Some issues re step-down mode
E1) CONE EXCURSION
E-V, JBL (in post # 26 data sheets ) and Kramer/Timbers' article did not provide in their documents cone excursion graphs/curves for the modelings they did. But we can refer to some excursion data provided in some posts here. One must remember there are TWO bumps on a cone travel graph with these B6 or so alignments, one on each side of tuning frequency, whereas the usual vented-box design has one bump above Fb and a continuing excursion increase level below Fb (e.g. post # 11).
Its interesting to note for these high level curves there is LESS cone travel reduction at Fb with assisted alignment VS conventional unassisted alignment. This is considered to be the result of applying a 6 db boost at Fb or so (subject to further explanation below), which is equivalent to supplying four times the power at that point. However, this lesser cone travel reduction by the vent at that point is somewhat compensated elsewhere with MORE cone travel reduction, below the assisted tuning frequency VS unassisted design. That reduction is the result of the subpassband 12db/oct. high-pass filter included with the 6 db filter boost (i.e. boost/cut filter).
The excursion bump above the new Fb (in the region of original tuning) may well reach driver excursion limit first because the driver's maximum acoustical output capability is usually reduced by a few db in the region of original box tuning, as a result of lower than normal box tuning. Consequently, more cone excursion there. On the other hand, the excursion bump below the new Fb, a result of bass EQ applied, has a slower rise than with usual vented box, thanks to the high-pass section of the bass EQ filter.
The older 15" 2235H (Xmax 8.4 mm) in post # 11 illustrates at the 150W power rating with assisted alignment the driver isn't out of bounds, excursion being a bit less than 8 mm on the right cone travel bump, and a bit more than 7 mm on the left excursion bump. Not a lot of loose left, but it doesn't have its back to the wall yet. That slower rise on left bump permits here to add, with an equalizer or DSP, maybe about 1-1.5 db or so of bass equalization around the boosting filter's frequency (Faux), then the driver is about done with Xmax on both bumps at the same time...
Vance Dickason also produced a computer simulation, in his Loudspeaker Design Cookbook 5th edition, of a Class I sixth-order alignment similar to others in this thread, "... to show the dynamic consequences..." of such. He used a 12" woofer (Qts 0.30) in a QB3 aligned vented-box, that he tuned at 24.8 hz, used an active filter frequency of 27.14 hz with a Q 1.77, giving a bass lift of 5.33 db.
So Dickason's boost was applied 2.34 hz above Fb, similar to Keele's and many of E-V's own boxes. Comparing his F3 numbers with/without filter shows the expected 10 hz F3 gain from assisted alignment (and the predicted half-octave extension). With regards to cone travel he writes: "Excursion only 5.8 mm maximum for the speaker/filter combination, plus the typical continuing increase in excursion rate below Fb has been attenuated to a low level. Given the 6 mm Xmax of the driver, the Class I system should provide good high SPL performance." (P. 63).
Also interesting from this simulation, are its cone excursion curves (btw similar shapes or so at higher level to 2235H's in post # 11). At the same low input power level the pre/post bass EQ excursion curves match more closely at Fb. But as the level increases they deviate more from each other at Fb (and elsewhere). Since the boosting filter is also present on the low level post EQ curve, and there's a relatively good match of curves at this level at Fb, then why less at higher level? As if the vent's cone travel reduction effect at Fb (and others) couldn't keep up with more input level.
E.g. On "forum.speakerplans.com/assisted-alignment-b6-loading_topic87595.html" one fellow that went the B6 way did mention "... and made port noise a big problem." No box, driver, port or power info being given its hard to even try to name the culprit with any reasonable accuracy.
It seems the bass EQ filter's sole presence/action, which is independent of power level, doesn't explain everything at higher level. Some Port compression also? There's a coincidence: as power input increases the shape of the post EQ excursion curve changes, and as power rises so does port compression according to Salvatti, Devantier and Button's Port Performance article I mentioned re port size in a previous post. They say:
"As the SPL of a port is increased, there is no escaping some degree of port compression." (P. 32) "Several complications occur in vented designs as the output is increased beyond the point where the air in the port is able to respond in a linear fashion." (P. 19). Examples given: extraneous noises made in the port, acoustic compression and distortion. Some food for thought regarding the above-mentioned issue.
More coming with other drivers.
Richard
Some issues re step-down mode
E2) CONE EXCURSION
No excursion data was given by E-V for the huge 30W driver, their only mention in the data sheet being that "Even at full power input, cone motion is within the linear range." Nice but not enough for a simulation.
The 18" 2245H (Xmax 9.65 mm, Pe 300W), "King of bass" for a long time, did have a limitation in one of the simulations done by JBL (third one below). The db numbers below are rounded for clarity.
In the 8 cu.ft. assisted box, Fb 26hz, the 2245 remains thermally limited (300W) down to 20hz, relative response of - 14 db and max output of 108 db (from Kramer/Timbers P.3). From JBL's data sheets in post # 26, same woofer in larger 12 cu.ft. unassisted box, Fb 25hz, the driver stays thermally limited until 22.5 hz, relative response being - 8 db and max output 114 db, plus at 20hz (displacement limited) response is - 11 db and max output is 109 db. Comparing both boxes, for comparable LF performance (20hz/-14db/108db VS 20hz/-11 db/109 db) you can get 4 cu.ft. smaller box with assisted alignment I mentioned previously. The steeper roll off of the assisted box, with a high-pass filter, usually explains the response difference of the two. If we compare both at 22.5 hz (lowest point where larger box is thermally limited), 8 cu.ft. response - 11 db, max output 111 db; 12 cu.ft. - 8 db, 114 db, so the larger box has a 3 db edge on response and output at that frequency.
Note the 8 cu.ft. Fb 26hz assisted box does not show (in JBL's modeling computer printout) any sign of reduced max output capability nor excursion issue a little higher in the bass range, since tuning is higher and box smaller (see below though).
Also note the 2245H max output of 122 db higher in the spectrum (150hz) is the same as for E-V's 30" woofer at 122 db (100+ hz)(post # 71). The former requires at least double the power input for this level, while the latter requires a huge box size for very deep bass...
In the third simulation done by JBL (12 cu.ft. assisted, Fb 20hz) the 2245H had a little more difficulty keeping up with power input even with its large cone travel rating (post # 26, last document). Here the driver remains thermally limited down to 40 hz (response -3 db, max output 119 db), then gets displacement limited from 37.5 to 27.5 hz, which would be represented by the right-hand excursion bump on a B6 alignment cone travel graph, showing it reached Xmax prematurely in that range. After, it resumes being thermally limited from 25 hz down to 18 hz (resp. -11 db, spl 111 db).
At the worst case point where it ran out of excursion (centered at 32.5 hz, resp. -4 db, max out. 117 db) it could take 243W instead of 300W, 57W less or a 19% input power reduction. In terms of relative db output loss caused by unsufficient excursion, this represents almost 1 db (see printout). (-1db is equivalent to 20% power reduction based on J. Eargle's power ratios table in Handbook of Sound System Design, p.9).
So this "super woofer", in that particular setup, suffered a little, but less than most, from the reduced max output capability in the region ABOVE Fb (region of original box tuning as E-V calls it). That reduced capability generated a little excess cone excursion here. 20hz tuning in a very large box with LF filter boost is a big order for about any driver. Graphically, the right excursion bump would be higher than the VLF left one at practical frequencies (not 15hz). (You'll see later another woofer didn't have its back to the wall, but its a different situation).
For the same 12 cu.ft. box size, comparing the assisted and unassisted 2245 versions, in that order, at the lowest high-spl frequency from each (thermal limit), there's a 4.5 hz difference (18 VS 22.5 hz). Not enough to be bragging about the LF gain here from B6 alignment. Moreover, if we stop the comparison at the LF human hearing threshold of 20 hz, then its only 2.5 hz difference: 20hz/response -9 db/spl 113 db vs 22.5hz/ -8 db, 114 db, giving a 1 db edge on both items to unassisted box, but at 2.5 hz higher. Not meaningful either way in my view.
Kramer/Timbers mentioned for their 300W 2245H boxes vent diameters of 5 7/8" (divided in two 4 1/8" or so equivalent vents seen on picture in original Audio magazine article I have) for 8 cu.ft. box, Fb 26hz, assisted box, and 9" dia. for both assisted (Fb 20hz) and unassisted (Fb 25hz) 12 cu.ft. box. No flared vents here.
They also stated "nominal maximum amplifier power guideline of 800 watts apply" (P.7) for the 2245, more than 2.5 times Pe (I know they used much higher power than that but said "at your own risk"). No concern here regarding the 2245 ability to sustain such program material peaks. But I am a lot more worrying about the 5 7/8" (or 2 X 4 1/8") port's ability to keep up delivering correctly at very high levels of 300-800 watts...
Dickason made no mention of vents for 18" drivers in his Loudspeaker Cookbook 5th ed., while he did mention 6" dia. ports are good for 12" and 15" types (p.53). Yet, If one extrapolates his own port size classification logic described, then one 6" vent would be usable but minimum for 18" woofer. Its difficult to reconcile a 5 7/8" vent dia. (split in two) here with very high power and port compression, specially at the very low frequencies involved... Larger diameter tubes, such as the 9" dia. port used in the 12 cu.ft. box, make more sense to me, though one reason for using a larger size vent was definitely the lower Fb of 20 hz on assisted box VS smaller assisted cabinet's Fb of 26hz. Fb is an item considered re minimum port area, whereas cabinet volume is NOT part of the vent's minimum required AREA equation, but Vb is part of the port LENGTH calculation. More on this later.
Finally, I have an issue with the response curve shown by JBL in post # 26 for the 12 cu.ft. Fb 20 hz assisted box (4th doc.) and response data in the relevant computer printout (5th doc.). For the 8 cu.ft. Fb 26hz assisted box Kramer/Timbers didn't show a response curve, but their computer printout of response data (p. 3) raises the same concern. Above, I've used the numbers at face value in absence of better ones.
I believe the response curve and data given do not reflect assisted alignment, as it should. These B6 with boost/cut filter should "... return the response back to a roughly flat condition ..." as Keele mentions and shows in his paper (P. 356). Obviously, this is not what we see in the above-mentioned documents. What we get are pretty dropping responses as a result of the low box tunings with no EQ boost. Plus the response curve for 12 cu.ft. 20 hz box shows no "knee" in the LF curve, as opposed to the curve for the 12 cu.ft. unassisted 25 hz box (this one is correct, but tuned a little too low in my view).
The "offending" curve and data are not typical at all of B6 response, i.e. relatively flat LF followed by a steep drop (as in posts # 11 & 29). Here its more typical of lower tuning with no EQ, than anything else.
Considering this matter, the accuracy of the db comparisons/results mentioned above for the assisted boxes should be taken with caution. In my book, they can still be SOMEWHAT compared but all should be considered as unassisted alignments I think, with two of them tuned lower than normal...
Next post getting closer to completion, should follow soon.
Richard
EDIT: Two examples to illustrate the offending data. For 8 cu.ft. assisted box boost applied at 26 hz so response should be about flat there, but at 25hz (closest point in the data) the printout shows its - 8.9 db. For 12 cu.ft. assisted box boost applied at 20 hz so again response should be about flat there, but at 20 hz the printout shows its - 9.4 db. However, the displacement and thermal limits indicated on this printout appear to be compatible with typical B6 alignment...
Some issues re step-down mode
E3A) CONE EXCURSION (THIS IS THE FIRST PART OF A TWO-PART POST, NEXT ONE TOMORROW)
The more recent 18" 2269H (Xmax 19 mm, Pe 1,200 w) was modeled at 600W in post # 29 with a 8.9 cu.ft. box, Fb 26hz, in both assisted and unassisted versions. Assisted alignment was with filter Q 1.5 for improved LF response flatness as later explained, instead of the usual Q 2, not a big difference though. Compared to JBL's assisted 12 cu. ft. Fb 20hz 2245H box discussed in my post E2), the present one has higher tuning, smaller box size and boost/cut filter Q 1.5 VS Q 2, both at Fb. This represents a less difficult "environment" to perform. In fact, here the 2269's situation is closer to the 2245's assisted 8 cu.ft. Fb 26hz filter Q 2 box.
The 600W assisted data & custom response curve from the above post shows the almost 8 hz lower F3 mentioned in a previous post, plus about 4 db higher output at 30hz (and about +3 db around 26hz) compared to the unassisted version (+ 4 db being equivalent to 2.5 times power according to Eargle's table). Around 22 hz both curves are basically at the same output level. This modeling, as some others, doesn't show a whole lot of LF gain from assisted alignment for this other "super woofer". Not really surprised by the result considering what I mentioned previously about some of these woofers who appear to reap less (not zero) VLF extension benefit from step-down, than lesser capable drivers. Still a nice 3-4 db output improvement here at 26-30hz.
The "comparable" 2245 assisted 8 cu.ft. Fb 26hz filter Q 2 box, had an F3 of 26 hz and unassisted F3 at 35 hz (9hz gain) VS 24.6 hz and 32.5 hz (8hz gain) respectively for the 2269H. These numbers for both drivers are pretty close to each other. BTW, I left aside for later the different box QL used in the modelings, QL 7 for 2245 VS QL 5.4 for 2269 and its impact on Vb, since I have an issue related to that (not re 2269 modeling).
At the same 600W power level, the vent air velocity curves around Fb, shown in that post, indicate velocity is notably higher for the assisted design which, IN PRINCIPLE, looks normal because of the lower frequencies involved. On the excursion graph we see the assisted 2269, though filter is Q 1.5 instead of Q 2, still shows a left bump a bit higher than the right one, contrary to some other drivers seen here where its the other way around. So this driver doesn't appear to be affected (or is less so) upper range by the small maximum output capability reduction, caused by step-down, in the "region of original box tuning" as E-V calls it. Probably due to driver's double cone travel capability and higher tuning in smaller box than the assisted 12 cu.ft. Fb 20hz 2245H box showing about 1 db decrease. The more comparable assisted 8 cu.ft. fb 26 hz 2245 box didn't show on the printout any output capability reduction either.
However, the cone excursion curves shown, also at 600W, are more of a concern as they bring the question of why the very little cone travel reduction seen around Fb, on both assisted and unassisted versions, compared to other vented box excursion curves?
Looking at a few data sheets for JBL made boxes using the 2269H (e.g. VT 4880A/4881A; ASB 7118/7128 these two btw use triangular type vents I mentioned in a previous post; MD 7 horn loaded, not relevant here), specific vent area or diameter used are not mentioned.
For comparison purposes, I noted the 2245H box example given in its own data sheet is Vb 10 cu.ft., Fb 30 hz and port area of 50 sq. in. This is equivalent to an 8" dia. tube! I modeled the same box in Winspeakerz and got a minimum recommended vent area of 45.7 sq. in. (7 3/4" dia.), reasonably close. Moreover, I've done another simulation for the same driver in the 8 cu.ft., Fb 26 hz, QL 7 box where the software tells me the minimum recommended vent area is 39.6 sq. in. or just above a 7" dia. port (38.5 sq. in.) VS JBL's 5 7/8" mentioned before.
By the same token, I modeled the 2269 in the 8.9 cu.ft. Fb 26 hz QL 5.4 box to see what would be the software's minimum recommended vent area for this one: 75.1 sq. in., close to a 10" dia. port (78.5 sq. in.) or to 2 X 7" vents (77 sq. in.). This looks pretty large, but remember this is a very high excursion (included in Vd calculation) 1.2 KW rated driver, modeled in post # 29 at half Pe (or double that of 2245). Nothing wrong with this, but "the pedal isn't yet to the metal" as there's still 600 more Watts to go (+ 3 db). And we haven't even mentioned the program material power/peaks (e.g. see JBL ASB 7118 data sheet).
If one finds such vent size(s) really inconvenient and has no intention whatsoever to use the driver anywhere near 1.2 KW, sure someone can get away with less port area at lower spl. Then you'll need to model accordingly in software with a lower driver power rating number in order to see what you can live with: vent size vs power. If you end up selling the cabinets one day its not a good idea though. Buyer looking for a very high power box gets so so vent capability (i.e. compression and noise at high level).
The second part of this post should follow tomorrow. It is written but I need to review it and add bits and pieces.
Richard
Some issues re step-down mode
E3B) CONE EXCURSION (PART TWO)
With regards to potential port compression in view of the cone excursion curve shape at 600 Watts shown in post # 29, I note the port diameter used in the modeling is 6" with two flared ends, more aerodynamic profile improving air entry and exit of the port. However, that doesn't change vent tube diameter required. It's still better than a standard straight vent (with "ifs and buts"), as show Salvatti, Devantier and Button.
In comparing both 2245/2269 port dia. info in my preceding post, the 6" vent used in the 2269H modeling appears too small considering the more or less 7+", 8", 9" and 10" vents mentioned for the 2245 or 2269, in order to minimize port compression at high spl. This quite possibly explains the pretty small cone travel reduction effect of the vent around Fb for the 2269 box in post # 29. If a 7+", 8" or 9" dia. vent tube is the way to go for the 18" 2245, then the 18" 2269 with much higher power and Xmax may very well justify the even larger 10" or so vent. There isn't a large Sd difference for the two drivers (.129 vs .123 sq. M.) but the very large one on Xmax gives the 2269 a notably higher Vd (1.23 vs 2.33 L) 89% more! (Sd x Xmax = Vd).
I extrapolated previously Dickason's vent size use classification, showing 6" dia. port would be usable but minimum for 18" drivers. This also depends on Fb and power input. The deeper the Fb and the higher the power, the closer trouble gets. In the specific modeling reffered to, power is very high and Fb pretty low, hence the issue. Driven with less power and/or using a higher Fb, would probably result in being able to get away with it, up to the point where the temptation of ever increasing the volume control gets too high...
I also did a quick modeling in Winspeakerz for the same 2269 box, but with the recommended vent area (75.1 sq. in., about 10"dia. port), and varied power input at 10, 100, 300, 600 and 1,200 watts to see the effect on excursion curve shape as power increases. The results I got, even up to 1,200 watts, are more typical of vented-box excursion curve shape with a "deeper diving" curve at Fb (i.e. more cone travel reduction). I believe this tends to confirm the port issue raised here regarding such 6" dia. vent at high power.
The minimum vent dia. formulas (e.g. R.H. Small; M. Engebretson) consider Fb and Vd to determine the maximum amount of air that will need to pass in the port. Port diameter is then determined as the minimum requirement to avoid port noise. In practice, input power level acts as a "modulator" of cone excursion level, which in turn varies the air flow in the port tube from a very low up to a very high amount.
While we're on the subject of a very high power/output driver, out of curiosity its interesting to learn from the same three Harman Engineers, how bad vent compression can get at some point: "Extensive benchmarking of current designs reveals that current attemps at high output ports suffer from compression effects at high drive, showing that at very high levels all ports eventually "lock up", limiting the maximum output. (...) At these levels the output from the port is 180° out of phase with the output of the cone, creating a nearly complete cancellation of low-frequency energy." (P.19, & 23) (!) In addition they mention "Also at lower frequencies (higher velocities) the effect [RMC: compression] is much more pronounced."(P. 29) So the quest for ever deeper and higher SPL bass has another obstacle on its course...
If you can get a clear picture of the JBL HLA Series 4897/A subwoofer (4K US$ in 1997!), as in the 1996 color Brochure, or in sketches shown at the end of data sheets/technical manuals, the unusually looking vent is a sample of JBL's work done on newer very high-power vents: very large rectangular or so with rounded corners and flared ends. In a way, the concept of the round flared vent, but increased substantially and adapted to a very high-power box.
The 4897/A's pretty large vent in close proximity to both woofers happens to coincide with Dickason's experiment where his larger vents "... very close proximity to the driver seems to cause the least amount of disturbance." The smaller of his vents having less problems at a distance from the driver, and double ports producing less problems when at a resonable distance from each other. (P. 55). The above should give viewers more food for thought.
Finally, the point here is the use of "super woofers" involves knowing much more about some important related issues behind the curtain, and seldom mentioned, in order to achieve or use a super driver's full potential. In my view its nicer to know ahead of time what this may entail, such as:
e.g. Speaker builders have to realize at these extreme power levels there's no room for flimsy cabinet construction. It must be "built like a tank", extremely strong, heavily braced and sound dead when hit with a mallet to avoid any panel vibrations at very high spl. In the 2006 Vertec subwoofer/2269H press realese and other docs, JBL mentions "The enclosure features advanced construction techniques, replete with JBL PlyMax engineered panels...". That should give some idea about the requirement.
Also, the amount of work involved to make or find a proper high-power vent(s) should not be underestimated, unless one is lucky to find the appropriate ready-made device, which may not happen.
More to come.
Richard