Symmetrical charged coupled crossovers

[Orgi1]

Good Morning!

Here speaks Georg from Germany and i wanted to ask, if anybody have some experiences with symmetrical charged coupled crossovers.
Last Month i have done mi first cc-crossover, and the way from ccc to sccc is not long...


Regards to a very interesting Forum!
Georg

[Orgi1]

Hi people!
Today I feed the discussion with a little food. The example of the midrange branch of JBL4345 I show what I mean. Has anyone ever seen something like this? A Balanced crossover sounds even better, and does not need a capacitor more than the CCC switch. Only inductors and resistors are required in slightly higher numbers. So if we already build a battery-charged crossover, then this can also be practically the same structure symmetrical with battery too.

6167761678

[fpitas]

Balanced circuitry is better at rejecting noise, but what advantage does it offer here? Lots of extra parts involved. Is there distortion cancellation, such as in push-pull stages?

[Orgi1]

Mathematically, there are no differences at all. Sonically, however, they are there - between the normal crossover and the symmetric there are differences. But my idea was a special: if we already use the CC-Crossover, THEN is the difference to the symmetrical-CCC not a big one.

[Orgi1]

Balanced circuitry is better at rejecting noise, but what advantage does it offer here? Lots of extra parts involved. Is there distortion cancellation, such as in push-pull stages?

I have absolutely no idea what the technical foundation is that speakers sounds better with symmetrical crossover. If anyone can help, please let us know what could be the basis for this circumstances.


And pls. My english not a good one - so if anyone find's a mistake, please make a note! Thanks!

[fpitas]

Your English is perfectly understandable so far.

Distortion measurements on each circuit would help understand why there might be an audible difference. I'll state the obvious, that simulations often don't tell the complete story.

[Orgi1]

PSpice is a program with which I draw circuit diagrams.
The fact that it converts the frequency responses congruent, gives me the confidence that I have the circuit properly implemented by asymmetrically to balanced.

You are totally right when you say that a simulation does not show all circumstances. I do rework speaker for the last 10 years. Our Team and I hear daily of the developed results, so that we can say very good if there are changes. However, unfortunately Listening clarifies not, in what way or why something has changed :banghead:.

[just4kinks]

Have you run a transient analysis? I would like to see the FFT. I suspect the improvement will not be as dramatic as the step up to CC, but I'm interested.

[Orgi1]

The simulation (transient or frequency response) does not bring any insights as only ideal components are present. The losses in the components are not mathematically writable.

So we only have the possibility to listen to the Speakers with S-CCC or with CCC. Only this is to be used for the results. And now, at this time, we are at the same place of my first Thread.... 'are there any experiences about...'
:(

[Orgi1]

Has anyone ever thought about to replace the battery against any other voltage generation? A high voltage transformer could help. There need not be 5000 volts yes, 30 would be enough. Which can be realized with high impedance, so that it may not change the music signal from the amplifier. Then you do not change a battery when it should ever be empty.


What do you think about?

[fpitas]

There might be an improvement from the higher voltage, although you could achieve that also by placing a few batteries in series. There's essentially no current draw, so the batteries will last for their shelf life. In their defense, batteries are very quiet electrically, simple, and inexpensive.

[Earl K]

Hi Georg,

The very first version of Charge Coupling ( I believe was used in the S9500 ) & was 18V ( using 2 , 9V batteries ) .

The latest development ( applied in a few of the horn circuits ) is to apply a much lower voltage than the stock 9 volts .

This lower voltage is derived from the movement of the woofer . A diode ( & an isolation resistor ) form the circuit for this style of DC biasing ( the diode allows voltage in / but not out ) . I've never tried it / but ( IME ) it may sound nicer that the stock circuit .

I myself, prefer the sound of DC biased caps after the constant drive ( battery ) has been removed from the circuit ( allowing the caps holding the DC charge to "float" above the reference ground / it's a subtle, but important difference to my ears ) . I typically use around only 2-3 volts .

:)

[Orgi1]

Good morning Earl!
This is an interesting idea - i will try it an report here!
Thank you!

[Orgi1]

Hi folks!
Today i will report a little bit of my experiences to the ccc-networks "without" a battery.

I have build a network like "ordinary" ccc-networks (symmetric-ccc's we will do later). Then i used a high-voltage-cascode, (http://de.wikipedia.org/wiki/Hochspannungskaskade) for generating voltages up to 25 - 35Volt from the Musicpower of the amplifier. No external Battery or Voltage-Source are used.

Then i hooked up a voltage-display, so i can "listen and see", what amount of voltage we get with normal music (not very loud, in German we say "Zimmerlautstärke").

First seconds of music the voltage is below one Volt. Than, up to a minute, the voltage increases up to 1 - 1.5 Volt. Below this mark the Sound is more smooth than without any voltage, but above one Volt the music is changing from smooth to a little more crisp, but without loosing the smoothnes. Then, during 30 minutes of listening to the Music, the voltage rises up to more than 10 Volt. Above 13 Volt the music-performance also change a little bit. Then the clarity and perfection of the reproduction gets more "live"-feeling. Below this 12-13 Volts the Music ist very bright and clear and dry, but above all of the positiv properties retained, but the sound is more live, more plausible.

I would be very interested in your Experiences! Please report, if you want to kindly share your experiences to me and all the others here.
thanks!
Georg

[Orgi1]

Nobody is interested in this fascinating subject? The CCC-user would listen to the music and not reading to find out how to improve the sound of their speakers yet, right?

[hjames]

Nobody is interested in this fascinating subject? The CCC-user would listen to the music and not reading to find out how to improve the sound of their speakers yet, right?

Sounds interesting, tho you may be the only person so far to have built such a crossover.
If the character of the sound changes as your caps build up a charge, that may be something folks would consider a flaw,
as opposed to CC crossovers that are charged up a single time by the battery and stay biased(?) until the battery goes flat or is removed.

Could be very subtle differences between the two, I would imagine.

No flames or snark implied or meant (!) - Keep us informed!

[ivica]

Nobody is interested in this fascinating subject? The CCC-user would listen to the music and not reading to find out how to improve the sound of their speakers yet, right?

Hi Orgi1,

It would be interesting to "measure" the capacitor value applying mentioned DC voltage (You have metoned in the previous post) on it.
May be 'under polarized condition' the capacitor would change it capacitance, and that would be the reason for the different sonic character depending of the DC voltage.....


Regards
ivica

[Ian Mackenzie]

Hi Georg,

What manufacture and model capacitor are you using?

The reason is JBL use Solen Fast Caps.

These capacitors have a known audible characteristic and the JBL approach for charge couple was made specific this capacitor.

Basically the DC bias voltage creates a tension on the capacitor windings and reduces "self" noise.


More recently JBL has used a zenor diode for a voltage source so it track the A/C voltage.

Like anything I have no doubt there is a sweet spot.

My thoughts are have customised DC voltage for Mid, HF and UHF that are scaled to the A/C drive voltages in pass band.


Your ideas are in use on current SOA amplifiers like Pass Labs where balanced symmetry is used to help cancel common mode distortion.

So your ideas have merit but perhaps you need a true balanced amplifier to do further test.

If you make your project as a model and contact Pass Labs DIY department they may well have a look at this idea and offer feedback. Refer to Kent English

Mr Pass is a well known advocate of DIY audio.





Ian

[Orgi1]

Hi Ian,
Thanks for the suggestions. We use Mundorf MCAP250/MACP400, these are also very good sounding capacitors. I'm going to order Solen Fastcaps, then I can try them there also.

To track the voltage by zenerdiodes is an interesting idea. This will I try and give feedback!

In the past I built many amplifiers by Nelson. These sounds great! (Speaker is Avantgarde-Acoustic Duo). Today, I use the amplifier by Gerd Sauermann from Germany, the best class-A amplifier i know, completely symmetrical. (www.sauermann-audio.eu)

Hi ivica,
the capacity varies only very slightly when a DC voltage is applied. For very small capacities in the Picofarad range up to a few Nanofarad, this is decisively - but not in loudspeaker constructions, where we use caps between 1uF and 47 uF. There is no change by adding DC.

[Orgi1]

Good Morning!

Now we have build again a pair of speakers with the symmetrical CCC-Network. It sounds very clear and absolutely free of any distortions in the sound. The reproduction of the room and the detailed location of the instruments is really stunning!

This little Speaker uses a wavecore-tweeter with 32 mm voice-coil (without ferrofluid) and a 17 cm Woofer made by vifa. The housing is a 15 liter bassreflex-box, and for best results we use WBT NextGen-0708-Terminals.

You will never belive how it sounds, if you not hear this!!!!!!

If we have finished the chrossover-board i will show you pics of them!

[Orgi1]

Now we have got the new coils from Jensen, Danmark.

This are beautiful and awesome! They ave two windings of one body, so the inductance of each winding is to double and add to the other winding: Two windings with 0,22 mH will give us the same effect like one winding with 0,88 mH. This ist cool (but mathematical correct).

Also we have got new capacitors from Jensen: Metallised Polypropylen in Oil - very good sounding but not so expensive like the paper-in-oil-alu or copper windings.
Both are the perfect parts for ccc-networks or symmetrical ccc-networks (we call it "s-ccc").
62784
So, please stay tuned for detailed description of the sound, if we have finished the new crossovers and played a little bit (for "burning in").

Georg from Germany

[Orgi1]

Today i will report to a new chapter of symmetric crossover-networks.
First we use Mundorf-Capacitors and we are lucky.
Then we think about it and add the fantastic Jensen Silver Paper in Oil Caps. Heavy $$, but a never heard sound quality.
Then we switch to symmetric networks with this fantastic parts i have told you from.

Now - at winter holidays - i got a cheap quadral-Box (called ASCENT 250) and got an idea: The efficiency of the tweeter is more than 10 dB higher then the efficiency of the woofer (diameter 13 cm). So i remove the caps to the tweeter....

Have a look at the circuit:
64303
The coil is the frequency-depdending resistor, and the both resistors with 2.7 ohm are the voltage-divider. This will work!
The Woofer of the little speaker has got an ordinary 6 dB coil.
64302
I have added the 0-dB-scale, so you can see the result for the tweeter. By the way: The coil in parallel to the tweeter needs to be very low impedance at low frequencies. I use there a "AWG14-copper-foil-coil".

The sound of this speaker is extremely detailed. The instruments are very "chrisp" and clear, the location of the instruments is extraordinary.
If you build such crossovers, you will be able to recognize, why we are state a new sentence: "it is nearly unbelievable, what informations get lost in (bad) crossover-networks".


Problem of this circuit: You can only use this, when the tweeter are much louder then the woofer (>10 dB).

Have a nice Day!

(R5 is needed for the simulation only)

[Orgi1]

Today i will report a new idea for generating the current for charging the capacitors.


The first idea was used at serveral Companies: A "High-Voltage-Cascade" (google: villard-cascade) with Diodes and Caps. This will work fine, because only very little energy is needed to fully charge the caps.


But we also have the possibility to use a Transformator-Effect. If you place a coil direct on top of an other coil, you can use the magnetic induction to put a little bit magnetism into this other coil.


If you have e.g. a big coil for the woofer in your crossover-network, you can put a e.g. 1mH-Coil with 0.71 mm copper-wire on top. Because of very little current you have to use diodes for rectifier with low drop-out voltage (shottky-diode), and then put an dc-dc-converter to step up the voltage up to e.g. 20 or 30 volt. This voltage is absolutely enough for charging your caps. The millihenry of the woofer-coil does not dramatically change, you do not have to worry about this.


Batteries are sometimes empty - especially the cheap chinees cans. But with some clever ideas you do not have to worry about your battery or somthing useless more in your speakers :-)


Does anybody have similair experiences in doing this?



Greetings from Germany,
Georg

[Legis]

Hi Orgi1, extremely interesting subject this balanced/symmetrical crossvers!

I'm also building symmetrical charge coupled crossovers for my horn setup. Here's the comp driver's xo. I also use electrical filtering for the super tweeters, but all three bass channels are acoustic filtered only, so they won't need CC'ing :). (Acoustic filtering would deserve another topic, also quite unknown/less used xo type.)

http://i817.photobucket.com/albums/zz95/LegisActio/Synergy%20Horn/Passive%20filtering/CC_zpsocooiylu.png

http://i817.photobucket.com/albums/zz95/LegisActio/Synergy%20Horn/Passive%20filtering/BCC8_zpstfmqbvlm.jpg


Maybe balanced xo cancels even order distortion that are caused in the xo components and voice coil's "electrical distortion" (it is also a coil, and higly nonlinear one), and this is also the reason why it does not show up in simulations because they do not simulate component nonlinearities? Maybe balanced xo also rejects outside noise better cancelling common mode noise?

I think balanced XO's are on to something based on my first experiences.

[Orgi1]

Hi Legis,
nice job you have done there!

My suggestion is: Change the "Audyn references" into "Jupiter copper wax" cap's - they are sounding much more natural than the audyn cap reference. We have had also the reference here, but we do not use them.

The Reference we will describe as a very bright and ultra detailed, but not harmonic and to much nervous. This Cap is useful to "hurry up" a system, with e.g. a british sound, a system with not much details, with dark and slow and clumsy sound.
But in a system with a very good tweeter, with a symmetrical, charged crossover, the copper-wax-jupiter (0nly 0.047 uF is needet) will bring you a very natrural but also detailed sound. I combine the clarity-cap with the jupiter-copper-cap and i also use the beeswax-copper-foil-coil.
You also can use russian paper in oil caps in a blue ceramic housing - not the k75-10, but this blue housing caps are alu-oil-paper-caps and they also sounds great. They are cheap but i do not know, if you can buy caps from russia in usa.

You can think of the following: charging the caps will bring you a clean sound - but symmetric chrossover is also a good thing. So if you want to charge the c9,c10,c11,c12 you can wire it in series, and place the resistor not wire by wire to the caps, but wire from plus of one cap to minus of the next cap - so you can charge the caps instead (by loosing the symmetry).

Also interesting: The EVO-Serie from Mundorf does not have the double-cap inside. So if you build a symmetric chrossover, it will be enough to use the evos, and place one of them from plus to minus, and the other from minus to plus for the same effect. This will reduce costs dramatically, because you will not need the expensive supreme-serie.


Regards,
Georg

[Legis]

Hi Orgi1, I have noticed that Audyn Reference does not sound right at the start and they need pretty much of burn in. The sound is too thin and slightly hard and aggressive top end at first but it mellows out and becomes whole after ~100-200 hours of listening. They are my favourite cap out of Clarity MR and Mundorf Supreme. Mundorf supremes on the contrary work "ok" right from the start but they also get better after some time.

I find Reference resebling the tonal neutrality of Clarity MR with the finely nuanced presentation like Supremes. They are also mechanically damped due to litz legs, they don't "ring" when given a knock with a finger. Caps with thick solid core legs tend to ring more (their legs vibrate).

I also noted that the CC'ing can make the sound harder at first compared to non-CC (with fresh caps, and 9V), the caps and voltages need time to soak in and find equillibrium perhaps? This CC is new thing to me, I'm not very familiar how it changes the sound yet, what bias voltage is optimal etc...

I don't have experience with the Mundorf EVO series but I put in the New Supreme EVO Oil in the balanced CC's. They seem to be nice like the basic Supremes.

I believe the "dual winding" caps or "dual caps" need to be polarized from both legs to work optimally. This way the polarity difference happens between the legs and the central winding.

http://i817.photobucket.com/albums/zz95/LegisActio/Synergy%20Horn/Passive%20filtering/Central%20winding_zpsvmw133u3.png

Avantgarde does this exactly the other way around and polarizes the central winding of dual caps. Their caps have 3rd leadout which connects to the central winding of the caps, where the bias is added.

http://www.6moons.com/audioreviews/duomega/hero_tweeterboard.jpg

I do it the other way around and apply the bias to both legs where it's possible. Thus my crossovers polarize both sides of the caps (except the very first big caps, only one of their legs are polarized) and also all speaker cables, crossover coils and even the voice coil is submerged in DC, floating above ground. This does not offset the coil since no direct current is going through the voice coil. The whole xo on the other side of the first caps is floating, polarizing every possible insulator in the crossover. I believe this also happens then: http://www.audioquest.com/resource_tools/LearningMods/DBS.pdf

You should try it out and see what you think!