How to use Martin J. King's software to design a TL loudspeaker?

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


I've downloaded mathcad explorer (12Mb with my 56K, ouch!) and Martin King's transmission line models


But I don't know what to do with it. I entered the driver's characteristics, and the experimental dimentions of my project



But how do I use it to tune my tranmission line dimentions?
I look at the curves, but since I'm still a noob, I don't know what each curve mean
And the way they sould look, after having optimaly set my parameters.

Can someone help me?
Thanks

Alex
 
frugal-phile™
Joined 2001
Paid Member
Bricolo said:
But how do I use it to tune my tranmission line dimentions?
I look at the curves, but since I'm still a noob, I don't know what each curve mean
And the way they sould look, after having optimaly set my parameters.

I'm sure Martin will have more tips, but it involves playing with the box parameters and looking at the output... i tend yo primamrily look at the summed response (system response) trying to achieve a curve that is extended towards the bottom (and above the sealed box line, with little ripple. Once i get close i'll start paying attention to the phase characteristics and the impulse response.

dave
 
Dave's method is almost exactly what I do when iterating a design. The only other curve I watch closely is the acoustic impedance. The first large hump tells me at what frequency the enclosure is tuned. I try and keep this peak at or just below the driver resonance.

When I am designing, I start with a first guess (I recommend reading Bob Brines' site for a better understanding and some recommended first guesses) and then change one variable at a time. If the response improves, I keep the change and try a little more. I continue to iterate until I have a design that looks to be optimized and can't get much better. Then I start again and work in a new direction. After a few sessions, I have several designs that I have saved. Based on other requirements (size, impulse response, construction challenge, low frequency response, ripple, and so on) I select the design I want to build. After a while you get a feel for what will work and the excercise moves along much faster.

No magic,
 
Ok; so I must tune my params to obtain a ripple free system response, that extends to the frequency I want


For the phase, how must it look? What phase "response" is bad, which one is good?


PS: what are the dotted blue line and the solid red line? According to which one do I tune my params?
 
As far as I understand, the blue line is a reference line, reflecting how your driver would react in an infinite baffle system.

The two big deals, when I built mine with MJKs mathcad stuff, were the very first graph and one further down. As MJK says, the "resulting Acoustic Impedance..." hump is the tuning frequency you're aiming for. The second graph of utmost importance is the second graph under the heading "Far Field Transmission Line System and Infinite Baffle Sound Pressure Level Responses." That particular graph reflects your drivers' bass response.

Once you have your driver specs entered, experiment with the length of the pipe and the amount of stuffing until you can smoothe out the ripples in the second one while maintaining the location of the "hump" in the first one where you want it.

Of course, you also want to keep your pipe relatively long, too. It's a great program.

Dave

Shamelessly advertising my completed project: http://www.diyaudio.com/forums/showthread.php?s=&threadid=12769
 
PS: what are the dotted blue line and the solid red line? According to which one do I tune my params?

In most cases the title just above the plot defines the variables being shown. The first item in the title corresponds to the solid red line and the second item in the title corresponds to the blue dashed line. In most plots I show the TL response by a red solid line and the equivalent infinite baffle response in a blue dashed line. I think that the only exception for this rule is the plot showing the driver output (red) and the terminus output (blue) for the TL system.

Hope that helps,
 
Thanks :)


Martin, what are your starting points in a TL designing?
I mean, after having picked up your driver's characteristics, how do you estimate the first value you enter for your TL lenght and area?

Do you have a rule of thumb for this, and then you use your software to tune it perfectly, or do you arbitrary set tue line lenght (I don't think so :))

And the same question for the bass response you want (I mean, the frequency you set the f3). Do you use the driver's free air resonance frequency, and keep this value for the aimed f3 frequency, something higher, lower, totally independant?
 
Martin, what are your starting points in a TL designing?

To be honest, I usually put the driver T/S parameters into an existing design worksheet for something I previously did with a similar driver. The response does not look very good so I start changing things. For my latest projest (Lowther DX series) I put together and optimized designs for a wide variety of enclosure styles including bass reflex, all of the various TL's including ML TQWT, back and front loaded horns, and double bass reflex. The drivers worked well in some and not so well in others. I ended up with three promising configurations and I picked the easiest to build.

For bass response I try and get as extended a bass response as possible allowing a 2 or 3 dB downward tilt below 100 Hz that will end up compensated for by room reinforcement. I try and minimize the ripple above 100 Hz. Recognize that in a real room you bass efficiency will be 3 - 6 dB below the value shown by the MathCad worksheet due to baffle step behavior.

My advise is to just jump in and try different conbinations. The first design will take a while to work out as you get a feel for what direction a change in geometry will take the response. Be patient, come back a few times after thinking for a couple of hours or even days, and don't expect that things will magically fall in place after an hour or two of experimenting. Sometimes all of the TL designs look bad if a driver is not suited for this type of enclosure.

I strongly recommend that you read Bob Brines' site (linked from mine) from start to finish to get a feel for what can be done. If you follow what he has done you will have a fairly decent understanding of what is possible and important in TL design.

Hope that helps,
 
MJK said:

The first design will take a while to work out as you get a feel for what direction a change in geometry will take the response. Be patient, come back a few times after thinking for a couple of hours or even days, and don't expect that things will magically fall in place after an hour or two of experimenting.



For what it's worth, my first try with the MJK Mathcad worksheets took me about 25 hours over several days, but the results are to die for.

Dave
 
MJK said:


To be honest, I usually put the driver T/S parameters into an existing design worksheet for something I previously did with a similar driver. The response does not look very good so I start changing things. For my latest projest (Lowther DX series) I put together and optimized designs for a wide variety of enclosure styles including bass reflex, all of the various TL's including ML TQWT, back and front loaded horns, and double bass reflex. The drivers worked well in some and not so well in others. I ended up with three promising configurations and I picked the easiest to build.

For bass response I try and get as extended a bass response as possible allowing a 2 or 3 dB downward tilt below 100 Hz that will end up compensated for by room reinforcement. I try and minimize the ripple above 100 Hz. Recognize that in a real room you bass efficiency will be 3 - 6 dB below the value shown by the MathCad worksheet due to baffle step behavior.

My advise is to just jump in and try different conbinations. The first design will take a while to work out as you get a feel for what direction a change in geometry will take the response. Be patient, come back a few times after thinking for a couple of hours or even days, and don't expect that things will magically fall in place after an hour or two of experimenting. Sometimes all of the TL designs look bad if a driver is not suited for this type of enclosure.

I strongly recommend that you read Bob Brines' site (linked from mine) from start to finish to get a feel for what can be done. If you follow what he has done you will have a fairly decent understanding of what is possible and important in TL design.

Hope that helps,
Actually, I know the design I want. Not the dimentions :)

I want to build a tapared ( 4:1) transmission line, folded once

Do you see how the Thor are made? My project looks the same, but with only 1 midbass driver (+ 1 tweeter), other dimentions, and a 4:1 taper
 
and with an image:
 

Attachments

  • sp.gif
    sp.gif
    3.1 KB · Views: 1,129
The dimensions will fall right into place, then. Enter a "4" in the SL "Area of the large end" Sd parameter, then a "1" in the SO "small end" Sd parameter.

After that, simply adjust pipe length, driver position, port radius and stuffing parameters until you see the desired tuning in the first graph and desired ripple in the response graph.

I'm assuming you're using the ML TQWT mathcad worksheet.

Bob Brines' website actually walks you through the mathcad worksheets, with full explanations and critique. http://geocities.com/rbrines1/

Dave
 
kneadle said:
The dimensions will fall right into place, then. Enter a "4" in the SL "Area of the large end" Sd parameter, then a "1" in the SO "small end" Sd parameter.

After that, simply adjust pipe length, driver position, port radius and stuffing parameters until you see the desired tuning in the first graph and desired ripple in the response graph.

I'm assuming you're using the ML TQWT mathcad worksheet.

Bob Brines' website actually walks you through the mathcad worksheets, with full explanations and critique. http://geocities.com/rbrines1/

Dave


I'm using the offset speaker worksheet
why should I use the ML TQWT one? If I understood correctly, TQWT are expanding transmission lines, with a port at the end.
My drawing shows a tapered TL
 
frugal-phile™
Joined 2001
Paid Member
Bricolo said:
I'm using the offset speaker worksheet
why should I use the ML TQWT one? If I understood correctly, TQWT are expanding transmission lines, with a port at the end.
My drawing shows a tapered TL

All the worksheets are actually the same but with the input area set up a bit different. The ML-TQWT sheet has the set-up that would be most convienient to simulate the line you are proposing.

dave
 
planet10 said:


All the worksheets are actually the same but with the input area set up a bit different. The ML-TQWT sheet has the set-up that would be most convienient to simulate the line you are proposing.

dave


To add a bit, "expanding" and "tapered" depends on the Sd parameters you enter. If you want an expanding-QWT, then your closed end to open end ratio will be less than 1; whereas, if you want a tapered-QWT, your closed end to open end ratio will be greater than 1. Yours would be 4:1.

The T/S specs you entered in the top part of the worksheet include the Sd (surface area) of the driver, so the number you enter in the SO and SL fields are multiplications of the area of your driver.

Dave

**Edit** I should note that I'm using the term "tapered" and "expanding" to describe the shape of the pipe as it moves away from the driver. I think I'm using "tapered" backwards from the common use of the term.
 
All the worksheets are actually the same but with the input area set up a bit different. The ML-TQWT sheet has the set-up that would be most convienient to simulate the line you are proposing.

BUSTED! Yes, almost all of the worksheets are the same calculation algorithm but with different input areas. People were getting confused trying to enter some of the standard geometries so I set up "dedicated" worksheets with a minimal amount of input required.

The only worksheet that really works a little different is the bass reflex model. In the bass reflex model you can move both the driver and the port to different, but not the same, locations along the long dimension of the enclosure. This will allow optimization of the port location which is typically not done in most bass reflex software models.

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