Isn't there another (hidden) bonus for achieving lower box losses (QL) ?

[RMC]

I think there is but can't prove it, so I hope other members will help me on that part.

This question has been bugging me for a log time and should do the same to other speaker builders who work hard to prevent air leaks. Is it really worth it to go the extra mile, for other than achieving the standard QL 7 assumption on box losses ? ( I leave aside for the time being Qa absorption and Qp vent losses).

When building a vented box I try to do better than standard construction: good plywood, glued & screwed, heavy bracing, not using cheap thin cardboard gaskets (sometimes supplied by driver manufacturer), but rather good quality rubber or foam ones on woofer and tweeter; siliconing ALL inside joints, including around the connectors and vent ducted tube junction with front panel, etc. As for the losses caused by the woofer (lossy surround or dust cap) there isn't much we can do about that, otherwise it may change woofer parameters by adding mass or modifying suspension compliance.

Quite often in speaker building literature the concept of QL is dicussed but not really the impact of it (good or bad), nor the possible benefit of reaching a higher number (QL 3 being a pretty lossy box and QL 15 being about as good as it usually gets). As standard practice most people (like me) design with QL 7, and they forget the rest. Not me, as I go further In trying to build meticulously to reach a higher QL number. The obvious standard reward is if you designed for flat response using QL 7 and worked "normally" on the box you'll get that flat response. If you did same but "sloppy" box work ending with a QL 3-5 then your low-frequency response will show an unforeseen dip. However, if you did same, but worked harder on the box and were successful at reaching QL 10, 12 or 15 then you're standard reward is an unforeseen low-end peak in response when compared to QL 7. A dip or peak is a curious form of reward one might say... You can try to model this in Win ISD Pro to see for yourself.

My encounter with this demonstration of dip/peak in response caused by various QL numbers came a few years ago on North Reading Engineeering's (NRE) web site (northreadingeng.com) with whom I have no association, but it is very informative, well-made and eye-opening. Though again it falls short of explaining what it means/implies or can do with it in practice for the speaker builder. NRE specifically states "The magnitude of QL must be determined after the enclosure is constructed by analyzing the motional impedance of the driver mounted in the enclosure." That's the bugger: to only know afterwards...

I'm venturing into new territory now. I'm pretty sure there has to be some other hidden bonus for my good box work, as far as the peak is concerned (not for the dip).

Those who enjoy a bump in low-frequencies would be served by higher QL. For others, how do you get rid of that high QL related peak ? Certainly not by making the box lossy which would defeat the purpose of hard work. EQ could be a possible option whose merit is to leave more headroom to amp/driver. Also, logically I think by increasing box volume and/or by tuning vent lower (Fb) one would flatten the peak leading to a lower F3. In other words, it looks like an air-tight box would allow deeper bass (if larger Vb and/or smaller Fb ?) , or maybe simply the same (F3) but from a smaller box ? Could this be the hidden bonus ? These possibilities are never mentioned in what I've seen about QL. Let's not forget that driver Xmax is not increased to deal with these lower frequencies, though the vent would help here.

I haven't had time to try to validate a bunch of scenarios in Win ISD Pro based on the above.

Your input is requested. Thanks.

Richard

[Ruediger]

To my knowledge QL is an indicator of box size (small, medium, large) and/or of build-(im)perfection, used to pick the appropriate nomogram out of several ones.

If you aren't designing by nomograms or if you have just one and not several nomograms then QL can be used as a check for the quality of your craftsmanship.

A real QL different from infinity comes from parasitic elements in the R/L/C circuits by which you describe a loudspeaker system (electromechanical analogies).


The benefit of a very high QL is that your electromechanical analogy actually describes your loudspeaker system and not one which you don't have. The blueprint matches the product.


<excerpt>
Excerpt from the chapter "Vented-Box Loudspeaker Design with a Given Driver" of the book "Introduction to Electroacoustics and Audio Amplifier Design" by W. Marshall Leach, Jr.:

The vented-box design procedure for a given driver requires a knowledge of the driver parameters fS, QTS, and VAS. A value for QL must be assumed. The rule of thumb that is commonly used is to assume the value QL = 7 for the initial design. After the required box volume is determined, it can be assumed that the value for QL is correct if the box volume is in the medium range, i.e. 2 to 3 cubic feet. For smaller volumes, a larger value of QL should be assumed and the design procedure repeated. For larger volumes, a smaller value of QL should be assumed and the design procedure repeated.
Vented-box design charts for QL = ∞, 20, 10, 7, 5, and 3 are given in Figs. 1 through 6.
The procedure for using these charts is as follows:
<end of excerpt>

The above excerpt is available on the net, search for the title.

Ruediger

[RMC]

Thanks Ruediger for your insight. Its a different approach to QL. Taking it from another angle is very clever. Plus Leach isn't the first "knows all Joe" met on the street corner...

My understanding now is that the assumed QL 7 figure is good initially (as a starting point). Once initial Vb is known after a first try, it is assumed QL 7 is ok for mid-size boxes, but must be higher for smaller boxes and another design try must be made with this higher QL, same two-step procedure for larger boxes but with a lower QL figure. This two-step process is a first time see for me. Sadly, Leach didn't suggest wich QL number to use for small or large boxes.

Doing it this way (in two steps with revised QL at second step) finally makes sense of all this and naturally gets rid of the low-frequency dip or bump since the second step of design considers a different QL number.

However, correct me if I'm wrong, from the above, Leach implies that larger boxes may have more losses (regardless of work done on it), and smaller boxes may have less.
By the way I have found the Leach excerpt (6 pages) on the Net, thanks again for that, plus I'm also going back to Small's Vented-box loudspeaker systems, part 1, where enclosure losses are also discussed.

You made my day ! Best Regards,

Richard

[grumpy]

IIRC, a person here, much more knowledgeable than I, suggested that JBL achieved a QL of ~10 with the better constructed cabinets... better than nominal industry std.

[RMC]

Hi Grumpy,
Thanks for that QL 10 number, this info is very valuable to me, since I finally get a ballpark figure of what a reputable manufacturer has reached with better construction. With your help I'll stop chasing ghosts and use QL 10 as my target from now on... By the way, on another Web site I saw last night someone mentioned something similar: that famous R. H. Small indicated in his papers that QL 7 and 10 are attainable, therefore implying that above this number its more elusive ... Thanks again.

This leads to the following correction to my article:

The phrase where I mentioned "... QL15 being about as good as it usually gets." should read instead: "... QL 15 being about as good as it theoretically may get."

Best regards,

Richard

[RMC]

I'm glad to report that there is in fact another bonus, as I had thought, for achieving lower box losses (high QL number). While searching for something else, I did find some proof of my theory of a possible smaller box volume required when losses are minimized with good box work.

RE SMALL ALIGNMENTS: here are the explanations followed by a measured example.

"Small observed that vented loudspeaker systems designed according to Thiele alignments did not always exhibit the frequency response predicted by the model. He referred to these ... as losses because they usually resulted in decreased output at certain frequencies, and accounted for them by ... leakage loss."

"... QL; the larger this Q number, the less effect the box has on the response." "... for initial design purposes we must assume a QL value. According to Small, most systems have a QL of between 5 and 10, with a general tendency for it to fall with increasing box volume. Thus, the assumptions of QL = 10 for small boxes, QL = 7 for moderate boxes, and QL = 5 for large boxes should be satisfactory. " (Robert M. Bullock, Bullock on boxes, 1991, pages 6 and 7).

In his "Design Box"on page 7, Bullock goes on with some examples, of which # 5 is the interesting one for our purpose. Using a driver's parameters, and a QL =7 assumption, he gets a Vb of 9,615 in^3 (5.56 cu. ft or 157.58 liters), an Fb of 28.6 Hz and F3 of 24.9 Hz. Then he states "This box is big enough that QL = 5 should probably be used. This gives Fb = 28.8 Hz, Vb = 11,343 in^3 (6.56 cu. ft. or 185.90 liters) and F3 = 25.6 Hz."

For pretty much the same Fb and F3, a QL = 5 box required 1 cu. ft. (28.32 L.) more volume compared to a QL = 7 box. That additionnal volume is needed to compensate for box air losses in order to achieve the desired results (Fb, F3). Or seen the other way around, the QL =7 less lossy box needed 1 cu. ft. (28.32 L.) or 18% less volume in this case to achieve comparable results of Fb and F3 !

"Rigid, well braced enclosure walls, minimal leakage at speaker mounting, port mounting and terminal installation will provide higher Ql values." (North Reading Engineering.com, re simulations with JBL 2226H).

I would add: using caulking to seal all the joints inside the box, even around the ducted port/front panel junction and terminal inputs, plus using a good quality gasket (rubber or foam) to seal the driver(s) to the box will definitely increase your chances of being rewarded by a smaller required enclosure Vb with low-end performance equivalent to that of a somewhat larger box.

That's an interesting incentive for going the extra mile to ensure cabinet leakage losses are minimized.

Richard

[RMC]

In fact, the less lossy box of QL = 7 (vs the QL = 5 box) in Bullock's example, does show a slightly lower F3 since the driver/enclosure combination makes better use of all the air in the cabinet, considering less of it is wasted in losses...

I bet that with a better QL figure, like QL= 10, this F3 number would be even a little lower!

Therefore, amateurs of "How low can we go" in terms of deep bass, have at least the incentive of improving their F3 results a little if they declare War to enclosure air leaks...

Richard

[Ian Mackenzie]

QL is lumped losses in the system

Driver, box , port, fill

Larger boxes tend to have have a lower QL ie 5 or 3

Programs like Leap enclosure shop simulate changes in QL

In reality measuring it like Bullock does is not easy

The Loudspeaker Cookbook covers QL in detail

A stiff box with flared ports is something to aim for but not loose sleep over.

[Ducatista47]

A stiff box with flared ports is something to aim for but not loose sleep over.

I'll back Ian up on this one. He suggested that a home build of mine (a speaker he had already constructed himself) include a doubled thickness baffle board. Best advice I ever got about building, along with Macaroonie's system for building with router joining.

[Ian Mackenzie]

I do have Bullocks book somewhere

To make sense of this you really need to
Test and measure the actual T/L parameters for your sample drivers

With LMS you then test not just small but large signal parameters

What the drivers does under warm large signals is not the same as small signal conditions

Jbl of cours take this into account when they design a box

In terms of low frequency extension the ear is quite sensitive to level changes in
The bass

But of course your room modes below 350 Hz with damp any of that like a 10 Db peak right where you sit at 40 hz

[1audiohack]

...With LMS you then test not just small but large signal parameters

What the drivers does under warm large signals is not the same as small signal conditions...

Agreed. LMS and a VI box were the tools needed to figure out several things that I could not put my finger on. It's a different world with a constant voltage, low impedance source at my typical listening levels.

Barry.

[RMC]

Hi Ian and Barry. Thanks for your input.

RE YOUR 1 rst POST:

I know QL is the sum of three things: leakage(Ql), absorption (Qa) and port losses (Qp). In my post #1 I mentioned in the first paragraph that I was leaving aside the last two items for the time being. Why? Because Bullock writes (page 22 footnote) "...the latter two are usually small enough to be ignored." And Dickason in his Loudspeaker Design Cookbook, 5 th. ed., 1995, on page 49 "In practice, Qa and Qp tend to be so low that they are not very significant." Naturally, assuming a normal enclosure as he writes. Box leakage is THE most significant item to work on when building a box (Bullock, p. 20).

I also quoted Bullock that larger boxes TEND to have a lower QL number. But nothing stops speaker builders from trying harder to achieve a QL figure of 7 or maybe 10, considering Small says most systems have a QL between 5 and 10, plus a member mentioned in this thread JBL achieved about QL 10 in its best boxes... Bullock has interesting tips to get rid of some leaks.

You can also simulate the effect of QL number changes on frequency response (if this is what you meant) in Win ISD Pro Alpha (the poor man's version of Pro software), but it is certainly not as extensive and professionnal as the LEAP program.

"In reality measuring it like Bullock does is not easy".

Agreed 200%. Bullock has the advantage of being a Mathematician. Scientists like him and Engineers, live in a world of equations, formulas and high flying mathematical/electrical models, where everything can be explained, represented or summarized by a model with quite curious figures... "Ordinary people" like us have the choice to try to follow by hanging in there, or just switch to gardening, painting or mechanics to occupy spare time. I try to hang in there...

Regarding the desirable stiff box, agreed, that's why I quoted North Reading Engineering. They did useful simulations on box QL a few years ago. Their Web site has changed, difficult to find the goodies now.

RE YOUR 2nd POST:

"To make sense of this you really need to Test and measure the actual T/L parameters for your sample drivers"

Bullock also appears to have low faith in spec sheet parameter values when he writes "In my opinion the spec sheet parameter values are not accurate enough to use for design. For the best design, the driver parameters of the actual units you will use should be determined." (p. 10) And "The vagaries of system parameter values make a need for corrections almost certain." (p. 19). Agreed to, for the best design.

However, there seems to be another point of view on this issue in Dickason's Cookbook (p. 18-19) where he writes about variances in Qts, Fs and Vas: "For a given production run of identical drivers..., values of Qts, Fs and Vas can exhibit considerable variation. Consequently, if the measured values for a small sample (usually two for most of us) are quite different, don't be overly alarmed. Although individual driver parameters can have a wide range of values, the ratios of Fs/Qts and the product of VasFs^2 are fairly constant. Thus, the final and essential results will also tend to be constant."

I don't disagree with small/large signal parameter differences, although I don't have the chance of owning Leap/Lms.
Finally, with regards to
"In terms of low frequency extension the ear is quite sensitive to level changes in The bass", past research has shown the human hear can distinguish a level change of 1 or more db, if my memory is correct. Regards,

Richard
EDIT: Looks like the software pulled a trick on me once again. Sorry for the changing size of the letters...

[Ian Mackenzie]

Yes it's an interesting academic discussion

Like most technical topics, the more you know the more you need to know

I must read Bullock again

It also depends if you are biamp ing your woofer or using s passive network ?

The net outcome is if you have high losses say Ql of 3, your going to have a non flat negative ripple and conversely a non flat positive ripple in amplitude at box resonance Fb

This is dependent on knowing the QL of the box and know the actual sets VAS etc of your actual driver?

I think Bullock is saying a correctly damped bass reflex design is the design aim

Why ? Because how will you know the baseline box volume without knowing the T/L parameters?

Only then can you assess QL

However as Vance points out the actual Qts and not box size errors or variations are more important

Any cable dcr or significant low output impedance of the amp is going to increase the Qts

Does your head ache...Lol

Now, add to this non maximally flat alignment like GTs banana curve

The idea of the banana is an over damped alignment that in a real room resembles a maximally flat response.

Try a maximally flat alignment in a real room unless the box is well out from room boundaries it will be audibly impaired

I have a suggestion :

If you reall are craving for extended response look into DB Keeles 6th order alignments

You can typically get a 1/2 octave extension in F3 and improve displacement output

Check out DB Keeles web site. The Electrovoice guys did this with their domestic system

Buy Woofer Tester and research this and report back how you go

[RMC]

Hi Ian,
With all due respect, and as a closing note on my part, I don't think its an academic discussion, as you mention. There's a real practical gain in achieving high QL: smaller box for equivalent F3 or lower F3 from the larger box. It would be a purely academic issue if higher QL numbers were out of reach or unattainable. But up to now, I have seen no "evidence" that a reasonably high QL (e.g. 10) would be a dream.

BTW, in post # 6 here, I forgot to cover that, in Dickason's Cookbook mentioned earlier (p. 51) he shows the effects of enclosure loss on frequency response (fig. 2.11) where it can be seen that compared to a "standard" QL 7 box, a QL 10 has a slight peak or positive ripple and a little lower F3 (as I anticipated in post # 7, second paragraph, "I bet..."), and a QL 3 has a dip or negative ripple along with a higher F3. Naturally, comparing the response of a QL 10 with a QL 3 will show significant differences from the same box.

Just under, (fig. 2.12) where Dickason shows the relation between QL value and box size, using QL 7 as a typical loss figure or reference, it can be seen where QL value is to the left of 7 (i.e. lower) it says "more than expected loss figure" "increase Vb"; on the other hand, where QL value is to the right of 7 (i.e. higher) it says "less than expected loss figure" "decrease Vb". This last item, like Bullock quoted in post # 6, confirms my initial theory in post #1, and also the smaller required box volume for comparable F3 with higher QL, in post #6.

"The enclosure's single most important requirement is that it contains no air passages except for the vent. I cannot emphasize this point enough; the slightest cabinet leak at a point other than the vent can significantly alter the response." (Bullock, p. 20). That's pretty strong language from the man.

Again, for those addicted to very low bass, load the caulking gun and get going on air leaks. Then, enjoy the ride. As for me, I do that routinely on all my boxes made, even if I'm not chasing very low bass results (± 40 hz OK) re lack of space for large boxes, the wife! and no real interest at this time in a 6th order alignment.

I know this doesn't settle the issue of the difficulty of measuring the actual QL reached... However, even if it can't be measured, one cannot loose by doing his very best on any possible air leaks, its a sure win with a likely lower F3. If a small + ripple (success) = pimples for you, don't re-do the box, try tuning it a bit lower...

I rest my case, as they say.

Richard

[Ian Mackenzie]

Hi Richard

It is a very interesting topic and no intent to offend on my part and l will refer to the exact text so it is in context of the relevant aspect within the chapter Vented Low Frequency Systems


As quoted in the Vance Dickason Cookbook p 67 2.61, 7th edition QL losses

Referring to the wider text the QL losses are provided in the underlying formulae and tables to assist in calculating with a degree of accurate the correct box size and tuning Fb. Quote "these losses are altogether unpredictable and must be measured with a working model." Un quote

Quote "to correct for losses begin by assuming the typical loss figure QL = 7, construct the appropriate sized enclosure, tune it to the indicated frequency, then measure the new enclosure for exact losses. If the measured loss is close to the target WL=7, no additional changes are necessary. If, however, the measured loss is less or more than QL=7 , you must recalculate and alter the enclosure size and tuning. Un quote

The aim of Dickasons text is to obtain an accurate alignment in practice, not aim for a lower F3 by virtue of a high QL = 10.

I understand what you are saying but a lower F3 in itself is not the intent of the Dickason text you are referring to. An accurate tracking of a specific alignment is the objective

In reading the article if you follow the process the box size is varied up or down to obtain the correct alignment.

It comes back to doing some real measurement.

If the aim or objective is a lower F3 there are other assisted alignments as l mentioned previously .

My concern with these types of discussions on an open forum is that to the un initiated a simplification or mis interpretation is mis understanding of someone else's original work can lead to mis information.

That leads to a repetition of mis information once posted and then you have everyone confused

As mentioned in another thread l hope to soon publish a series of documents on a website that will cover some topics such as this for the benefit of obtaining the pinnacle of performance from the 43xx and the 44xx systems

Enclosure tuning will be an area that l will research and illustrate with real examples using highly accurate models from LMS in LEAP Enclosure shop. This cover theory, real world applications and the challenges and compromises facing the JBL user.

You will be therefore be better informed and equipped to construct, tune and set up your system so you can obtain the intended performance.