CauYem
12-02-2008, 02:51 PM
I build this network based on the simulations on "13.5 uF" thread. I bought parts from Part Express. Each components were individually measured (using a digital Philips LCR) to make sure they are within the specs, and also recorded its ESR & Q for simulations. They are within 2% of the nominal values, some are right on (Dayton Caps). I used (2) 440uF polarized cap in series (+--+) to make a 200uF cap (actual measurement is 216.3 uF @1KHz) since (2) Dayton 100uF took half of the board area. I'll repalce it with a 200uF bipolar later.
This is my first time ever to build an audio crossover networks, even though I have build many filters & dividing networks (Multiplexers) for military & space programs (up to 5 channels/5-ways & frequency up to 34 GHz). This circuit based on the same techniques I used for communication systems, which provide a constant input impedance & steeper slope/octave with minimal components used.
All the components are standard values (1.8, 2.4, 3.3, 4.7 .. 68, 82... not odd number like "13.5 uF":)). The simulations of the circuit are on the "13.5 uF" thread along with both JBL L300 networks (regular & "charge coupled").
Each crossover took almost half of a day to build since I have to work off & on, and laid them out on the board (10" X 6" X 0.5" piece of wood) as I soldered them along. I should have measured each component dimensions then layout on CAD program so it doesn't take up a lot of area on the board. I wound the 0.068 mH myself so it doesn't look as pretty as the inductors I bought. They don't sell the 0.068mH. The network will replace the JBL network in my pseudo L300 (L200 + N8000 + 077).
I was worry & anxious before I plug it in. What happen if it doesn't work?. Then I removed the JBL LX16 & N8000, and connect all the wires, and turn on the test CD (Hary Belafonte Live in concert at the Carnegie Hall). My worry was gone when it singed beautifully. Even my wife noticed the different between the two. The mid from my network is a bit dominant, so I lowered it about 1.5 dB. Overall sound level is higher than the JBL's, I think the hi-Q components cause less loss on transmission, thus more power to the drivers.
I wish I have room to set up the RTA lab to measure the exact electrical & acoustical performance of my network even I know the performance will be very close (if not exact) as in simulation.
I own many audio systems & speakers, but never thought I would tinker with the Xover network before. I assumed that the speaker manufactures have the optimal condition for their products, and thought those DIY are kinda nuts. Until I came across these audio forums and made contact with those who like sound production like me, but has all kind of knowledge & experience with sound. Now, I officially joined the group of nuts.
I've spent sometimes to check out most of passive crossover networks from different manufactures that are on the net. JBL has the most networks posted. There are much room for improvements.
This is my first time ever to build an audio crossover networks, even though I have build many filters & dividing networks (Multiplexers) for military & space programs (up to 5 channels/5-ways & frequency up to 34 GHz). This circuit based on the same techniques I used for communication systems, which provide a constant input impedance & steeper slope/octave with minimal components used.
All the components are standard values (1.8, 2.4, 3.3, 4.7 .. 68, 82... not odd number like "13.5 uF":)). The simulations of the circuit are on the "13.5 uF" thread along with both JBL L300 networks (regular & "charge coupled").
Each crossover took almost half of a day to build since I have to work off & on, and laid them out on the board (10" X 6" X 0.5" piece of wood) as I soldered them along. I should have measured each component dimensions then layout on CAD program so it doesn't take up a lot of area on the board. I wound the 0.068 mH myself so it doesn't look as pretty as the inductors I bought. They don't sell the 0.068mH. The network will replace the JBL network in my pseudo L300 (L200 + N8000 + 077).
I was worry & anxious before I plug it in. What happen if it doesn't work?. Then I removed the JBL LX16 & N8000, and connect all the wires, and turn on the test CD (Hary Belafonte Live in concert at the Carnegie Hall). My worry was gone when it singed beautifully. Even my wife noticed the different between the two. The mid from my network is a bit dominant, so I lowered it about 1.5 dB. Overall sound level is higher than the JBL's, I think the hi-Q components cause less loss on transmission, thus more power to the drivers.
I wish I have room to set up the RTA lab to measure the exact electrical & acoustical performance of my network even I know the performance will be very close (if not exact) as in simulation.
I own many audio systems & speakers, but never thought I would tinker with the Xover network before. I assumed that the speaker manufactures have the optimal condition for their products, and thought those DIY are kinda nuts. Until I came across these audio forums and made contact with those who like sound production like me, but has all kind of knowledge & experience with sound. Now, I officially joined the group of nuts.
I've spent sometimes to check out most of passive crossover networks from different manufactures that are on the net. JBL has the most networks posted. There are much room for improvements.