TL tapered design help/comments

[sunil]

Hello,

Firstly I would like to thank Martin J. King for this contribution to the DIY community & Bob Brines for helping with getting this design to a start. Mathcad is not easy for the non-initiated. I'm using a Vifa 8" driver, the P21WO-39-08 with the foll. published specs.

fs=26, Qts=0.38, Vas=85 lit.

Bob's suggestion of a tapered TL with an S0=3Sd & SL=0.25 Sd works very well. The width of the TL is 8". The f3 is around 30hz ( I hope I have read the graphs correctly ). The cabinet is not too big & scores points with the WAF. The design is a single-fold tapered TL with the wider part at the bottom & the fold at the top. This follows thru' down to the terminus. There is no port per-se. It is more like a slot (Bob's peerless pipes). Since the end is 0.25 Sd I have taken that area & used the equated radius of 3.4" as the "rport" ( Thanks Bob ). The lport is left at 0.0001". My problem is that testing this with the ported-box (PB) worksheets don't match. The PB worksheets show an f3 closer to 50hz . What has gone wrong here ?

Is it necessary to have a port with a proper radius & length ?
Most importantly should the port be raised above the floor ?

Any help/comments are welcome. I really would like to get this sub-thing going.

Cheers,
sunil

[sunil]

Here's a skeletal image of the thing.

[BillFitzmaurice]

I see a problem in that you have the driver location relatively close to the center of the line. An offset driver should be located from 1/5 to 1/3 of the way down the line. To accomplish this with a single fold and maintain the driver height above the floor you can flip the box upside down.

Your anticipated F3 of 30 Hz is very optimistic. The F3/Fp ratio will be about .8. A 66 inch line has an Fp of about 51 Hz, so you can expect an F3 of 63 Hz. With the .38 Qts The F3 to Fs ratio is about 1.8, so even with a long enough line (90 inches) the best you can expect with this driver is about 47 Hz. With room lift it will probably work fine to 35Hz or so, but I wouldn't call that true subwoofer peformance.

Your stumbling point here is the driver Qts. With Qts of .6 the F3 /Fs ratio goes to 1.2, so a 26 Hz Fs would be good for a 31 Hz F3 from a 90 inch line and with room lift it would run to 20 Hz useable no problem.

[GM]

Originally posted by sunil


Bob's suggestion of a tapered TL with an S0=3Sd & SL=0.25 Sd works very well. The width of the TL is 8". The f3 is around 30hz ( I hope I have read the graphs correctly ).

I get ~28Hz using published specs/your design, assuming 'Epsilon = 0.33 Sd' actually means the driver is at 0.33*L. This is close enough to my calc'd 0.3247, so should perform well if the published specs are reasonably close.

Since the end is 0.25 Sd I have taken that area & used the equated radius of 3.4" as the "rport" ( Thanks Bob ). The lport is left at 0.0001".

Hmm, I calc the radius as ~1.703", which yields the ~28Hz F3......

My problem is that testing this with the ported-box (PB) worksheets don't match. The PB worksheets show an f3 closer to 50hz . What has gone wrong here ?

For whatever reason, when the 'Lport' value in the MLTQWT worksheet is some very small length, you have to change the density value at D2-D5 to what is listed up in the 'density' field for it to match the PORTED plot, and why I do not use the latter for simming TLs.

Most importantly should the port be raised above the floor ?

Only if it fires out the bottom, which yours does not.

Originally posted by BillFitzmaurice
Your anticipated F3 of 30 Hz is very optimistic. The F3/Fp ratio will be about .8. A 66 inch line has an Fp of about 51 Hz, so you can expect an F3 of 63 Hz. With the .38 Qts The F3 to Fs ratio is about 1.8, so even with a long enough line (90 inches) the best you can expect with this driver is about 47 Hz. With room lift it will probably work fine to 35Hz or so......

Being a reverse tapered pipe, it has a mass loaded vent of sorts, so if the published specs are reasonably close it should have no trouble reaching its T/S half space prediction of ~28Hz F3. Of course if the room/speaker location does not give a true half space loading then some EQ will be required to flatten it out in-room.

That said, a shorter, slightly smoother response straight (or folded if height is an issue) ML-TL would be my preference and use stuffing at the vent to lower the rolloff as required to blend to any room gain:

L = 44.5"

zdriver = 16.75"

zport = 41"

SO/SL = 90.121"^2

density = 0.3lbs/ft^3

rport = 1.5"

Lport = 6"

GM

[sunil]

Hi GM,

Thanks for that. Yup, you're right about the radius being 1.703"., but I don't get how you got the f3 of 28hz.

I used the specs you've recommended & yes, it has a nice curve. One thing is I did'nt get is the S0/SL part where you've mentioned 90.121"^2, so I went ahead & tried a model of 3sd:0.5sd ie....109.26:9 "^2 & it has a nice curve. I tried it with a line length of 66" which is OK by me, sizewise. Do you reckon these figures are good ?

The mistake I was making was with the port size/dimensions. Do I need to worry about phase errors between the port & the driver ?

Once again, thankyou for the help.

cheers,
sunil

[GM]

Originally posted by sunil
Hi GM,

Thanks for that. Yup, you're right about the radius being 1.703"., but I don't get how you got the f3 of 28hz.

Greets!

?? I just plugged your numbers and the correct radius into the MLTQWT worksheet.

I used the specs you've recommended & yes, it has a nice curve. One thing is I did'nt get is the S0/SL part where you've mentioned 90.121"^2,

?? How could you plug them in if you did not understand the input values? Anyway, SO and SL is the top and bottom cross sectional area (CSA) and I use square inches so that folks can easily figure out width/depth dimensions. When basing it on Sd you have to divide it into the CSA to get the multiplier, or 2.4741 for both SO and SL, i.e. no taper.

so I went ahead & tried a model of 3sd:0.5sd ie....109.26:9 "^2 & it has a nice curve. I tried it with a line length of 66" which is OK by me, sizewise. Do you reckon these figures are good ?

Well, 0.5Sd is not 9"^2, but using 0.5Sd it is ~max flat, but the straight ML-TL pipe I proposed is smaller, smoother, and is flat to a little lower Fb, but how audible the difference is I cannot say. Anyway, since we do not have measured specs, who knows which is actually superior.

The mistake I was making was with the port size/dimensions. Do I need to worry about phase errors between the port & the driver.

Once again, thankyou for the help.

cheers,
sunil

You are welcome! Not at this low a tuning frequency, though some stuffing my be required at the vent to roll the bottom off if there is too much room gain.

GM

[sunil]

GM,

I did'nt undersatnd it was a straight TL. The discussion bordered so-much on the tapered bit I assumed the figure was either related to S0 or SL. Anyway, yes, the straight TL has a nicer taper.

Do you reckon maximally flat might overload the room ? I ask, because, opinion on the forum varies quite a bit & I understand the TL is a different beast when it comes to bass. Since the roll-off is much gentler, with room-gain, it adds a lot to the output.

cheers,
sunil

[GM]

Greets!

Depends on the room/app, but I prefer to design for max acoustic gain and 'dial it back' as required rather than not have enough and have to overdrive the speaker with EQ boost or lower the rising response to flatten it.

Obviously, it is best to measure the room response and design a speaker response that is ~the inverse of it to achieve a ~flat in-room response, but few DIYers have this capability.

WRT ML-TLs or lightly damped TLs, as the sims indicate, there is not much difference to a reflex design WRT the roll off slope/impulse response.

Stuffed TLs that mimic an infinite baffle's performance in a semi-manageable size are a 'different beast' and require considerable room gain to match the others WRT LF output unless drivers with a ~1.0 Qts is used with a much larger CSA, so usually require corner loading in a relatively small, well sealed room for best performance. These work best with 0.5-up Qts drivers, otherwise some series resistance or digital EQ is required to flatten their response.

GM

[sunil]

Hi GM,

Phew, looks like you're the only one willing to comment.

From what you've described, I understand this design is still in BR territory, not a true TL. Is it any different from a BR in terms of unloading at really low frequencies. I have no experience & thats why I ask.

I put in the numbers you suggested & they make a slightly large cabinet. So I tapered & folded it a bit. Problem now is the vent, which is 6" long gets pretty close to the back plane of the cabinet. Is this alright ? Does there need to be space behind the vent ? If so how much is good ?

I modeled a basic cabinet with a fold & have managed to put the port next to the fold, so I get more space behind. Please take a look at the attachement & comment. The figures used for the design are the same as you suggested. The graphs look really nice.

Thanks,
sunil

[sunil]

Attachment: