Understanding crossover networks

[Kay Pirinha]

Hello,

I'm referring to the 4430's 3134 network schematics (http://www.jblproservice.com/pdf/Network%20Schematics/3134%20Network.pdf). The 4435's is rather similar. The more I look at it, the less I don't understand it, honestly:

There's a L102/C3 LP filter that drives the 2135H woofer. So far so good. But there's C8/R2, which resemble a Boucherot cell in an power amplifier's output, but with 100 times the capacitance. What does this series network do?

The HP section is even more complicated. Ok, the HP itself is C11/L2 that drive an L pad. This L pad may be arranged as usual, i.e. a 8 Ω rheostat in series with the tweeter and a about 42 Ω one (according to my own measurements) in parallel with it - which puzzles me, as the other way round should have to be assumed, in order to have a almost constant input impedance, hence a predictable rolloff frequency. In the given arrangement the tweeter sees an almost constant impedance. What's that good for? And why does the 42 Ω rheo give no output in it's fully ccw position?

Then there's a series resonant bypass, consisting of C9, L1 and the 30 Ω rheo and the voice coil's inductance and DC resistance. C9 and L1 alone would have a resonant frequency of 17.8 kHz. But there's the VC inductance. With my newly acquired De Ree DE-5000 LCR meter I measured my pair of 2421A's and got very different results between both - and surprisingly depending on the measurement frequency in the same unit. The readings were about 0.3 and 0.4 resp. @ 100, 120 and 1000 Hz, and 13.2 and 15.2 µH resp. @ 10 kHz. Ok, let's compute with the mean value of 14.2 µH and get a resonant frequency of 15.3 kHz. The Q(uality) factor depends on the rheo's plus the VC's DC resistances. This series resonant circuit bypasses the rheostat even in it's ccw position. But what's the use for the rheo to get open in fully ccw, as said in the underline?

Finally, what's the controls for mids and treble? I guess the rheo is the treble, and the L pad is the mid control. Am I right?

Best regards, and pardon me for causing you quite a headache ;)!

Edit, additional question: Does anyone know the DC resistance values of the involved inductances? Are they air cored?

[Kay Pirinha]

Ok, thank you for your remarkable participation so far :D!

So I have another question: The 3134 document says that the rheo is open when fully ccw (note 3). This means that the compression driver would completely be off if the L pad is fully ccw at the same time, and the 2235H rolls of at 1 kHz. This doesn' seen to be compatible with the frequency ranges given in Fig. 7 and Fig. 8 in this document (http://www.jblpro.com/pub/obsolete/443035.pdf) and doesn't make any sense at all. So, where's the crux?

Best regards!

[Ed Kreamer]

Are you building your own network? I looked at the schematic and and in an instant of brilliance and clarity, I knew I was out of my league.

[Kay Pirinha]

Yes, I'm about to do so, still collecting parts. I had hoped that at least someone could tell me which of the potentiometers is the mid and which is the brilliance control.

Some progress, though: Recently my DeRee DR-5000 LCR meter has arrived that allows me to modify stock inductors to the required values.

Best regands!

[turnitdown]

The 30 ohm pot addresses the mids. Good luck. I have built this network, it is not hard.

[Earl K]

Kay,

83898

The title of this thread is pretty misleading ( IMHO ).

You would ( most likely ) have received quicker response if you hadn't ( effectively telegraphed through the title ) that you wanted an education on the theory, design and implementation of passive networks.
- That's a multi-year learning process that isn't achievable from a handful of posts through any forum.

Had you simply announced that you were building a pair of 3134's and didn't understand which knob/attenuator did what ( & needed some clarification ), I'm confident you would have received more timely help.

:)

[Kay Pirinha]

Most probably you're right :o:!

Best regards!

[Earl K]

Kay,

If you want to further your education about how some parts of these networks work, then download XSim crossover simulator (http://libinst.com/Xsim/XSimSetup.exe) and use it to play around with the files that I've provided below.

Right-Click on any part and then choose "Normal" or "Short" or "Open" to see the effect on the circuit when that part is bypassed or removed or in the circuit with its assigned value ( feel free to change the values > just observe the net effect for the sake of education ).

Note ( re; File Management ) ; It's best to keep the primary files that I provided and the sub-set of ( frd + zma ) folders/files that XSim auto-creates ( upon opening & closing the main file ) all within the same folder.

http://www.audioheritage.org/photopost/data//500/TAD2001_2344_Mod-3134_.PNG

The pic above uses custom FRD + ZMA files that I created using some of Guidos graphs ( for his TAD 2001 on a 2344 from his 4435 rebuild project ).

The layout above is clearly inspired by the 3134 network ( with appropriate changes and some additions ).

I've included 2 XSim files for you to play with ( while the responses look almost identical they'll sound slightly different > & I can't tell you which I would prefer without actually building them and listening at length ).

Plus, since I don't own any TAD 2001 drivers, this is just an academic design exercise.

:)

[Jaska01]

If you want to make it sound way better, than consider buying a digital crossover. Those DSP ones are really low cost now days and you can fix response and timing way more correctly than in any passive crossover. In the other hand if you try just keep it original as possible than understandable.

[DogBox]

Hello,

I'm referring to the 4430's 3134 network schematics (http://www.jblproservice.com/pdf/Network%20Schematics/3134%20Network.pdf). . Ok, the HP itself is C11/L2 that drive an L pad. This L pad may be arranged as usual, i.e. a 8 Ω rheostat in series... Am I right?

Hi Kay,

I see you are still in love with "rheostats!" I'm not so sure about swapping out/rearranging them in place of a L-Pad is such a good idea, especially when you have so many Questions. I found the first two sub-headings in this explanation very useful :(Speaker L-Pad and Audio Frequency [AF] operation)... https://en.wikipedia.org/wiki/L_pad

Are they air cored?

If you look at the differing symbols on the schematic, you will see the 2.6mH [L102] and 1.6mH [L2] have two parallel lines above the coil representation - this suggests an iron-cored inductor whereas the plain 0.04mH [L1] is the air core inductor, coil.

Hope this helps a little. Re-creating a crossover exactly as per the schematic, is a great way to start 'before' trying to change anything. You may even answer your own questions by building one just like the original!

Kind Regards,
DogBox [just trying to help!] :)
.