Legis' Horny Tales

I presume the diffraction slot in the wavequide is the one responsible 90% of the widening of the radiation, not the small port/throat.

What I suspect to be the main reason of the increased SPL, is that the (quite large) front chamber and the very small hole/port (which enters the wavequide) might have created an acoutical low pass filter and also have limited the SPL output by being so small. They gained most SPL at 20kHz (+6dB) and somewhat less (+4dB) at 10kHz and around 3dB below that.

I did not measure the polars response prior to modding but we can trust JBL's PDF. In after-measurements they are still very "CD". I'd say pretty close to the 2404H polars:
Legis,

It certainly appears like the before and after polars are very similar.

It seems you have got "something for nothing" (other than a lot of work!) but it also always seems that "there is no free lunch".

Since the dispersion is unchanged, that leaves another possibility of increased sensitivity, distortion.

Even if you did not measure distortion before, I'd be curious as to what it is now, at levels up to a watt or two.

Art
 
Legis,

It certainly appears like the before and after polars are very similar.

It seems you have got "something for nothing" (other than a lot of work!) but it also always seems that "there is no free lunch".

Since the dispersion is unchanged, that leaves another possibility of increased sensitivity, distortion.

Even if you did not measure distortion before, I'd be curious as to what it is now, at levels up to a watt or two.

Art

I did also measure the distortion with REW sweeps.

Since the sensitivity is higher the motor and suspension induced distiortion will also be higher at reference power just because the output is greater, but not in same reference level in decibels. At same level the distortion was pretty much the same. Some areas even got better. Also now the compression ratio is much lower than before (very low, propably 2:1) so the driver will have less H2 at the same decibel level (given high enough level) due to smaller levels or air nonlinearization generated 2nd harmonic. I naturally could not measure that difference since my measurement levels were quite low (for a pro driver anyway), but at some point it should happen that the modded would show less H2 especially at high freqs since the air non-linearization effects are stronger at high frequencies I believe.

The reqion where there was the most negative difference was at the drivers Fs (at 1,5khz as modded). It got relatively louder (and dropped in frequency) due to less air spring damping the resonance. As you know the distortion usually gets wild in metal suspension comp driver's Fs. It does not matter since the driver is supposed to be crossed over >7kHz and definitely sounds best only up high crossed at least 2nd order, preferebly 3rd order. I use them at ~10khz/3rd order.
 
Last edited:
Looks good.

Did you do a write-up about your Synergy builds? What drivers are you using?

Hi bwaslo! I have not done in-depth writing of them but there was a so-called build thread, which covers some details: http://www.diyaudio.com/forums/subwoofers/258706-study-dipole-cardioid-bass-horn.html

The drivers I'm using (in synergy) are JBL 2446J with Truextent Beryllium 'phragms and modded Eminence Deltalite 2515 (2x per synergy). They have their baskets damped with bitumen, removed dustcaps and slitted and Pva glue treated cones like in Scanspeak Revelators. The slits&cuts I did just a week ago. :)

That measurement was taken at 15cm so the reqion above 1kHz is little down because of that (balances out when measured further away). Very good behaving cone resonance region.

An externally hosted image should be here but it was not working when we last tested it.


An externally hosted image should be here but it was not working when we last tested it.


Other drivers I'm using are Fountek Neo Pro 5i (will change the ribbons to 5µm/2,5µm) and JBL 2404H as super tweeters and BMS 18" woofers (4x 18N860, 2x18S430v2 and 2x18N850) in the 33Hz and ~14-16Hz (tuning changes with different loading) tapped horns in psuh-pull configurations. Neo Pro 5i is a direct radiating super tweeter and 2404H's are ambience super tweeters to balance the power response, or something like that. :)

Every bass channel/band uses only acoustical crossover/filtering (even tapped horns), no electrical filtering at all, a detail which I'm quite proud of. I have also tuned horn's sensitivities so that they do not basically need electrical level adjustments in single amped confiq, although the room and placemet also affects that.

Even though there are 6 separate bands per channel, out of which 5 are also electrically separate, parallel bands, the speakers are designed to be tube friendly even single amped. All woofers are connected in series and minum impedance is around 5,5 ohms. Whole system can be single amped with a non-feedback SET amplifier, which I like to use. The frequency response is also not very output impedance dependent becauser the impedance plot is quite flat from all the parallel bands which damp the impedance peaks automatically. Also the acoustical filtering of the bass channels (mainly the tapped horn) damps the resonance peaks.

System is basically extended 2-way even though it's 6-way, since the Synergy dominates everything and covers ~8,5 octaves out of the 10 (~50-60Hz - 18,5khz in my room). The system sounds very "same" with just the synergies playing, everything else just extends the response in both extremes. And they do quite alot, those extremes. :)

edit. On a second read, I wrote quite much information.:rolleyes::D
 
Last edited:
Founder of XSA-Labs
Joined 2012
Paid Member
removed dustcaps and slitted cones like in Scanspeak Revelators. The slits&cuts I did just a week ago!

You are a brave man! I tried slitting a much less expensive $18 PA130-8 and found not much of a difference - maybe its all in the adhesive used to re-glue the slits together. Do you have before and after measurements?

Here was my attempt, which ended up being a custom coaxial:
http://www.diyaudio.com/forums/full-range/279043-daytonator-pa130-8x-4.html
 
You are a brave man! I tried slitting a much less expensive $18 PA130-8 and found not much of a difference - maybe its all in the adhesive used to re-glue the slits together. Do you have before and after measurements?

Here was my attempt, which ended up being a custom coaxial:
http://www.diyaudio.com/forums/full-range/279043-daytonator-pa130-8x-4.html

Hi X, thanks I had not seen that. We got the same idea almost the same time :D. I had been thinking this particular modification for some time, weighting risks and possible benefits etc.

The measured differences were quite small, although there was some. But the biggest problem with measuring was how to get repeatable results. I found it to be quite a challenge. I don't have any test baffle which makes measuring such things repeatably quite hard. :)

Listening revealed that the slitted cone has better downward dynamics. Slitted cone was/is "quieter", ie. has blacker background, and was more clear. The difference was audible in both as a direct radiator and horn loaded. Measuring differences in spectrum decay was again quite hard and moving the mic a little resulted in larger changes than between drivers.

I would still need aluminum phase plugs for the Deltalites for them to be "ready". Aluminum phase plugs act as a shorting rings and reduce inductance and lower higher order harmonics by linearizing the Le(X).
 
Crazy world. I have been listening to JBL 2404Hs and ribbons all radiating directly to the listening spot. Sounds good, multiple tweeters seem to make the tweeter harder lo localize than a single. Solo Ribbon as direct super and 2404H's as ambience tweeters were more localizing that this "comb-filter setup" :D (which you can't hear).

An externally hosted image should be here but it was not working when we last tested it.



So i decided to start making super tweeter modules. The frame is made out of 2x4" :D. Front and back walls will be detachable and secured with M6 bolts.

An externally hosted image should be here but it was not working when we last tested it.



On a side note, Infected Mushroom's new album Converting Vegetarians 2 blew my mind. Massively atmospheric, layered and deep album, simply brilliant. Take a listen asap! :cheers:

converting-vegetarians-II.jpg
 
Last edited:
Updated view from listening position? :)

They are still without binding posts, better inside cabling and rear tweeters (some point in the future...) but are ready to play. :) I need to tilt them downwards to face the listener...

I like how they turned out and the fact that they make the cabling easier and visually more discreet.


An externally hosted image should be here but it was not working when we last tested it.
 
First of all Legis - thanks for your tip off on Infected Mushroom - downloaded it, and totally LOVE it! :D:D

Just a quick question regarding you 'ambiance tweeters' though. I was just wondering why after going to the trouble of building highly controlled directivity synergy horns, which have fantastic phase alignment, single acoustic source etc.

Why then fire high frequency ambiance - which presumably have a very different directivity and are from a distant point from the horns? Is this simply a preference or is there something I am missing?

Please excuse my ignorance - I have built a fair number of direct radiator systems which have very wide-band uniform directivity and am thinking about my next project being more narrow band -AKA horns/waveguides so this is a genuine question.

Thanks.
 
First of all Legis - thanks for your tip off on Infected Mushroom - downloaded it, and totally LOVE it! :D:D

Just a quick question regarding you 'ambiance tweeters' though. I was just wondering why after going to the trouble of building highly controlled directivity synergy horns, which have fantastic phase alignment, single acoustic source etc.

Why then fire high frequency ambiance - which presumably have a very different directivity and are from a distant point from the horns? Is this simply a preference or is there something I am missing?

Please excuse my ignorance - I have built a fair number of direct radiator systems which have very wide-band uniform directivity and am thinking about my next project being more narrow band -AKA horns/waveguides so this is a genuine question.

Thanks.

Hi bushmeister, IM did almost perfect album there for my taste. I'm totally hooked to it like to some drug. :D

Super tweeters are to (over) compensate the narrowing directivity of the synergy horns at the top octave. Ambience super tweeters, when the sit well in the system, really open up the imaging making some things in the image much larger, and make the top end sparkly and ethereal in general.

I have always approached sound production/creation (and reproduction, if that is even possible) in forms of both art and science. How you place the supertweeters in the system can make the system sound either "impressionistic" or accurate (if you turn them off completely) and all forms in between. I like the sound to be a bit of both, be accurate but still be very expansive and "larger than life", but not going over the board artistic. My dream sound would be something like a wall (or buble) of sound that has accurate imaging, but the image is big like a 300" projector screen and envelopes the listener also from the sides, and has depth in front of the listener.

The top end we hear in nature is always diffuse - this is also one thing to consider. Very accurate and non-diffuse top end sounds quite impressive for some time, but I have heard the best music production from systems that are not dead-dry sounding.

Thus best out of both worlds is my goal, hard to achieve, but sonically rewarding. :)
 
Last edited:
Thanks for clarification! There is always that blend of science and art in speaker design.

My speakers are very wide dispersion controlled directivity, with very flat power response - four way design with low crossover points (sub to 80, 8" woofer to 280, 4" mid to 2000) so they never beam. I do love the enveloping sound wall they provide, but have started wondering about narrow dispersion controlled directivity speakers......

I think I need to listen to some before taking the plunge.

PS - I can't stop playing the album either, I have it in my car, on my main stereo, on my i-pod......
 
Did some "wrong-measured" :D polars with the synergies. I did both mouth measurements (horizontal and vertical) and "far field" measurement at 160cm (little over 5 feet). Only horizontals with far field measurement...

Mouth measurements cover only the horn flare's angle (~28deg, the horns open at approx 55 x 45deg). Far field measurement goes beyond the flare to approx 49-55deg.

Easier to explain with pictures...

Mic's ending postion of near field horizontal measurements:
DSC05068_zps2uxnpog7.jpg


Ending position of near field vertical measurements:
DSC05070_zps0nrwrgth.jpg


Ending position of far field horizontal measurements (used a chair as a mic stand as i don't own a proper one yet, lol):
DSC05073_zpsi3udqysc.jpg


Near field horizontal (0-28deg):
horizontal%20124%20500ms_zpsrh8btfmg.jpg


Near field vertical (0-28deg):
vertical%20124%20500ms_zpsodd36rld.jpg


Far field horizontal (0-49...55deg) windowed to 3ms:
horizontal%20160cm_zpsjbwafmem.jpg


Averaged response of far field measurements (I dunno if REW does it correctly):
hor%20average_zps5idmglz9.jpg


Horizontal measurements' step responses, just checking if the time coherency is still applying to all angles :D:
Steps%20horizontal_zpsh4hjclpj.jpg


Vertical measurements' step responses:
Vertical%20steps_zpsxazqgpza.jpg


Far field step responses:
Horizontal%20steps%20160cm_zpsv8ufoazx.jpg



I wish REW had a polar plot tool that could draw all sorts of directivity diaphragms...

Horns waisband a little at 500Hz-1,5khz and there are some "anomalies" at 2khz and 5khz meaning the strongers radiation appers at off-axis. I made a sonogram with ARTA at the start of the year, in which it can be seen clearer:

polars_212_10ms_2_13_zpsb397b375.png


I think it's somewhat normal to a conical horn (according to HR), but I also suspect the injection ports might exaggerate them by couple of dB's since the vertical polars are smoother. But they also have points where the off-axis radiation surpasses the 0deg, just not as big as horizontal polars. Or then it's the opening angle of the horn that influences them and narrower opening vertical behaves better with compression driver that has a small horn in itself (transition from the driver is less abrupt vertically than horizontally). Or something.

Conical3_zps1f446015.png


conical2_zps431a547d.png


Conical_zpsbd549032.png
 
Last edited:
My dream sound would be something like a wall (or buble) of sound that has accurate imaging, but the image is big like a 300" projector screen and envelopes the listener also from the sides, and has depth in front of the listener.

One word: diffusors http://www.diyaudio.com/forums/everything-else/269366-making-easy-diy-depot-sound-diffuser-panels-step-step.html

I have mine set to the sides and behind the listening seat, and fire the CD waveguides more or less into them (which also crossfires them relative to the seat). Gets me both imaging and envelopment.

BTW, if your REW files can be changed to FRD format, you can use the OmniMic software (even without having a microphone, free) to generate your polar plots from them.
Dayton Audio Dayton Audio OmniMic V2
 
This might not be the best thread for passive charge coupled (CC) crossover discussion but here goes anyway.

I have been theory crafting a CC crossover a little.

1) The best type of capacitor for a CC crossover?

As you know, many types of higher end capacitortos like Mundorf Supreme, Audyn Plus, Jantzen Superiors etc are made from two capacitor connected in series to get higher voltage rating and low inductance. A picture from Wima MPK10 pdf clarifies the basic construction types from single cap to three caps in series:

An externally hosted image should be here but it was not working when we last tested it.


Such a multi "winding" capacitor could also be represented like this, two caps in series.

super-cap.png


When this kind of capacitors are used in CC network, the effect is - I believe - that only the other half of the capacitor gets biased and other half won't (even from static electricity). This might be partly the reason why some "inferior" caps sound better in CC network than the other, higher costing one. But of course not all high end caps are with multi windings.

So the best capacitor for CC would be, according to my "theory", a very good capacitor that is made out of single cap. What a revelation! ;) Like the top most one in the Wima picture. Naturally the capacitor should also have good sound to it before charge coupling. :)

2) CC XO turned into a "balanced CC crossover" almost for free

Balanced crossover networks are very rarely used. I have used them myself in the past and have found them to be good. Whenever you put a current through a coil or a transformer primary, and take it out from the other side, it is not the same anymore as the input. This warrants that balanced crossovers could have some audible benefit to them by "balancind the drive". I don't think it matters whether the amplifier is balanced or not.

Normally balanced xo reguires much more investment in the crossover components than asymmetrical crossover. But this is not the case with CC crossover, where you already have doubled the capacitors! You can convert a CC xo into balanced CC xo almost directly.

In the picture there are three different filtering schemes, that all give the same mathematical transfer function with a given driver. At the bottom is normal asymmetrical xo, in the middle is normal asymmetrical CC xo and at the top is balanced/symmetrical CC xo.

An externally hosted image should be here but it was not working when we last tested it.


When CC is used in balanced network, the DC is automatically blocked in amplifier's direction and this might have some benefit. Normally CC is used only to bias the capacitors. In shown balanced CC xo, with high shelving filter, one could also bias the whole filter's + voice coil's and speaker wires' dielectrics. I believe this would negate zero crossing in all the dielectrics (not just caps) if the bias voltage is higher than the drive voltage peak amplitude. Dielectric would also have much less impact on sound since they would be saturated all the time from electrons and the current that passes by when driving the speaker would not "excite" the dielectric. In short the electrons would not move in or out from the dielectric like they normally would, in non-biased conditions.

The second 3.6µF capacitors in my picture would not be "normally biased" since the same DC would affect both their sides. I believe this would not give them the "squeeze" effect that normal biasing does (and reduces microphonics), but instead the dielectrics would be even more non-reactive than in normal biasing.

Can the CC crossover be done this way that it biases the whole xo, speaker cables and voice coil's dielectrics, that is another question. I don't see harm done floating the crossover and speaker above the ground. But I'm a layman what comes to understanding electricity.

Does anybody some insight on discussed matters? Single winding cap recommendations are also welcome!
 
Last edited:
Ordered the caps for "balanced charge coupled" XO's. We'll see if they sound better than single ended xo with boutique caps (same caps will be used in CC so the tonality does not change too much, maybe :eek:).

Also upgraded my DAC to Teac UD-503 and the transistor setup (alternative to tubes) to Denon PMA-2020AE. The mechanical construction of Denon (of both, actually) is straight from the good ol' days. They have really seen some effort in this flagship model amp in making every little detail right imo, especially in the price point. PCB's are populated widely with Elna Silmic II and small polypropylene caps, there is only one mosfet output transistor pair per channel, dual mono construction, clever mechanics regarding internal resonance control and electrical field control/shielding. Both the transformers are "floating" (risen above the bottom and coupled lossy) so they won't vibrate other components, amplifier boards and heatsinks are floating/lossy coupled as well.

This is one very fine sounding amplifier, sound has good drive and feel of torgue and power, inviting tone (slightly sweet and romantic just as I like it) and the picture is "3D wall of sound", expansive in every direction.

An externally hosted image should be here but it was not working when we last tested it.


Some pics from the net:

An externally hosted image should be here but it was not working when we last tested it.


2020-TR.jpg




An externally hosted image should be here but it was not working when we last tested it.




The amp would not be anything this good without the Teac DAC, which is giving it the sweetest and purest signal to amplify. It can also do non over sampling output (NOS) when PCM output filter it set to OFF. Quite rare feat and I like NOS output very much and use it almost all the time. Check it out if you are not "fully DIY" ;).

44,1kHz/16bit NOS and 88,2khz/16bit NOS waveforms:

An externally hosted image should be here but it was not working when we last tested it.
 
Last edited: