with such a low efficiency driver IMO.
WRT the mathcad problem, if you have Norton AV 2003 or other comprehensive AV program, try turning off 'scripting' to load the spreadsheets. Of course don't do this while on-line or get back on without resetting it. This is the only way I can load his later spreadsheets.
GM
GM said:
Thank you very much for the good information.
I am not familiar with the term "max flat alignment." Remember, I'm a tyro.
On the subject of tuning the pipe, I was aware that the pipe would be shorter than 1/4 Fs, both because of the box, and also because of damping material and pipe taper if any. But I still don't know how to tune it.
If you could say a bit more on the subject, I would appreciate it very much.
Dave
P.s. Are the plans for the MLTQWT available? Online?
Augspurger defines a line geometry, where given the volume of the TL, you use a 1/3 of the volume for a chamber behind the driver and the rest of the volume for a line tuned to the driver Fs. This geometry improves the low pass function on the terminus vrs a straight line.
I always associated this geometry with a daline.
dave
I always associated this geometry with a daline.
dave
Attachments
planet10 said:
Glancing back at http://www.ejjordan.co.uk/enclosures.html , it does appear that about 1/3 of the volume is in the box, 2/3 in the pipe.
In the end it all means plug the numbers into the model and see how it does... Augspurger has a parallel model to King's. They were both developed at the same time, and the 2 models generate similar results. Augspurger's model is not freely available, so we we're back to Martin's software...
or i guess you could get in the ballpark using Martin's tables to model a straight line and then rejigger the subsequent volume.
The length of the line (+ end correction) should be 4L=speed of sound/Fs
dave
or i guess you could get in the ballpark using Martin's tables to model a straight line and then rejigger the subsequent volume.
The length of the line (+ end correction) should be 4L=speed of sound/Fs
dave
It's the T/S alignment (tuning) that's yields the flattest response the driver can achieve in a golden ratio vented cab. Drivers with a Qts = 0.403 has its flattest response at Vb = Vas and Fb = Fs. Below this point, Vb = <Vas and Fb = >Fs. Conversely, Above this point, Vb = >Vas and Fb = <Fs. Every box program I've used has either a default or max flat choice, calculating the Vb/Fb, so it's not like you have to 'do the math' anymore beyond multiplying the Vb*1.414.
Damping material should be limited to the basic cab, so it won't affect tuning. if a pipe taper is used then this will rquire you to model it in the Mathcad program for best results WRT sims. Once built, if it's tuned too low and folded up in a cab, then nothing comes to mind at the moment to tune it higher. If too high, then an extension can be made or tuning it aperiodically will lower it.
The actual measurement requires doing an impedance plot, or if absolute accuracy isn't required you can just do a slow sinewave sweep and when the driver basically stops and all the output is coming out of the vent, you've found the tuning frequency.
No, they're not on-line, I just gave him this info and he built them:
31.04"L x 6.97"w x 4.31"d
driver down from top 11.14"
vent = 2"dia x 5"Long
stuffing = 0.54lb R-19 fiberglass or equal distributed behind driver and towards the top
All dims are and inside and of course can be rounded off to the next common fraction, and since the driver's well below ear height when seated and the vent's longer than either the width/depth, I recommend extending it out the bottom and incorporating it into the stand. I don't know if he had to adjust the vent length or not, but since this is mostly dependent on the room it's in, not really relevant for you.
GM
Damping material should be limited to the basic cab, so it won't affect tuning. if a pipe taper is used then this will rquire you to model it in the Mathcad program for best results WRT sims. Once built, if it's tuned too low and folded up in a cab, then nothing comes to mind at the moment to tune it higher. If too high, then an extension can be made or tuning it aperiodically will lower it.
The actual measurement requires doing an impedance plot, or if absolute accuracy isn't required you can just do a slow sinewave sweep and when the driver basically stops and all the output is coming out of the vent, you've found the tuning frequency.
No, they're not on-line, I just gave him this info and he built them:
31.04"L x 6.97"w x 4.31"d
driver down from top 11.14"
vent = 2"dia x 5"Long
stuffing = 0.54lb R-19 fiberglass or equal distributed behind driver and towards the top
All dims are and inside and of course can be rounded off to the next common fraction, and since the driver's well below ear height when seated and the vent's longer than either the width/depth, I recommend extending it out the bottom and incorporating it into the stand. I don't know if he had to adjust the vent length or not, but since this is mostly dependent on the room it's in, not really relevant for you.
GM
planet10 said:
I used the Augspurger's alignment tables in Loudspeaker Design Cookbook by Dickason, and I came up with these numbers:
F3 = 72Hz (yuck)
Line length = 38" (seems short. I used the table on page 93, which assumes the line is stuffed.)
Total volume = 11.25 liters
I don't know what to make of those numbers.
I'm trying to use Martin King's alignment tables. They call for a speaker parameter Bl. Is that the "force factor" in these parms? http://www.ejjordan.co.uk/jx92.html If not, how do I calculate or estimate Bl?
Thanks.
Thanks.
Yes, B*L is usually abbreviated to BL or Bl. When not spec'd it can be calculated:
Cms = Vas/(Sd^2*rho*c^2)
where:
Vas needs to be in m^3, so multiply liters by 0.001
Sd needs to be in m^2, so multiply cm^2 by 0.0001
rho = density of air in kg/m^3, or ~1.21 at a typical room temp
c = speed of sound in meters, or ~344.42 at room temperature
Mms = 1/[(2*pi*Fs^2)*Cms)]
BL = {[2*Pi*Fs*Re*(Mms*10^-3)]/Qes}^0.5
Note that you'll get higher values than Lspcad and some other programs
predict since they use an unrealistically low rho/c IMO. I mean, how many of
us have a listening room temp/humidity down to where the drivers are in
a standard temp/pressure (59deg/29.92") environment? If anything they're in
a higher temp/pressure than I use, but if it's good enough for Tom Danley,
it's good enough for me. :^)
GM
Cms = Vas/(Sd^2*rho*c^2)
where:
Vas needs to be in m^3, so multiply liters by 0.001
Sd needs to be in m^2, so multiply cm^2 by 0.0001
rho = density of air in kg/m^3, or ~1.21 at a typical room temp
c = speed of sound in meters, or ~344.42 at room temperature
Mms = 1/[(2*pi*Fs^2)*Cms)]
BL = {[2*Pi*Fs*Re*(Mms*10^-3)]/Qes}^0.5
Note that you'll get higher values than Lspcad and some other programs
predict since they use an unrealistically low rho/c IMO. I mean, how many of
us have a listening room temp/humidity down to where the drivers are in
a standard temp/pressure (59deg/29.92") environment? If anything they're in
a higher temp/pressure than I use, but if it's good enough for Tom Danley,
it's good enough for me. :^)
GM
GM said:
He uses his software model. His approach seems to be to calculate the volume and then play with the geometry to improve the terminus low-pass function.
The alphaTL presented by Rick Shultz would also be a starting point. Rick uses's George's software. I have has the alphaTL paper for sometime and made up a spreadsheet to calculate lines. I have no explicit readout for the line volume, but i have noted that as you shorten the line the cross-section gets bigger (this harks back to John Cockroft's empirical stuff -- he too has been getting input from George of late -- i'm sitting on an unpublished article i need to OCR in and put up on my site of a project with GA's input).
I get a chance. i'll run a little experiment to see if the volume of an alpha TL stays the same as the geometry is changed.
dave
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.