mbeards
11-29-2013, 04:02 PM
Hi Everyone,
I wanted to share my progress with redesigning crossovers for my DIY JBL Hartsfield project to document and help others with crossovers as well as a sanity check for my own sake.
(However I feel my sanity is long gone for taking on this project in the first place. ....:D )
My goal is a maximum flat design with a Q of 0.5 for each filter
I am starting with the JBL 3115 (500Hz) and N7000 (7KHz) crossovers.
For drivers I am currently using 130A woofer, 2441 horn, and 075 tweeter for a three way system. I soon realized that the E145 (with suggestions form the forum) would be a much better choice.
I don't have them in yet so I am staring from the HF crossover point down.
I recently purchased the DATS speaker impedance analyzer from Dayton audio to aid with the design.
The 2441 impedance curve looks predictably inductive at high frequencies. DATS indicates and Re = 7.5ohms and Ie = .176mH. I modeled this in SPICE and sure enough, Re in series with Ie created a spot on model for the impedance of this driver. Note: this is only accurate for the higher frequencies. At lower frequencies, there are a lot of resonances most likely due to the horn and lens assembly. However, for the 7KHz crossover, the inductive model is what I will use. Due to the non negligible inductive rise, I added a Zobel consisting of a 9.5ohm and a 2uf cap.
The 075 impedance curve at the frequencies of interest are more "periodic" and not as predictable as the 2441. Therefore I modeled the 075 as an average fixed resistance load for the crossover network. I plan to use an 8ohm L-pad (as in the N7000) and since the 075 average impedance is close to 8ohms, I modeled the load as just 8ohms.
2441 w/ Hartsfield horn and lens Impedance Plot:
60757
075 Impedance Plot:
60758
Crossover Circuit (Ignore the rise in the LF 2441 driver since this has yet to be optimized with the e145):
60759
Response Curve:
60760
Next step is modeling of the LF 500Hz crossover point from the E145 to the 2441. For this, the 2441 is no longer described by an inductive model due to multiple resonances. For the LF design, I will use an average resistance model (as with the 075 in the HF Crossover) of around 17ohms.
Any comments on changing the impedance model of the drivers for the frequencies in question?
That is all for now,
Matt
I wanted to share my progress with redesigning crossovers for my DIY JBL Hartsfield project to document and help others with crossovers as well as a sanity check for my own sake.
(However I feel my sanity is long gone for taking on this project in the first place. ....:D )
My goal is a maximum flat design with a Q of 0.5 for each filter
I am starting with the JBL 3115 (500Hz) and N7000 (7KHz) crossovers.
For drivers I am currently using 130A woofer, 2441 horn, and 075 tweeter for a three way system. I soon realized that the E145 (with suggestions form the forum) would be a much better choice.
I don't have them in yet so I am staring from the HF crossover point down.
I recently purchased the DATS speaker impedance analyzer from Dayton audio to aid with the design.
The 2441 impedance curve looks predictably inductive at high frequencies. DATS indicates and Re = 7.5ohms and Ie = .176mH. I modeled this in SPICE and sure enough, Re in series with Ie created a spot on model for the impedance of this driver. Note: this is only accurate for the higher frequencies. At lower frequencies, there are a lot of resonances most likely due to the horn and lens assembly. However, for the 7KHz crossover, the inductive model is what I will use. Due to the non negligible inductive rise, I added a Zobel consisting of a 9.5ohm and a 2uf cap.
The 075 impedance curve at the frequencies of interest are more "periodic" and not as predictable as the 2441. Therefore I modeled the 075 as an average fixed resistance load for the crossover network. I plan to use an 8ohm L-pad (as in the N7000) and since the 075 average impedance is close to 8ohms, I modeled the load as just 8ohms.
2441 w/ Hartsfield horn and lens Impedance Plot:
60757
075 Impedance Plot:
60758
Crossover Circuit (Ignore the rise in the LF 2441 driver since this has yet to be optimized with the e145):
60759
Response Curve:
60760
Next step is modeling of the LF 500Hz crossover point from the E145 to the 2441. For this, the 2441 is no longer described by an inductive model due to multiple resonances. For the LF design, I will use an average resistance model (as with the 075 in the HF Crossover) of around 17ohms.
Any comments on changing the impedance model of the drivers for the frequencies in question?
That is all for now,
Matt