Something stirs in the Undergrowth

[Lynn Olson]

The last time I talked to Gary, the box volume for the 416B (Alnico) is around 2.5 cubic feet, and the UltraTouch filling has the effect of raising it to 3 cubic feet or so. I would imagine teased-out wool (suitably moth-proofed) would have a similar effect.

The impedance curve is considerably smoother in the UltraTouch filled box, as you might expect.

[IslandPink]

Thanks Lynn. I have plenty of long-haired wool to use ( no, not just because I live in North Wales ) , I will add some more ( I used a little ) to the box next time it's open .
I also have a plan to add a zobel to the input ( eg. 7.5R/10uF ) to bring down the response above crossover , as from sig.gen checks lst night there's lots of energy at 3-4kHz ; make it easier to evaluate .

Beyond this, I now hear more confusing voices in my head telling me to reconsider the whole plan, bearing in mind what Dick Olsher did with the Basszilla . ( I have fairly reliable data from Jon Noble that this is a very good solution ) .
I do know that the best midrange I've heard has always been from the OB solutions using Fostex or AER drivers . It makes sense that these light-coned very sensitive drivers are also very articulate on vocals and work very well at least up to 2-3kHz . The roadblock to doing this is that my current bass idea has always revolved around the tapped horn which can realistically only be crossed over at 100Hz max . This is what's causing me to struggle and compromise with drivers for the next stage up . Maybe my optimum solution might be a 15" driver either OB or Sealed-box to 200Hz, then something like a FE208Ez ( let's not break the bank ! ) from 200 to 750Hz ?

Plenty of work to do in the meantime.
I'll have a better idea if I can live with a sealed-box midrange when I've tried the zobel and the 8PE21 driver substituted onto the 30L enclosure.

[slowmotion]

Beyond this, I now hear more confusing voices in my head

You too, huh ...

[IslandPink]

Sure thing - goes with the territory ( speakers ) !
Now, I got to playing around with Hornresp and trying out drivers ( if I could find anything suitable ) for OB and sealed-box. This started to confuse me, so then I went upstairs for a bath, to relax.
Then the voices started again, in a different way. They started to say "transmission-line bass ... transmission-line bass" .

So this adds another potentially useful permutation to the matrix : transmission-line ( vertical tower - good ) to 2-300Hz , then OB with wide-ranger 8" from there to 700-800Hz, then ... you know the rest .

Now, Lynn would know a thing or two about this subject.
Seems to me you might get more bass extension for a given driver size, with good fidelty , using TL . Any thoughts, Lynn ? Am I looking for Qts about 0.35 to 0.45 for TL ?

It's funny because I started out my DIY speaker life with an IPL S4 kit - which had great bass extension, but limits elsewhere . Latest ones use Kevlar drivers - not sure about that .... but design can be worked-on of course. Ivan has tried various units over the years and perhaps could advise.
http://www.iplacoustics.co.uk/

Now, how to get the TL out of the way of the OB with the fullranger ....

[Lynn Olson]

Well, TL's do work right through the free-air value of Fs; the Ariel drivers have a free-air Fs of 60 Hz (and sometimes higher, for some models of Vifa), yet in subjective terms, the Ariel has pretty decent bass down to 50 Hz or so. (The internal lowpass filter for the REL Strata II subwoofer has to be set no higher than 40 Hz ... there's too much overlap at 50 or 60 Hz.)

The upper bass has a noticeably free and open quality compared to closed or vented boxes, and I have no explanation for this. Planet10 in Vancouver, Canada, has also noticed this, and favours TLs and resistive-vent enclosures for that reason.

The gotcha with TLs are quite large enclosure size compared to comparable vented boxes ... but pretty much every TL I've heard has much better bass than conventional T/S aligned vented boxes, no matter what the models and mathematics say. If I could tolerate the box size (which would be huge), I'd put the GPA/Altec 416 Alnico in a TL and not look back. But a TL for a 15" driver is probably going to be bigger than I can accept.

It's a myth that TL's depress efficiency. Not true at all. The only way you can do this is go crazy on the stuffing, which then starts to mass-load the cone. If you do that, not only does the efficiency go down, but dynamics are destroyed as well. Drivers do not like additional mass-loading, in my experience.

Gary Dahl has been around the block a few times with Altec A7's, 414's, 416's, and 515's. He feels (and I agree) the notorious flabby Altec bass is nothing more than putting superb drivers into cabinets that are just too big and not that well damped. The Altec Model 19 has about 10 cubic feet, and the A7 is pretty similar. By contrast, Gary's new cabinets have 1/3 the volume, and the bass is crisp and fast, not blurry at all. There's nothing wrong with the drivers: they just have short-throw voice coils, and putting them in underdamped bass-reflex cabinets lets the VC go into the nonlinear region.

Returning to TL's, they also have pretty good bandwidth, since the interior volume is constrained to a narrow pipe which is heavily damped along its length. They are basically free of the box modes that are so troublesome for conventional cabinets in the 300~800 Hz region, so there is a similarity to OB sound (but with more power, dynamics, and bass extension). In subjective terms, the 200~500 Hz region is better than other types of cabinets ... more vivid tone colours, freedom from the usual Helmholtz box murk-n-mush, and a solidity to the deep bass that very different than the overhang of vented boxes.

Hmm ... along those lines, although a TL for 416 would be huge, a TL for the 414 would be more reasonable, and the GPA 414 (Alnico) has the same efficiency (97 dB/meter/watt) as the 15-inch 416. TL's for 8" drivers would be correspondingly smaller, of course, at the expense of dynamic range.

I feel a little reluctant to recommend a TL for a single 8" driver since that's pretty much the same emissive area as the pair of 5.5" drivers for the Ariel, and you've been there, done that, got the T-shirt. I would guess you want a lot more dynamic range and more vivid tone colours than the Ariels ... which are a twenty-year-old design by now.

[chris661]

I'm part-way through building a transmission-line/quarter wave pipe for a 15" PA driver. We're talking 97dB@1w, -3dB just below 40Hz, and will stand 1.2kW within driver Xmax.

Driver is Beyma 15P1200ND, and the cabinet is around 120L. Mark, if you happen to be round this way, gimme a shout for a demo. It'll be driven off a 3.4hp (yes, horsepower) Peavey amp

Hornresp suggests it'll get nasty above 150Hz, but I suspect the folded pipe will keep most of the mush at bay. Testing will tell more, so I'll be getting some frequency response charts done and see how high this can be used.

Offset Driver,

S1=550
S2=550
S3=550
L12=62
L23=128

http://www.beyma.com.ua/wp-content/uplo ... 200NdN.pdf

Chris

[IslandPink]

Just added a 10uF/7R zobel to the 'Box' unit. Sounds quite a bit better .
I'll try & do some measurements Friday, then change to the B&C 8PE21 for comparison and do a bit more bracing and add some more wool .

Thanks for the offer Chris, sounds like a fun system. Rather industrial-strength compared to what I'd want to do, but I bet it'll be impressive.
I think the first option I'd like to check though would be one of the IPL kits using the 10" driver, so I might check if there are any in the area. It'll give me an idea what TL low-bass sounds like, in the 30Hz range, to compare against the Ariel .
I reckon it should be OK to use something like a 10" unit with eg. 40Hz Fs for my purposes. That should get me into the mid-30's Hz which will be fine. More important is that if it's handing over to a light-coned driver eg. Fostex or Lowther in the mids, there would have to be some consideration of cone material to get an acceptable 'blend' . I'll have a look around & see what that might be, but any suggestions would be useful .

[chris661]

Hi Mark,

Industrial strength indeed, though the PA drivers I'm using have quite a lot in common with higher-end HiFi drivers (shorting rings in the magnetic circuit), so the quality ought to be there...

The Hornresp inputs I've given above might be worth trying for a 10" driver (halve all the areas, and play from there). The overall length tunes to the quarter-wave frequency of the low-end cutoff, and the cross-sectional areas seem to alter the Q of the resonance. Huge area (like an oversized ported box) gives a really peaky, underdamped response, and an undersized area will "choke" the bottom end.
If you add stuffing/lining, that'll stop most of the mush getting through the transmission line, so you'd be free to run up to a compression driver.

In terms of recommending drivers, which frequency range are you trying to cover?

IIRC, you wanted to use a tapped horn, but now you're talking about going to 35Hz with a 10"...

[Lynn Olson]

Good luck, Chris661 and IslandPink. TL's sound better than the sims indicate ... they sound less resonant than the equivalent T/S box, despite the predicted FR curves. No good idea why (although I have some speculation).

Chris661, what do the sims say about the old trick I used in the Ariel, a folded labyrinth at the exit, used as an acoustic lowpass filter? By measurement, the labyrinth works, since the FR of the vent output is smooth from 20 Hz to 100 Hz, with very rapid rolloff above that. It looks just like a subwoofer output.

[chris661]

Hi Lynn,

I'd need to learn to use Akabak to answer that question.
All I can say for Hornresp is that the above-band mush isn't anywhere near as bad as it thinks, and adding more folds (as you say, creating an acoustic low-pass) seems to take this further.

Once my build is done, I'll be doing some measurements, ranging from a quick pink noise test, to taking SPL readings on test-tones every couple of Hz. For curiosity, I'll measure the output at the transmission-line exit and compare to the Hornresp sim of what'll be coming out there.

Had a quick look online, and it looks like there's quite a few different TL labyrinths drawn. Do they all measure similar?

Chris

[Lynn Olson]

Hi Lynn,

I'd need to learn to use Akabak to answer that question.
All I can say for Hornresp is that the above-band mush isn't anywhere near as bad as it thinks, and adding more folds (as you say, creating an acoustic low-pass) seems to take this further.

Once my build is done, I'll be doing some measurements, ranging from a quick pink noise test, to taking SPL readings on test-tones every couple of Hz. For curiosity, I'll measure the output at the transmission-line exit and compare to the Hornresp sim of what'll be coming out there.

Had a quick look online, and it looks like there's quite a few different TL labyrinths drawn. Do they all measure similar?

Chris

I've only measured the Ariel, no idea how others measure.

I would certainly measure the vent output as well as the driver output (nearfield); there's a good chance it will not be the same as the simulation. This is partly because damping materials like UltraTouch and long-fiber wool are hard to model, and I'm not sure models of multi-fold labyrinths are accurate either.

Sometimes you just have to build it and see ... and listen.

[IslandPink]

Chris
The application of the TL would be to hand over to a light-coned high-efficiency driver either in a sealed box, or on OB, for the midrange . On that basis, the TL will be crossed at something like 200-300Hz, and have to be pretty clean to 1 to 1.5k at least , as the crossover slope will be low'ish' . I don't believe any biggish driver with a low Fs suitable for a TL will be able to keep up with the compression driver in the 500-1000Hz area. What the TL would do is extend comfortably up into the 200-300Hz crossover area - unlike the tapped horn . It wil free-up driver choice for the midrange - allow me to use lower-Qts units that are more efficient and also rolling-off at eg. 200Hz on an OB . With the tapped-horn there'd be a gap I couldn't fill from around 100 to 200Hz .

I exchanged some email again with JamesD in the last couple of days . He's remaining in the middle-east for 6 more months now, but he's OK health-wise . He was very complementary about a TQWT project he'd done with Scott's help, for a Mission 770, and encouraged me to look at those sort of designs too .
It's clearly a big area of endeavour, with a lot of design variants out there . One consideration I'll have to make is that an odd-looking cabinet would ideally be required, something that isn't more than about 60cm high in one area at least ( folded in 3D or laid on the floor ) so that I could mount a decent-sized perspex baffle above if I go for the OB option for midrange. For boxed (sealed) midrange , that could be incorporated into a more normal TL enclosure, somehow.
Either way this would be an odd-looking beast, but if it works it'll be fun to have found a combination that hasn't been brought together before ( I think ! ) .

On the other hand it's probably safer from an integration point of view to go down the OB route and just have a 15" driver low-down on a larger OB to augment the midrange driver. This would give me speed and tone for sure, with a little less 'slam' I guess.

ps . In all of these cases the Bass option will be driven from an SS amp , so I'm not especially restricted to high-efficiency drivers or matching anything cross-over-wise to the midrange unit. However, I think treated paper cone would be best suited to the bass driver to assist integration with the mid.

[chris661]

I've only measured the Ariel, no idea how others measure.

I would certainly measure the vent output as well as the driver output (nearfield); there's a good chance it will not be the same as the simulation. This is partly because damping materials like UltraTouch and long-fiber wool are hard to model, and I'm not sure models of multi-fold labyrinths are accurate either.

Sometimes you just have to build it and see ... and listen.

Ah, I wasn't clear. There would appear to be lots of variants on the Ariel transmission line.


Mark, you might well be able to go from the transmission line straight to the compression driver. A 10" driver might do nicely. You're looking for Qts ~0.3, Fs ~40Hz.
Open baffles present their own set of challenges. There's a lot of ripple towards the top of the pass band.
The 8" B&C (8PE21?) driver you have will do something down to 40Hz in a transmission line like mine, but needs a few ohms of series resistance to raise Qts. You could also try a ported cabinet. Again, some resistance would be needed to get Qts up. If you can get away with one of those drivers per side, covering 40Hz up to the compression driver, I think you'll be laughing.

Chris

[Lynn Olson]

The 8" B&C (8PE21?) driver you have will do something down to 40Hz in a transmission line like mine, but needs a few ohms of series resistance to raise Qts. You could also try a ported cabinet. Again, some resistance would be needed to get Qts up. If you can get away with one of those drivers per side, covering 40Hz up to the compression driver, I think you'll be laughing.

Chris

I second what Chris661 is saying. Drivers sound remarkably different in closed boxes compared to transmission lines.

I first heard the Vifa 5.5" in a pint-sized minimonitor that Lineaum in Portland was prototyping. I had leased my MLSSA system to them for a year, they were able to develop their new 1" tweeter with it during that time (which they then licensed to Radio Shack), and when I returned to collect the MLSSA system, they let me hear the new speaker they were working on.

I was impressed with the smoothness and natural tone of the Vifa 5.5" midbass driver ... the best I'd heard from polypropylene ... but the sound was very bass-shy and had even less dynamics that the LS3/5a. Doubling up the drivers and putting them in twin transmission line solved both problems, and had the unexpected effect of much better tonality as well. I credit the TL/labyrinth for that.

The classic early-Seventies transmission lines with the KEF B139 gave a lot of people the idea that TL's only work with heavy cones. Actually, they thrive on light cones, and have particularly good midrange qualities. I suspect the much more rigid construction (compared to closed boxes) and freedom from the usual standing-wave box modes are a large part of the characteristic sound of TL systems.

TL versus box is one of those areas where measurements can diverge from subjective perception. The little wiggles on the impedance curve that are created by internal standing waves appear innocent, but that's where the characteristic "shut-in" sound of a closed-box (compared to an OB) is coming from. I know people who have experimented with simply removing the back of a closed-box and noticed a remarkable decrease in box colouration, which is a hint it's the Q of the internal standing-waves that are audible.

The ear is extremely sensitive to box colourations that barely appear on a MLS waterfall ... perhaps because musical instruments use box colourations to define the sound of the instrument, and the closed-box modes conflict with this perception. I think that's why both transmission lines and OB's seem to sound better than they measure ... they may have delayed-sound effects, but it doesn't directly contradict musical impressions.


Technical Digression: I view TL's, resistive-vent enclosures, and open-back boxes that are filled with UltraTouch or long-fiber wool as part of a spectrum of resistive-loaded enclosures.

Both closed and vented-box enclosures are almost pure reactances as seen from the driver; a closed-box is an air-spring, or acoustic compliance, in series with the compliance of the driver. In typical closed-box enclosures with system Q's in the 0.65 to 0.9 range, the box compliance dominates by a factor of 5 to 10 over driver compliance, the classical acoustic-suspension alignment.

A modern T/S vented-box is a more complex 2nd-order spring/mass system tuned to match the 2nd-order spring/mass system of the driver. Line everything up perfectly and you have a 4th-order highpass filter ... however, the driver parameters are not stable at all signal levels, so the system can go out of tune under dynamic conditions.

Resistive (not reactive) enclosures have an interesting property that driver parameters do not need to be tightly matched to enclosure characteristics, so if Fs, Qts, or Vas shift under dynamic conditions, the driver/enclosure system does not become misaligned. This property is also shared with OB, where driver characteristics alone control the Q and shape of the lower corner frequency.

Misalignment under dynamic conditions might sound like a trivial, rather theoretical problem, but it should be remembered that the group-delay of the lower F3 corner is quite long, in the tens of milliseconds, and if this shifts with a sharp transient, the change in group delay may be audible at much higher frequencies.

The upper-frequency characteristics are different as well. Although reactive at very low frequencies (1/10th of a wavelength or less), closed and vented-boxes become very complex at lower-midrange frequencies of one wavelength or less. Without damping, the Q of standing waves is very high, and they exist in three dimensions. Acoustic damping reduces the Q and magnitude of the standing waves, but it does not completely remove them, and I suspect they remain audible at fairly low levels compared to the direct sound from the driver. Boxes are not as simple as they appear.


P. S. The only version of the Ariels I've ever measured have been my own, which are technically the 1.9 version, a bit simpler than the ones on the Web page. As far as I can tell, the differences between them are very small.

[chris661]

Hmmmm...

The T-nuts arrived a few hours ago, so the transmission-line PA sub is up and running.
It is, in a word, silly.
It'll turn your head inside-out down to an honest 40Hz. Currently powered off a bridged Peavey PV2600, for laughs.
I suspect it'd even cope at Owston!


That said, there's no lining or stuffing anywhere in the cabinet, so there's something weird going on in the >150Hz range.
No problem for me, since I'm crossing down at 70Hz anyway, but might be worth bearing in mind if you do decide to go down this route - lining/stuffing will lose some LF output, as well as smoothing the top end. Its all compromises...

Chris