Active crossovers are actively discussed but I have not found any directions on module arrangement.
Look at the attached picture. Until today I've planned the serial way, designed neat HP/LP crossover modules etc. But then I started to question myself. That's what I came up with:
Serial arrangement
Pros
neat structure
phase-coherent
Cons
possibly higher distortion due to number of active elements in signal path
Parallel arrangment:
Pros
possibly lower distortion
Cons
not so modular (ie modules and wiring become a bit more complex)
possible phase incoherencies (not sure about this one)
So, which approach is better in Your opinion and why?
Look at the attached picture. Until today I've planned the serial way, designed neat HP/LP crossover modules etc. But then I started to question myself. That's what I came up with:
Serial arrangement
Pros
neat structure
phase-coherent
Cons
possibly higher distortion due to number of active elements in signal path
Parallel arrangment:
Pros
possibly lower distortion
Cons
not so modular (ie modules and wiring become a bit more complex)
possible phase incoherencies (not sure about this one)
So, which approach is better in Your opinion and why?
Attachments
I tend to prefer what you describe as a parallel arrangement for the reason you stated as a con in the serial category - it has more active devices and potential distortion in the signal path. the M-T low pass does nothing for the woofer circuit, in typical corssovers the woofer is already down >30 dB by the M-T XO freq.
Also, the parallel arrangement makes phase compensation easier to do.
Note that Linkwitz does a little serial work in his active XO - incorporating baffle step compensation into his input buffer on his MT1 board, so that the tweeter gets the compensation, too. He also takes the tweeter from signal after the midrange's HP filter on his ASP (Orion XO) board. this is a departure from his early active XO articles in Speakerbuilder. In that article the tweeter XO was a standard 12db/octave dilter and a shelving 12 dB/octave filter from the Xo point to the tweeter resonant frequency, so that the acoustic slope would continue at 24 dB/octave. the current design seems to show that this was less important than he first thought, since he now has a second discontinuity in the tweeter's high pass slope.
From this I'd say that there is no single right way, just the one that works best for your application. With decent opamps and board design either configuration should work.
Also, the parallel arrangement makes phase compensation easier to do.
Note that Linkwitz does a little serial work in his active XO - incorporating baffle step compensation into his input buffer on his MT1 board, so that the tweeter gets the compensation, too. He also takes the tweeter from signal after the midrange's HP filter on his ASP (Orion XO) board. this is a departure from his early active XO articles in Speakerbuilder. In that article the tweeter XO was a standard 12db/octave dilter and a shelving 12 dB/octave filter from the Xo point to the tweeter resonant frequency, so that the acoustic slope would continue at 24 dB/octave. the current design seems to show that this was less important than he first thought, since he now has a second discontinuity in the tweeter's high pass slope.
From this I'd say that there is no single right way, just the one that works best for your application. With decent opamps and board design either configuration should work.
- Status
- Not open for further replies.