Help needed from you network wizards

[Guido]

In one of the previous threads I posted about my idea about building a 4435 based Studio Monitor. Up to 1 kHz everything should be the same then in the 4435. This because it is said that the 4435 low end is very impressive and not least because 4 pcs. of 2234H speakers are waiting to play.
The region above 1 kHz is improvable. I'm even more convinced about that after having a informative talk with a very knowledgeable technic guy from JBL Germany at the Frankfurt music fair last weekend. He told me that the combination of the newer waveguides from the Application Engineered series (especially the 100°x100° version) with the new 2431 Driver will be superior to the 2344 / 2426 combination. Also he recommends not to go with the Horn EQ from the 4435 network although this should be possible.
He recommended to go with high pass filter from the network of the AM6212_00 (file see below) modify the steepness of the xover point but keep the excellent EQ.

Heres what I did with that information and what for I need your input, suggestions, opinions etc.....

First of all respect that until know everything is theoretical as the waveguides and 2431 drivers will show up earliest in 3 to 4 weeks. Then the testing will start.

1.
I modeled the HF section from the 4435 network in circuit maker. This is the result:

[Guido]

Closer look around 1kHz

[Guido]

2. I modeled the HF section from the AM6212_00 in circuit maker. This is the result:

The shape in the lower region is different from the original 4435 but I think the HF EQ is nice and it will fit better to the 2431 driver as it is made for it :-)

[Guido]

Again closer look around 1kHz

[Guido]

3. I modified the AM6212_00 HF section.

Here's the result:

[Guido]

closer look around 1 kHz and fitting:)

[Guido]

5. As circuit maker doesn't support impedance simulation I hacked everything in a speaker simulator program (WinBoxSimu). I'm not sure with this. At least this should be checked from you.
Here's the result:

[Guido]

What do you guys think about it????

[4313B]

"As circuit maker doesn't support impedance simulation"

Set it up just like you would measure impedance in real life. Use a 1000 ohm resistor. Calibrate the voltage source so an 8 ohm resistor measures 8 mV.

[4313B]

3134 and 3135 voltage drives
Red traces are with L-Pads set to "0"

[4313B]

3134 and 3135 simulation schematics

[4313B]

"I modeled the HF section from the 4435 network in circuit maker. This is the result:"

R1 and R2 should come after R6, not before it. In your model, if you have R1 set to 2 ohms then R2 would be ~ 24 ohms and both would come after R6 instead of before it.

Rotary L-Pads act the same way a hard wired L-Pad works. An "8 ohm" L-Pad means the L-Pad is looking for an 8 ohm load behind it so that it presents an 8 ohm load at it's input. So, with the knob at a particular position on the L-Pad if the series portion measures ~2 ohms the parallel portion would then measure ~24 ohms. With an 8 ohm load "behind" the L-Pad the result would be an 8 ohm load "in front" of the L-Pad. 2 + (24 * 8) / (24 + 8). I had to learn this the hard way early on when I was playing around with the then brand new 4315 (~1977) and it's 16 ohm L-Pad in front of an 8 ohm load :D

[4313B]

On the subject of L-Pads... here's a trick for bandpass circuits courtesy of G.T. at JBL.

Build your hard wired L-Pad attenuation for your bandpass driver. Let's say you have a 108 dB horn/cd and you need it padded to 94 dB. Build the neccessary L-Pad. Now build the high pass filter in front of the L-Pad like usual but put the low pass after the L-Pad. That way they can be tweeked independently of each other :) Very cool!

[Guido]

Thanks Giskard for correction!

Just didn't have the experience yet.

But what do you think about the implementation of the 2431 and the slightly different HF section from the AM6212_00?

[4313B]

I think anything that can improve the 23 year old 4430/4435 is a good thing. Let us know what you come up with :)