Supravox 8" TLonken... while city of 22M laid flat

[wchang]

(Addendum)

Supravox still wins the "smile-on-face" contest. Bass (A-) is clear and deep just lighter than some. In comparison, the original labyrinth bass would be B- and the simplified TLonken shown below B+. The gap between 1.5m and 1.25m line-length is partially mitigated by the "Area Ratio" effect (MJK Table 1, courtesy @planet10 ) and more/easier stuffing: 1.5m 1:6.25 taper effective 1.47X 2.2m quarterwave frequency~40hz; 1.25m 1:7.25 taper effective 1.52X 1.9m quarterwave frequency~45hz. Eyeballing said table with a calculator, TL (reducing) taper 1:R has effective-length multiplier R^(1/4.75) independent of frequency. Horn (expanding) taper R:1 has effective-length divider 2-1/R^(1/4.75). VERY CLOSE TO, anyway! So the theory goes

 

[bob53]

Hello wchang,
If I put TB1772 in your box will it be OK?

 

[wchang]

Hi bob53, I looked up the TB's specs: Qts=0.27 Vas=90L Fs=42hz ventedVol=23L ventedF3=65hz (compare F200A specs in post #20). The 17L 1.5m 1:6.25 <40hz TLonken should work if there's good room-reinforcement; 23L (make cab deeper or taller simplified) would be better and allow the tapered line to be fully fleshed out. If you started with a run-of-the-mill cab and inserted cardboard dividers to create & lengthen the line, the jaw-dropping "good bass" will go incrementally deeper. I stop when my music requirements are met (organ/piano/bass-drum climaxes). Also, my one TB (W5-1880) takes/taking a very long time to run-in.

BTW, tapered ML-TL isn't new just somehow overlooked by MJK, Bob Brines et al. So-called "Fat Thor" ML Beta is 83% the volume of "equivalent" TL design.
FatThor tapered TL 191x1111x582=123.5L https://www.diyaudio.com/community/threads/clarity-on-seas-thor-kit.64799/post-743582
FatThor ML tapered TL 191x1093x492=102.7L https://www.diyaudio.com/community/threads/clarity-on-seas-thor-kit.64799/post-744315
(I don't know how many FatThor/etc. got built -- it's a long thread -- after MJK's charitable comment: "I don't know for sure but I would think that combining the drivers and modeling them at one location is providing a better response then what really will occur, very optimistic result." https://www.diyaudio.com/community/threads/clarity-on-seas-thor-kit.64799/post-822290)

 

[Scottmoose]

For reference, I would be very unlikely to design a box that way again. Were I to do so, I would be more likely to simply use a higher taper ratio for a similar result. Tapering TLs are by definition mass-loaded, so if you want that configuration, you should be able to achieve the desired tuning by adjusting that and Vb to provide the desired gain BW (assuming the driver is capable of it).

Regarding Martin's comment about the driver positions, as frequency rises, he's undoubtedly correct. However -Thor was designed as an Augspurger type TL (notwithstanding a somewhat excessive reduction in Vb); it's not far removed from being an aperiodic line, the object, as shown in both George's, and d'Appolito's measurements, being no more than 1dB null at F3 and at the very most 1/2dB at F5 etc., so given the higher output is almost completely attenuated, what differentials that would exist in an un / lightly damped pipe should not have a major impact on behaviour for a TL of this variety.

 

[wchang]

Don't know if the 20% larger 123.5L Fat Thor was a highly suboptimal tapered-TL design in terms of volume.
Citing MJK, Bob Brines has shown (ppt or pdf, see https://www.diyaudio.com/community/threads/low-sensitivity-coaxial.396170/post-7282796): Tuned to the same alignment frequency, all five types deliver equal SPL and fall-off, but line volumes are straight-TL(100%), TQWT(70% or 76%), ML-TQWT(62% or 50%), ML-TL(57% or 51%), tapered-TL(51%). Too bad no ML-tapered-TL (TLonken).

Thought experiment 1: TLonken is an "approximate" tapered TL. One can Hornresp a few hundred times over (area ratio, average CSA, slot dimensions if any) to optimize/minimize volume given target QW alignment and output, thereby quantify whether/how much it helps to have the ML slot as a design freedom.

Thought experiment 2: The bends and folds in the line are empirically important to the sound, in two ways. First as filter to reduce output, more so the higher the frequency including QW harmonic peaks. Second, if a bend is treated as a halfwave pipe to the exit terminus with reduced length and CSA, and derived, mulitply-reflected paths are similarly considered for standing-wave resonances, the calculated summed response might be somewhat different from a straight line. Perhaps at least the first fold/bend can be simulated.

 

[bob53]

Hi wchang,
Thank you for comprehensive answer.I must explain why I asked for your opinion.A few years ago I made, based on Hiraga calculator, TQWT speakers
with Jamo IO165(wall spk.)+Jantzen tweter.They mesured good and sounded decent. I made my own TQWT design for TB 1772,but it is BIG box.
My room is small and speakers (SBA+RALL-on the picture) are about 1.5m away from my ears.
When I saw your design,I liked very much.But I know that the line must be lengthen to ,2,15m for 40Hz and volume has to be risen to 23l as you wrote.
I shall make it with 2 addtional "tube" and see what will hapen.
Regards

 

[wchang]

You can most likely stay with 1.5m tapered and ~20L once the TB are broken-in. For a while there probably won't be much bass in any cab.

 

[Scottmoose]

Don't know if the 20% larger 123.5L Fat Thor was a highly suboptimal tapered-TL design in terms of volume.

Well, I designed it, so I'm in a reasonable position to say on that front. As far as Vb goes, it's better for LF gain than the original Thor enclosure, which is loosely based on Augspurger's alignments, albeit with volume reduced from the nominal 82 litres George's formulas provide in order to keep box size down. However, were I to design it again, I would revise both the volume & internal taper ratio. While it's OK as it is, it's not as good as it could be. I'm a better designer now than I was two decades ago.

Citing MJK, Bob Brines has shown (ppt or pdf, see https://www.diyaudio.com/community/threads/low-sensitivity-coaxial.396170/post-7282796): Tuned to the same alignment frequency, all five types deliver equal SPL and fall-off, but line volumes are straight-TL(100%), TQWT(70% or 76%), ML-TQWT(62% or 50%), ML-TL(57% or 51%), tapered-TL(51%). Too bad no ML-tapered-TL (TLonken).

Except they don't, quite. Bob's example suffers a little as it's not entirely like-for-like, and acoustically large to begin with so in relative terms gain should not be a problem for the chosen Fb. However, you can in fact see that gain is reduced for the tapered line relative to the straight, with its greater Vb, and clawed back with the expanding line (horn). As noted, your 'ML-tapered-TL' is just a tapered TL (tapered TLs are mass-loaded by definition). It's just a slightly over-complicated geometry thereof; you'd get more or less the same results by using the same Vb & simply adjusting the length & taper ratio to achieve your desired Fb.

Thought experiment 1: TLonken is an "approximate" tapered TL. One can Hornresp a few hundred times over (area ratio, average CSA, slot dimensions if any) to optimize/minimize volume given target QW alignment and output, thereby quantify whether/how much it helps to have the ML slot as a design freedom.

You can, but as noted, ducted vents tend to be most useful with untapered pipes and horns. Since a tapered pipe is innately mass loaded, there doesn't seem a whole lot of value in providing only a small amount of loading via that, and then adding a duct to provide what the basic geometry fails to provide -assuming you're holding Vb static, you might as well just just a higher taper ratio to achieve your desired tuning & have done with it. No harm in it of course, it's just a somewhat over-complicated way of going about it. The caveat there I'd add is for tapers created by (in this case a contracting) cascade of untapered pipes that are approximating a general target taper ratio, e.g. 6:3:1 or whatever. But in that case, the final part is an integral stage in the overall line rather than viewed as an addition.

Thought experiment 2: The bends and folds in the line are empirically important to the sound,

'doesn't have to be empirical, except in the sense that all science is ultimately derived from empirical testing & observation of behaviour.

in two ways. First as filter to reduce output, more so the higher the frequency including QW harmonic peaks. Second, if a bend is treated as a halfwave pipe to the exit terminus with reduced length and CSA, and derived, mulitply-reflected paths are similarly considered for standing-wave resonances, the calculated summed response might be somewhat different from a straight line. Perhaps at least the first fold/bend can be simulated.

They all can be, if you've the software to do so. That's how most of my commercial horns are modelled at some stage of their design process.

 

[wchang]

I'm a better designer now than I was two decades ago.

We would each of us like to think that were the case.... But some youngster had come up with 102.7L FatThor ML tapered TL, the record to beat.

I checked again Bob Brines' pdf and (to my eyes) the 1:3-tapered-TL/ML-TL/ML-TQWT bass curves were all indistinguishable (75/85/91/94dB at 40/50/60/70hz) and <1dB lower than the straight-TL. I didn't re-check his ppt because the two TLs were unchanged from his pdf. Not that it was a definitive study -- only one driver example and the taper area-ratio 1:3 was probably not fully utilizing that type of line geometry.

They all can be, if you've the software to do so. That's how most of my commercial horns are modelled at some stage of their design process.

More info would be appreciated by all!

 

[planet10]

with volume reduced from the nominal 82 litres George's formulas provide in order to keep box size down

The reason why the original line was good for 1 driver, as he confirmed when he made the mistake in measuring them. But not for the 2 used. Hence the 3 alternate enclosures, 1 i did straight off MJK tables and 2 that Scott designed using a more comprehensive set of tools.

dave

 

[techsi]

Dear wchang,

Thank you so much for sharing your work! You have been a source of great inspiration to me.

I would like to ask you how you see the use of a new Supravox 215 Heritage bi-cone driver with field-coil magnet in your TLonken 17L 1.5m cabinet. The possibility of varying the operating voltage of the field-coil magnet should allow fine tuning of the Qts value (I plan to use a battery supply).

Do you think it would work properly without applying modifications to your design?

Thanks again and congratulations,
Gianluca

The driver : https://www.supravox.fr/products/215-heritage-bicone-exc-haut-parleur-large-bande-25w-96-5db-8-ohms

 

[GM]

Greets!

??? It's not a field coil and didn't see it as an option except the EXC series.

 

[wchang]

Heart takes a leap seeing those beauties. The 215 heritage (bi) cone itself should be very similar to mine. I don't have first-hand knowledge of field coils though, especially how to lower Qts to near ~0.3 like the vintage permanent magnet 215. Since you are not constrained by a ready-made cabinet, I would suggest a slightly wider and deeper box, so there's more space for the field coil, and the first divider board tilted more to redirect reflection. Greater width should lower baffle-step a bit and raise deep bass sound level by adding volume. Finally, the stepped tapering of the line could be tweaked (middle divider lowered fractionally) to be more "geometric", though this might not be audible. Better material (hardwood or good ply) will lessen the work needed to pad the driver-chamber surfaces for max-clarity sound (MDF untreated imparts "warmth"). Those are my initial thoughts. Your kind words are appreciated.

(Choice of amplification of course matters a bit.)

 

[techsi]

Greets!

??? It's not a field coil and didn't see it as an option except the EXC series.

Yes, the 215 Heritage EXC is a field-coil driver actually! Their website is wrong by reporting the same specs of the non-EXC version for the EXC version of the 215 Heritage!

 

[techsi]

Heart takes a leap seeing those beauties. The 215 heritage (bi) cone itself should be very similar to mine. I don't have first-hand knowledge of field coils though, especially how to lower Qts to near ~0.3 like the vintage permanent magnet 215. Since you are not constrained by a ready-made cabinet, I would suggest a slightly wider and deeper box, so there's more space for the field coil, and the first divider board tilted more to redirect reflection. Greater width should lower baffle-step a bit and raise deep bass sound level by adding volume. Finally, the stepped tapering of the line could be tweaked (middle divider lowered fractionally) to be more "geometric", though this might not be audible. Better material (hardwood or good ply) will lessen the work needed to pad the driver-chamber surfaces for max-clarity sound (MDF untreated imparts "warmth"). Those are my initial thoughts. Your kind words are appreciated.

(Choice of amplification of course matters a bit.)

Dear wchang, the encounter with your TLonken thread was enlightening for me. In fact, the use of the 'usual' TQWT for this project of mine based on the Supravox 215 Heritage EXC did not convince me all that much because of the 'comb' response in the mid-low range (it seems to have unresolvable pressure holes that I have experienced myself) and because of the dimensions in height, which are in any case important. Your solution is compact and as I understand it gives a more linear response in the mid-bass and seems to be perfect for my idea of complementing the bass by means of a dedicated module on which the "TLonken Heritage EXC" will rest.

The bass module will turn the system into a floorstanding system and will be equipped with two new 38cm Supravox 380-2000 EXCs (yes, they are Field-Coils too!) in push-push configuration opposed on the sides of the cabinet and mechanically linked by threaded rods. Closed enclosure. The bass module will be the only one to be filtered to unobtrusively cross with the 'TLonken Heritage EXC' module above (a crossover point at 80 or 85Hz would also be good in my opinion).

Both the operating voltage of the 215's magnet and the voltage of the two 380's will be independently adjustable in order to find the preferred solution.

Bringing the 215 to a Qts of 0.30 will be easy, it seems, by sending a voltage of 12 to 13 volts (on the extreme high side) to its magnet!

Following your invaluable advice I am thinking of increasing the width of your original TLonken to 30cm and proportionally enlarging the depth and height. I will use 2.4cm birch plywood and try to calculate exactly the proportional increase in internal dimensions according to your design.

If I understand correctly, I will have to angle the first partition board just behind the driver a bit more and leave everything else untouched (I don't think I will be able to apply fine modifications to your design!).

What do you think? Could it work?

Now I first have to find the funds (and the courage) to buy the 6 very expensive electro-excited drivers!

This is the link to the Supravox 380-2000 EXC (specs at 12 volts to the elctromagnet) :

https://www.supravox.fr/products/copie-de-380-2000-lc-haut-parleur-de-grave-95db-8-ohms-1

 

[GM]

30-500 kHz! Looks suspiciously like an early 515.........

 

[techsi]

Ahah! They're terrible at the specs!

 

[wchang]

Following your invaluable advice I am thinking of increasing the width of your original TLonken to 30cm and proportionally enlarging the depth and height. I will use 2.4cm birch plywood and try to calculate exactly the proportional increase in internal dimensions according to your design.

If I understand correctly, I will have to angle the first partition board just behind the driver a bit more and leave everything else untouched (I don't think I will be able to apply fine modifications to your design!).

What do you think? Could it work?

Now I first have to find the funds (and the courage) to buy the 6 very expensive electro-excited drivers!

More drool... As you double the volume (each dimension ~5/4) and lengthen the line (~5/4 proportionally lowering the quarterwave frequency), there may well be enough boost to the low-end to not need those very expensive subs. Especially given the adjustability of your intended field coil (relax on the Qts a bit due to cab volume) and amplification means.

 

[techsi]

wchang, your words are very encouraging and certainly the first part of the project will be to build the 5/4 enlarged TLonken for the 215 Heritage EXC. I will post the result in the coming months when I have finished!

Only then will I decide whether or not it is worth spending a really large amount of money to build the electro-excited sub module.

For now, I thank you enormously for your help and inspiration.

Gianluca
Italy