New ML TQWT project

[tda]

Hello guys,
I've just finished my first ML TQWT MathCAD simulations.
I would like to know your more advised opinion about the results.
Any suggestion will be great !
Thanx

[sreten]

Your area looks very large for for the size of driver.

Recommend you use two drivers and roll one off for
baffle step compensation.

Recommend a 5cm and 10cm port.

(I haven't actually checked your results)


sreten

[Timn8ter]

At first glance I think that 100 cm is very short for a driver with 49 Hz Fs. You should consider going 150 - 175 cm. I'm not familiar with this driver being in the US but I'm assuming it's a mid-woofer with a top end around 5 Khz, is that correct? Are you using another driver with it?

[Hefty]

Where's the grafs?
Your placement of the driver is a little odd? Usually one strives to place the driver at the middle.
A TQWT at 100cm? what's the use of doing a TL if you're not going to use it to extrend the bass responce?
I would very much like to se the frequency responce...
I'm building a ML-TQWT myself right now so I'm always interested in more people doing similar work.

[Timn8ter]

Your placement of the driver is a little odd?

Others including myself have used distance from the top equal to 43% of the length of the line.

[GM]

It sims fine, though the stuffing has to be what I consider somewhat excessive to achieve the desired results. This is an unfortunate side effect of using too long a pipe and too much Vb to get down to a lower Fb than the driver is comfortable with based on its specs. FWIW, the way I calculate pipes, this length = 31.55Hz and for a Xo = 0.35, I get a So/SL = 254.86cm^2, which yields a flatter response with less stuffing.

One thing though, if played loud enough for > ~18W to be required it will go non-linear in the octave above Fb, so if using this driver to its limits, a smaller/higher Fb will be needed for best performance.

Since I don't normally push a driver this much in a pipe, I don't know how accurate it will sound WRT the sim, and of course the driver's actual specs may vary enough to require a completely different alignment, rendering this all moot.

====

>At first glance I think that 100 cm is very short for a driver with 49 Hz Fs. You should consider going 150 - 175 cm.

====

Actually, it's much too long for an optimum ML-TL. You make them longer to tune them lower, but at 0.707*49 = 34.63Hz, a point of diminishing returns is reached. At 100cm he's skirting right at the limit of ~flat response. At 175cm, Vb must increase 5.34x over that required for 100cm, dropping Fb to 30hz, well below what this little driver can handle without a huge horn.

====

>Your placement of the driver is a little odd? Usually one strives to place the driver at the middle.


====

In a ML-TL, the objective is to put it where it suppresses the 3rd harmonic, which is a complex function of column length Vs width, not a linear dimension, so for the theoretically best performance it will vary with each design, and tapering further complicates it. Fortunately, MJK's worksheet allows you to move the driver around till the right spot is found if you don't have a mindset that limits you to one point along its length regardless of the alignment.

GM

[Hefty]

I must admit that I'm a little baffled by the design?
It pretty much goes against all that I've learnt about a ML-TQWT.
It would be nice to see the graphs though so that I know what you're trying to achieve. Your goals may very well differ quite a bit from mine...
I'm a n0ob at this as well...

[Timn8ter]

Boy, I'd better hit the books some more. Is there a discussion of how these graphs compare (or what they really should look like)? I'll review Martin's comments but a breakdown of what those graphs mean and what they should look like would be a big help.

[tda]

Ok guys,
I didn't expect to have so soon even one reply; so first I will put my response frequency graph.
To achieve this results I used Martin alignment tables, with S0/Sl=1.
The response frequency looks pretty good to me, but I’m not 100 percent confident that the results will be the same as in theory.
I’ve reduced the stuffing density to 0.5 and the results are almost exactly the same, I don’t think the stuffing will be an issue.
Anyway, what is the best way to use Martin King MathCAD sheets? Personally I used alignment tables provided by him. Are there any other methods?
Thanks to everybody for every reply.

[GM]

>I must admit that I'm a little baffled by the design?

====

In what way?

====

>It pretty much goes against all that I've learnt about a ML-TQWT.

====

OK, I'll bite, what have you learned that's so contrary to what I've said?

====

>It would be nice to see the graphs though so that I know what you're trying to achieve.

====

I don't have a website to post from and don't feel like messing with Yahoo/whatever, but if you have Mathcad 8 or newer I can send you the .mcd worksheets to view, or you can load the data into the freebie demo from MJK's site to view them, though you can't save them.

====

>Your goals may very well differ quite a bit from mine...

====

In what way? I'm into high performance designs, but I understand the limitations of the various systems due to physics of the situation.

====

>I'm a n0ob at this as well...

====

I assumed as much.

====
>Boy, I'd better hit the books some more. Is there a discussion of how these graphs compare (or what they really should look like)? I'll review Martin's comments but a breakdown of what those graphs mean and what they should look like would be a big help.

====

When you find a book that discusses this stuff in depth, let us know? Maybe some aeronautical engr. textbooks? I learned alot from perusing mechanical/structural eng. textbooks in that they pointed me in the right directions WRT building numerous prototypes looking for trends back in the '60s/early '70s. I could have saved myself a lot of time/wood/$$ had my math skills been better and I had known someone with good math skills that was willing to help me.

====

>I'll review Martin's comments but a breakdown of what those graphs mean and what they should look like would be a big help.

====

Hmm, WRT to pipe design, some of his graphs appear to be just his way of 'thinking out loud' while he was developing his theories, IOW, showing the intermediate steps that comprise the bottom line, i.e. the FR plot.

I mostly look at the 'Far Field Transmission Line System and Infinite Baffle Sound Pressure Level Responses' and the 'Woofer and Terminus Far Field Sound Pressure Level Responses'. The former to view if I've hit the target alignment, and the latter to ensure there's little/no vent harmonics affecting the driver's phase response.

I use the 'System Time Response for an Impulse Input' plot to tell me if there's adequate stuffing.

====

>To achieve this results I used Martin alignment tables, with S0/Sl=1.

====

And then just added the vent? Interesting!

====

>The response frequency looks pretty good to me, but I’m not 100 percent confident that the results will be the same as in theory.

====

His worksheets are extremely close to reality when accurate driver specs are inputted and measured nearfield to keep the room interaction to a minimum.

====

>I’ve reduced the stuffing density to 0.5 and the results are almost exactly the same, I don’t think the stuffing will be an issue.

====

Actual stuffing density is so room/personal preference dependent that the effects of the stuffing is strictly a baseline.

====

>Anyway, what is the best way to use Martin King MathCAD sheets? Personally I used alignment tables provided by him.

====

I made a spreadsheet based on his original release to see how it compared to mine and wasn't surprised to see a considerable difference, but I didn't try them with a vent. If they all work as well as this one did, then for a ML-TL these would probably be 'close enough', at least as a first approximation.

BTW, did you sim it with the reduced So/SL I recommended?

====

>Are there any other methods?

====

Well, mine is currently proprietary, though I've dropped enough info/hints on the FR forum for someone to probably figure it out, but who knows what the future may bring.

There's Augsburger way of course, and Rick Shultz's Exolinear format. Whether one is audibly superior to the others, I can't say since I've only built my versions.

====

>Thanks to everybody for every reply.

====

You're welcome!

GM