Value of charge coupling 3105 network??

[scorpio]

I'm preparing to build a 3105 like network to install instead of the simple series capacitor I'm running now between a 2450 and a 2405, I'm going to use by-pass capacitors out of curiosity, since it's dead simple to implement on this network. While I'm at it, I was wondering about the value of charge coupling such a simple system, has anyone tried this and would care to comment??

I'm using a 5235 crossover between the woofer and the mid/high, so I have no other passive component in the system.

Thanks for any comment - and by the way, happy new year to all,

Cheers

[Robh3606]

Perfect test case for you. I would try bypass caps, CC and potentialy a high quality audiophille capacitor type like Clarity's. See what you like and how much of a diference it make in your set-up. I CC all my networks but that's my choice.

Rob:)

[Mr. Widget]

Perfect test case for you. I would try bypass caps, CC and potentialy a high quality audiophille capacitor type like Clarity's. See what you like......and report back to us.:D

I've been curious about just such a comparison... tho not enough to pull a Zilch and mock up 16 designs.;)


Widget

[Baron030]

Hi Scorpio

I have found that bypassing and charge coupling did make a very noticeable difference in my own system. So, I would recommend that you try it with the 3105 network. Since, these capacitor values are small; the additional cost is not going to be too much.

Could you describe your system in more detail?
It sounds like it’s a bi-amplified 3 way system.
Which horn are you using with the 2450?

Depending on the 2450's horn efficiency, you may need to alter the 3105 network and move the L-Pad from the high pass section to the low pass section of the 3105 network. This will reduce the 2450's efficiency to better match that of the 2405s.

Also, worth noting is that the 3105 network does not offer any low frequency protection to the 2450 driver. So, I would assume that you are planning on using a high pass protection capacitor between amplifier and the 3105 network.
In which case, bypassing and charge coupling the protection cap will also make a world of difference.

Baron030 :)

[scorpio]

Hi Baron030,

The 2450 are for the moment on circular horns with a profile close to tractrix (Le Cleach'), I've got some pics in this:

http://audioheritage.org/vbulletin/showthread.php?t=18475&page=2

I've just started building a set of radial 300Hz a little similar to the TAD TH-4001, but that will take a little while to complete. The 2450 have alu 2441 diaphragms in them, so I'd guess the standard values of the 3105 should work OK.

Right now, I have an L-Pad in the 2450 circuit, just like you suggested, and will start that way in the 3105 'look alike' too. I'm using a bundle of capacitors coming to about 57µF to protect the 2450, good suggestion to by-pass those as well, didn't think of it.

OK, I may try to go through all the steps, but be patient guys, as I've mentionned here before, I'm not the fastest in making changes, and I like to take my time to assess the results (unless obvious, I'm always worried about the better vs. different debate when making a change...). I'll go by first by-passing the protection caps, then adding the 'standard' 3105, then by-pass that one, and if all this keeps getting better, CC - phew... It should still be complete before I finish those radial horns though...

One question for my CC 101 training, the + of the battery goes in between the doubled up capacitors, but where the h**k goes the -, to the common ground? I didn't catch that in what I read - sorry if obvious.

[Chas]

One question for my CC 101 training, the + of the battery goes in between the doubled up capacitors, but where the h**k goes the -, to the common ground? I didn't catch that in what I read - sorry if obvious.

+ goes to one side of a 2.0-2.2 Meg resistor that is tied to the junction between the capacitors.
- goes to common.

[scorpio]

Thanks Chas,

[Baron030]

Hi Scorpio

Here is a schematic for a CC 3105 network.

Baron030 :)

[toddalin]

You guys need to pay more attention..., you're missing the big picture here.

Scorpio wants to run 2405 on a 3105. Big mistake! He should be building a CC 3106 network. Change the chokes to 0.3 mH.

[scorpio]

Hi Toddalin,
Is this the mod that pushes up the X-over frequency to >8k ? I've been trying to calculate the needed change from the stock values to do this, but could not make sense of it - probably the formula I have is wrong, did not get the right values for the components at all.

I take it that the value of the caps does not need changing? I'll dig the 3106 schematic from the library on the JBL site to check.

Thanks Baron030 for the developped schematic, at least I'm sure now I won't make mistakes! (besides the value of the choke...)

Cheers

[Zilch]

http://audioheritage.org/vbulletin/showthread.php?p=130246#post130246

[There's the "Cascade" debate, Todd.... :p ]

[Mr. Widget]

In which case, bypassing and charge coupling the protection cap will also make a world of difference.Really... how do you set up your protection caps?

On my TAD TD-4003 I am actively crossing over at 800Hz. I have a 47uF cap in series (first order ~425Hz @ 8 ohms) that I keep in the system for protection... I have listened with it in the circuit, out of the circuit, and even shorting it out of the circuit on the fly and have heard no contribution from it. It is a relatively inexpensive Bennic Polypropylene cap.

At $1850 a pop (sorry:D) for replacement diaphragms, I am not comfortable flying without a net.


Widget

[toddalin]

Hi Toddalin,
Is this the mod that pushes up the X-over frequency to >8k ? I've been trying to calculate the needed change from the stock values to do this, but could not make sense of it - probably the formula I have is wrong, did not get the right values for the components at all.

Cheers

According to the JBL schematics, this pushes it out to 8K. But the L300 and 4333 use a 1.0 mfd cap, followed by the choke (between about ~0.18 and ~0.27 mH, depending), followed by the 1.5 mfd cap to push it out further (~8.5 kHz) and create a "double peak." (Follow Zilches link regarding cascading for more info on this setup.)

BTW, when I tried this approach (i.e., 1.0/choke/1.5)with the 2402 (but using a 0.5 mH choke) I was less than satisfied, ended up with a peak at about 10K that overlapped a horn peak at 10K causing a massive peak at 10K (and much HF fatigue) and went back to the simple 1.5 mfd and 0.5 mH chokes (ala N7000/N8000), with a voltage divider to reduce its output and increase the slope. I also replaced the 1.5 mfd low pass on the mid with 2.5 mfd to kill the 10K horn peak (LE175 on HL91). (From experience, I would now recommended somewhere between 2.0 and 2.2 mfd here.)

I tried this both cascaded and non-cascaded. Cascaded, I lost too much volume out of the 2402 (but would probably have beem alright if I left off the voltage divider). The simpler N7000/N8000 schematic with the voltage divider actually provided a slightly smoother transition frequency between the mid and high non-cascaded, so that's what I did.

[Baron030]

I am sorry for straying a little bit off of the 3105 network discussion. But, I needed to answer Mr. Widget's question about protection caps. I am currently using a two stage system for protecting my 2446H drivers.

Stage One: I have a relay box which sequentially connects and then disconnects my 2446H drivers from my bridged-mono Crown75a amplifiers to prevent any start up or shut down noises from reaching the drivers. The resistors pictured in the relay box are to maintain a constant load on the amplifier's outputs only and don't impact the sound in any way. For a while, the relay box allowed me to hard wire the 2446H drivers directly to the D75A's output transistors. But, one of my D75A's started generating a very low level of DC current. So, I decided that I needed to go back to using a protection caps a well.

Stage Two: Pictured below are four 82uF Solen caps wired is series parallel with 0.01uF Dayton Film & Foil bi-pass caps. And after the protection caps, I have added a 20 ohm damping resistor across the drivers voice coil. And the addition of a damping resistor made very noticeable improvement. The Clio impedance plots show how the damping resistor reduces the size of the horn’s resonance peaks.

Baron030:)

[scorpio]

One more questions on something I don't understand, why does the 3106 network have 20Ohms resistors in parallel with the drivers, when the 3105 does not?

I don't understand the intent of these.

Thanks!

[ratitifb]

and perhaps ... the 8/16 ohms JBL impedance labelling mistake :blink: along time ?

[Baron030]

Quoted by Scorpio:
One more questions on something I don't understand, why does the 3106 network have 20 Ohms resistors in parallel with the drivers, when the 3105 does not?
I don't understand the intent of these.
Thanks!

Hi Scorpio

The 3105 network has a load impedance of 16 ohms.
And the 3106 network has a load impedance of 8 ohms.

Assuming that the 3106 network was originally used with a 16 ohm “J” high frequency compression driver. The 20 ohm resistor in the low freq section was used to bring the load impedance down to 8 ohms. And the 20 ohm resistor in the high freq section does the same thing for any of the JBL UHF drivers, which all generally have an impedance of about 12 ohms.

For your own project, you could add 20 ohm resistors to the 3105 network and it would convert its load impedance from 16 ohms down to 8 ohms. This is assuming that you are using a 16 ohm driver in your project. And I would think that the crossover point would still remain at 5 KHz. But, without changing any of the other component values the crossover point attenuation might go higher then the published -3db. If you notice the 3106 network has a crossover point attenuation of -7db. So, adding a 20 ohm resistor might be a way to fine tune a 3105 network and remove a slight peak at the crossover point. This is assuming that should you find that to be a problem in your own project.

Remember crossover design is a much an art as it is a science.
So, don’t be surprised that you find yourself testing and tweaking things for sometime to come.

Baron030 :)

[scorpio]

So, don’t be surprised that you find yourself testing and tweaking things for sometime to come.

Baron030 :)

Yep, exactly what I'm concerned with...not what I prefer, but I really want to get these babies to sing their best.

[scorpio]

Hum, many months later, I think it's time to report back.

I almost did what I was talking about some posts higher in this thread. First thing I did was to replace the simple capacitor on the 2405 that I was using as 'crossover' so far with a standard 3106 network using reasonable but cheap wire selfs and Audyn PP caps. I found it not easy to sort our the relative levels of the 3 drivers in this new configuration but the overal result was positive, mostly by having a better defined top end, and less directive mid/highs, good progress in doing so, but not dramatic. For the less than 50 euros it costed me, shipping included, a positive change in any case.

Then I tried by-passing the caps in the HF section of the 3106 with 10nF polystyrene, no surprise for most of you here, yes, I agree about this providing additional 'air' to the top end, again an incremental improvement worth the very small outlay needed. This improvement I could best feel when later removing the by-pass, beyond the 'air' issue mentioned, back came some nasty sibilants on some records that were smoothed out with the by-pass.

Why remove the by-pass then? to change all the caps and try a CC version using some entry level MKP caps I'm sourcing locally, overall cost about 30 euros for 8*3.3 µF. Not sure about the brands, they look very much like Audyn, only red, not bearing a manufacturer name.

This has now been installed in the system for about 4 days, and it's definitely the best of the alternatives I tried so far in terms of clarity, lightness/naturalness (sorry, I'm basically a French speaker, so I might be inventing words along the way), dynamics. I have not had a chance yet to listen to very many records, but I also have the feeling that this helps make the system more alive and dynamic at lower playback level.

It's difficult to put this on a scale if you don't know what 10 represents, but I'd put the starting point at 4, using the 3106 network would bring it to 6, the by pass to 7 and the CC to 9, based on the drivers I have. I'm wondering now if using higher grade (e.g. more expensive) selfs and caps would bring it to 10, but that will have to wait for a while!

I've never been the good one to use adjectives to describe my listening impressions, and I'm certainly not a golden ear, but based on that experience, I would say that using CC approach on entry level (in an audiophile sense) components system is well worth the relatively small extra cost.

This sounds like a little wast of broad band, but I though that it was wort sharing. Thanks all who helped me getting where I am so far with my speakers!

[sekess]

Hey Baron030,

I'd like to comment on the following quote I took from your post (post#14):

"But, one of my D75A's started generating a very low level of DC current."

I had the same problem myself on a Crown DC300A. I adjusted the DC offset and all was perfect again. Don't know if you remedied this on your D75A. But if you didn't, I'd recommend you lok into it. It was a pretty easy adjustment on my DC300A. However, I don't know if it is the same on the D75A.

Good Luck,
Steve

[Baron030]

Hi Steve

The Crown D75A amplifier that had the small DC output current was purchased new and was under warranty. So, I sent it into a crown authorized service center and that problem was fixed under warranty.


While, I could go back to hard wiring the D75A outputs directly to the 2446 & 2405 driver's voice coils. I have decided that the risks just don't out weight the benefits. The large charge coupled protection caps that I am using don't seem have any detrimental effects on the sound.

Baron030 :)

[scorpio]

Well, the network has now been charging for one week, and although I am not one to normally use strong or bad language (at least in writing :blink:), it's one heck of a f******g improvement, low level detail retrieval is nothing short of amazing!

Besides, I made another very noticeable improvement in all aspects by finally doing one of the small thing that everybody suggests in terms of improving the room acoustics, I removed the low level coffee table that was standing between me and the speaker :applaud::applaud::applaud:

This hobby will never stop to amaze me,

[lgvenable]

Really... how do you set up your protection caps?

On my TAD TD-4003 I am actively crossing over at 800Hz. I have a 47uF cap in series (first order ~425Hz @ 8 ohms) that I keep in the system for protection... I have listened with it in the circuit, out of the circuit, and even shorting it out of the circuit on the fly and have heard no contribution from it. It is a relatively inexpensive Bennic Polypropylene cap.

At $1850 a pop (sorry:D) for replacement diaphragms, I am not comfortable flying without a net.


Widget
Mr Widget:
How did you setup your cap on the compression driver ("in series"), see the attached jpeg> like this?