Missing Link - Box Tuning

[Zilch]

O.K., who posted the link to a page (Altec Tech Info, I think) describing placing a 1 kOhm resistor in series with the woofer to measure impedance as a method of box tuning, and where the heckie-darn is it?

I wanted to discuss this, but did I bookmark it?

NOOoooo.... :banghead:

[Robh3606]

I didn't see the post but you have BassBox Pro?? When you measure FS same set up if I remember right.

http://alteclansingunofficial.nlenet.net/publications/techletters/TL_226.pdf

This it??


http://alteclansingunofficial.nlenet.net/publications/techletters/TL_264A.pdf

This cause it's a gem.

Rob:)

[Zilch]

That's the method, but the page I'm looking for showed measured impedance curves as examples for optimal box tuning.

I believe the point it made was that "optimal" meant that both peaks were the same height, a new concept to me, the basis and implications of which I wanted to discuss further....

[moldyoldy]

Sounds like this, pp13-16

http://www.lansingheritage.org/html/altec/plans/1974-enc-manual.htm

[Zilch]

Thanks, Moldy. That's IT!

First, I don't see a difference between figs. 12A and 12B. 12A should not have the little second "hump" at 190 Hz, I presume? Otherwise, where is the leak shown there?

Figure 13 and associated discussion indicates that for a given driver and box, "optimum" tuning may be achieved by making the impedance peaks symmetrical about Fmin?

In the example cited, 25 Hz (Fmin) is too low, 42 Hz is too high, and 33 Hz is optimal.

What tuning does this produce?

I'm gonna have to pay more attention to the impedance curves generated in BB6P, looks like.

Any insight on the basis of this approach to tuning? AES reference(s)?

[moldyoldy]

I read the stuff, and even once in awhile understand bits and pieces, but am far from qualified to disscuss it.

One I've repeatedly try to 'grok' is;

http://www.dbkeele.com/PDF/Keele%20(1972-05%20AES%20Preprint)%20-%20Vented%20Loudspeaker%20A%20Restatement.pdf

with little success, though you may find it more enlightening than I did.;)

[Zilch]

***Zilch konks Moldy with a 79-page root locus plot...***


:p

[moldyoldy]

So that's what it is. Ain't the first time it's konked me.:D

[Zilch]

O.K., so I follow the Keele footnote 6 to the Ashley and Swan reference [8] for "an excellent way to measure box resonance frequency."

Ashley illustrates phase zero-crossing as an accurate means to measure the vented box frequencies Flo, Fhi, and Fmin, but does not translate them to box tuning.

Fine, that's the WT2 method, and Fsb is calculated:

Fsb = Flo * Fhi/Fmin

and termed "Vented box tuning frequency. Compare this to design."

4507 with 2235H, two ports closed measures:

Flo = 11.50 Hz
Fhi = 47.96 Hz
Fmin = 27.60 Hz

WT2 therefore reports Fsb = 19.98 Hz

Thiele definition of Fsb is "effective value of Fs when mounted in an enclosure,"

as distinct from Fb, "resonant frequency of vented enclosure."

There is NO way these boxes are tuned to 20 Hz as measured by WT2.

BB6P says Fb = 27.49 Hz, i.e., WT2's measured Fmin, essentially.

So, do I conclude WT2's Fsb is NOT the box tuning frequency? :dont-know

[I've had this problem before, unresolved.]

Note: CLIO gets basically the same impedance curve as WT2....

[4313B]

What is the Fc of the box with all ports closed tight.

BTW - Besides Loudspeaker Enclosures Their Design and Use (http://www.lansingheritage.org/html/altec/plans/1974-enc-manual.htm) I also posted JBL Enclosure Information Manual (http://www.lansingheritage.org/html/jbl/plans/1979-manual.htm) which has tuning methods as well.

[4313B]

What is the Fc of the box with all ports closed tight.

[4313B]

Ok.

[Zilch]

Fc = resonance frequency of closed-box system.

Fc = 40.12 Hz, top.

Also, vented system results, bottom.

[4313B]

sqrt(Fh^2 + Fl^2 - Fc^2) = 28.9 Hz so I would say both BB6P and WT2 are "correct" at ~ 27.5 Hz.

I checked WT2 awhile back against known volumes/tunings and I do believe Fmin in WT2 is Fb of the system. I can't remember WT2's actual variable labels offhand.

[Zilch]

Thank you, Giskard, that's what I was thinking, too: WT2 doesn't report Fb as such; it apparently uses Fmin as Fb to calculate Fsb and Ha.

The measurements are good; it's just a matter of definition and interpretation.

There's lots of information on how to tune boxes.

They presume I know how to measure box tuning once it's done. :p

[Well, there's what I know, now....]

Fb = sqrt(Fh^2 + Fl^2 - Fc^2)

[Zilch]

My peaks are unequal and asymmetric about Fmin.

According to the references, my tuning is "X BASS."

Which alignment it actually is, I have not yet ascertained.... :p

F3 is 37.62 Hz, and I don't run out of cone displacement at 150 W 'til 20 Hz.

[Sounds good, too....]

[4313B]

Which alignment it actually is, I have not yet ascertained....What is the measured Qts of the vented system?

It looks like an 18 dB/octave roll-off in your BB6P plot so it's a quasi-butterworth 3rd order which is typical of vented JBL's.

It looks like your sealed box/driver combo has a Qtc of 0.45

[4313B]

What is the measured Qts of the vented system?

[Zilch]

What is the measured Qts of the vented system?I'm confused, now. Qts is a "free-air" driver parameter, no?

Running Q/Fs on the driver in the vented enclosure, it reports data derived from the lower impedance peak at ~11.5 Hz.

For two runs:

Qts = 0.3959, 0.3962.

Is this what you're asking? Please explain....

[4313B]

Fine. Qt then. What driver is this? Is there a network in the mix too?

[Mr. Widget]

What driver is this?I'll bet it is a 2235 kit in the E130 frame... do I get the prize?


Widget

[Zilch]

I'll bet it is a 2235 kit in the E130 frame... do I get the prize?Nope.

Well, not YET anyway. :p

It's 2235H kit in 2225H frame, so, the "real deal."

There's no crossover, just driver in JBL 4507 box, two of four ports closed.

The 2235H/E140 I'm puttin' in a second 4507 for A/B comparison, once I establish the boxes measure the same....

[4313B]

Why? Do you have a bunch of E130/E140 baskets that have been reconed with 2235H kits?

[Mr. Widget]

Nope.

Well, not YET anyway. :p

It's 2235H kit in 2225H frame, so, the "real deal."Well at first I thought it was a 2235... I haven't modeled them in awhile and I remembered reading your posts about the E130/140 core issues so I leapt.

Why are you using those frames? Daniel ignorantly stuck me with a pair of them a couple of years back... I told him they were different and he didn't believe me until he measured the thickness of the magnet... they work pretty well, but certainly aren't as good as a true 2235... and now that the price of a 2235 kit has gone up a bit I certainly wouldn't waste one on an "E" frame. There are plenty of 2225 frames out there if you can't find an original 2235.

Widget

[hapy._.face]

BTW- Thanks for all the info guys! :applaud:

[Zilch]

I had a smoked E140 here. The forum was having the discussion regarding what motors and baskets were compatible with 2235H kits. I was curious as to what the difference would measure out to be. I also made some up with 2225H frames.

One's not good enough for listening, of course, so I went shopping on eBay for a mating E140 basket. I had to purchase 4 of them (from different sellers) over the past couple of months to get one that the magnet wasn't fractured. I had the cone kits.

So now, they're done, and being run in, today, actually. I have four 4507 boxes, too, so I'll be able to do reasonable comparisons and measurements in identical alignments, once I test them all.

The next fun thing is figuring out how to do reasonable (and easy) LF response measurements to characterize them. I want to try Small's "in-box" approach, but I'm having difficulty getting the CLIO distributor to sell me just the lite mic02 to use for that. :(

I'll get back on procuring one next week, now that the drivers are done and ready. Someone should clue Andre in that we're promoting his business here....

[WTPRO]

Hello to everyone.

You guys always seem to be coming up with good discussions! Ill try to add some of my own meager understandings and hopefully Brian will pitch in later with some of his knowledge regarding alignment tables (Loudspeaker Design Cookbook, Bullock On Boxes etc.) as this was the root cause for adding Fsb. My own personal way of looking at things is electro-mechanicaly with motors, back-emf, masses and springs... all rolled into a nifty simulator. Anyhow this is going to take some explaining.

Measuring and comparing to a simulated Box
------------------------------------------
Box simulators typically work with the user inputting a 'Box Tuning' frequency or a desired Helmholtz frequency. In this case the box is assumed to be a solid enclosure of air having a port acting like a mass. The air in the box then acts as a spring and will have a characteristic Helmholtz frequency Fh. The Helmholtz frequency is however NOT the final box frequency since there will be an interaction between the driver, box and port.

When drivers are added to a box, the box is no longer a sealed solid enclosure with a hole/port in it. Each driver is a mechanical element acting as a piston radiator. Morover, there is a moving mass with a suspension spring constant attaching the piston to the frame/box. The cone area is important since this is attached to one end of the box air spring. The other end is connected to the ports moving mass. That is, the actual box frequency is no longer Fh, but rather a new 'effective' box frequency.

Flo, Fhi and Fmid are points of interest that can then be extracted from an impedance curve and *compared* to a simulation or a family of alignment tables. Furthermore, Fsb is calculated and from this Ha and Alpha. I am saying 'points of interest' here rather deliberately since these require some interpretation with respect to a particular methodology.

Another methodology is to compare the impedance curves to a box simulator (EG the one in the data overlay window). In the end, if you adjust Fh (and maybe Vb in your simulator) the two impedance peaks should line up and when they do, the effective overall box tuning should be evident. Then you go back and adjust the port (or box) as needed. BTW, the peak amplitude can be off a bit due to Qloss but aligning the peaks in frequency is more important.

Keep in mind this assumes the starting point is a clean Z/P curve. Zilches curves look great, but keep in mind that if you ever come across a mechanical resonance (box on a flimsy chair) this can shift Flo, Fmid or Fhi throwing off the calculations.

Another interesting sticking point is that for some particularly massive drivers the effect of Le is hardly negligible and can tilt the phase curve enough to push the in box tuning calculations. That is, if your phase curve looks like Le has Fbox shifted up a bit, the measured zero phase crossing is probably no longer the mechanically equivalent zero phase crossing. I can fudge the data in the WT software, but this is not considered the standard operating procedure so I have declined to do this. Interestingly, a comparison to the simulator is a rather nifty solution.

Anyhow... You got me thinking again the Fsb thing again

[Zilch]

Thanks, WTPRO. The math using Fsb all works out correctly, it's just that I misinterpreted that as being the system (including the box) tuning frequency. Fmin is cited in the literature as a close approximation of Fb, except when large driver inductance is involved.

So, I'm now using Fmin as Fb. If I want more accuracy, I'll also measure the closed system Fc and calculate using that with Fhi and Flo as Giskard suggests here. That approach is shown in D'Appolito's Testing Loudspeakers. I don't know that contemplation of any change is called for. It's just not well covered in "Help" or FAQ yet, is all.

You see where I'm coming from. I haven't interactively linked WT2 to a box sim program. In this instance, I just want to measure the stand-alone system design result, i.e., hook it up, hit "Vented Box," and get the most info possible from just that. It seems that measuring Re in the process might be a worthwhile addition, since it's reported in the test results....

[WTPRO]

Hi Zilch

I was still hanging around tonight when I saw your post (seems I got another message too).

I had do to some head scratching too the first time I came accross Fsb. And, I agree that the help/faq could use some (ok a lot of here) help. More than likely some of my last post will show up in it.

Anyhow, if you open the Data Overlay Window you will find that the box simulator is set up so you can tweek up/down Vb and Fb until you get the impedance curves to line up. You can aither import data from a previous test or enter the TS data via a dialog box. The dialog box takes some getting used to however as the order in which you enter the data matters (note to self).

The part that weirds people out (including me) is that tuning a box is in the end interactive. That is, how can it be that what you put in for the Helmholtz frequency is in the end not what you get? What I would like/need to do is find a good way to make this intuitive.

I have also taken note about the fact that Re is not remeasured (goto Options->remeasure Revc). Actually one of my 'to do' things was to add a reverse annotate button for the T/S entry dialog, but Im not too satisified with that dialog just yet...

And you can bet Im looking at what Giskard posts too. That man is an asset!

Anyhow, its getting late here on the East coast.

Best regards
Keith

[spkrman57]

I have found the best results for me are derived by using a sealed cabinet with the WT2. When using ported boxes the results seem to be less accurate. :blink:

Then again, that could be because I am less knowledgable than most here!:p

Ron

[4313B]

When using ported boxes the results seem to be less accurate.
Really? How so? I guess I'd have to drag all that stuff out again and give it another going over.

I'd measure something easy like a B380 and post all the results that might be illuminating. I'd also include the internal box dimensions, bracing and ducted port dimensions so everyone had a thorough understanding of what we are dealing with.

1 run stock, woofer facing out
1 run stock, woofer facing in
1 run sealed, woofer facing in
1 run sealed, woofer facing out
1 run without any fiberglass, woofer facing in
1 run without any fiberglass, woofer facing out

[spkrman57]

Giskard,

It seems that I get "Q" factors (losses) that keep me from getting a good reading.

With the sealed box I use for most the values stay really close every time I test. I usually test one day and then a few days later and see what different results I come up with.

With the ported cabinet the values change more than the sealed for the same driver and setup.

Like I said, it is probably me as I am only a rookie in this stage of the game!;)

Ron